Quantitative Risk Assessment of Hazardous Materials Transport Systems

QUANTITATIVE RISK ASSESSMENT OF
HAZARDOUS MATERIALS TRANSPORT SYSTEMS

TOPICS IN SAFETY, RISK, RELIABILITY AND QUALITY
VOLUME 5
Editor
A.Z. Keller, Department of Industrial Technology and Management,
University ofBradford, U.K.
Editorial Advisory board

P. Sander, Technical University of Eindhoven, The Netherlands
D.C. Barrie, Lakehead University, Ontario, Canada
A.V. Gheorghe, Swiss Federal Institute of Technology, Zurich, Switzerland
R. Leitch, Royal Military College of Science (Cranfield), Shriverham, U.K.
Aims and Scope. Fundamental questions which are being asked these days of all
products, processes and services with ever increasing frequency are:
What is the risk?
How safe is it?
How reliable is it?
How good is the quality?
How much does it cost?
This is particularly true as the government, industry, public, customers and society
become increasingly informed and articulate.
In practice none of the three topics can be considered in isolation as they all interact
and interrelate in very complex and subtle ways and require a range of disciplines
for their description and application;, they encompass the social, engineering and
physical sciences and quantitative disciplines including mathematics, probability
theory and statistics.
The major objective of the series is to provide a series of authoritative texts suitable
for academic taught courses, reference purposes, post graduate and other research
and practitioners generally working or strongly associated with areas such as:
Safety Assessment and Management
Emergency Planning
Risk Management
Reliability Analysis and Assessment
Quality Assurance and Management
Special emphasis is placed on texts with regard to readability, relevance, clarity,
applicability, rigour and generally sound quantitative content.
The titles published in this series are listed at the end of this volume.
Quantitative Risk Assessment
of Hazardous Materials
Transport Systems
Rail, Road, Pipelines and Ship
by
MICHEL NICOLET-MONNIER
PSI. Paul Scherrer Institute. Villigen. Switzerland
and
ADRIAN V. GHEORGHE
ETHZ. Swiss Federal Institute ofTechlWlogy. Zurich. Switzerland
SPRINGER-SCIENCE+BUSINESS MEDIA, B.V.

A C.I.P. Catalogue record for this book is available from the Library of Congress.
ISBN 978-90-481-4666-6 ISBN 978-94-017-2821-8 (eBook)

DOI 10.1007/978-94-017-2821-8
Printed on acid-fru paper
Ali Rights Reserved

© 1996 Springer Science+Business Media Dordrecht
Originally published by Kluwer Academic Publishers in 1996
Softcover reprint ofthe hardcover Ist edition 1996
No part of the material protected by this copyright notice may be reproduced or
utiJized in any form or by any means, electronic or mechanical,
including photocopying, recording or by any information storage and
retrieval system, without written permis sion from the copyright owner.
CONTENTS
PREFACE xm
FOREWORD xvn
ACKNOWLEDGEMENTS XIX
LIST OF ABBREVIATIONS XX
CHAPTER 1: ANALYSIS AND ASSESSMENT OF TRANSPORTATION

RISK ·ENVIRONMENTALANDSAFETYFACTORS 1
1.1. A Methodological Overview 1
1.1.1. GATHERING INFORMATION ON THE REGION OF
INTEREST 4
1.1.2. SELECfING DANGEROUS GOODS FOR REGIONAL
RISK ASSESSMENT 6
1.1.3. THE NATURE OF THE MAJOR RISK TO BE ASSESSED 8
1.1.4. MANDATORYROUTINGFACTORS 11
1.1.5. ENVIRONMENTAL AND LAND USE SAFETY FACTORS 12
1.1.6, SUBJECTIVE FACTORS 23
1.1.7. IMPLEMENTATION GUIDANCE 24
1.2. Traffic and Operational Factors for Road and Rail Transport 25
1.2.1. TRAFFIC VOLUME AND COMPOSmON 25
1.2.2. CARRIAGE-WAY LEVEL OF SERVICE 25
1.2.3. STRUCTURAL AND GEOMETRIC ADEQUACY OF
ROADS 26
1.2.4. NUMBER OF TRAFFIC SIGNALS 27
1.2.5. TRAVELTIME AND TRAVEL SPEED 27
1.2.6. AVAILABILITY OF ALTERNATIVE EMERGENCY
ROUTE 27
1.3. Procedures for the Development of Estimated Truck Accident
Rates 28
1.4. Risk Management for Dangerous Goods 32
1.4.1. A PATHWAY APPROACH APPLYING TO THE
TRANSPORTATION OF DANGEROUS GOODS 34
1.4.2. TRANSPORT OPERATIONAL COSTS AND OPERATOR'S
REQUIREMENTS 34
Summary 3S
References (Chapter 1) 3S
CHAPTER 2: ASSESSMENT OF TRANSPORTATION RISKS IN A

COMPLEX AREA 37
2.1. Overview and Scope of Transportation Risk Analysis 37
2.2. Risk Estimation and Risk Assessment 38
VI QRA OF HAZARDOUS MAlERIALS TRANSPORT SYSlEMS
2.2.1. QUANTITATIVE RISK CRI1ERIA 38

2.2.2. DEFINmON OF INDIVIDUAL AND SOCIETAL RISK
CRI1ERIA 40
2.2.3. ENVIRONMENTAL RISK CRI1ERIA 43
2.2.4. HOW TO COMBINE CONSEQUENCES AND FREQUENCY
ESTIMAlES TO PRODUCE RISK MATRICES 44
2.2.5. DEFINmON AND ACCEPTANCE OF RISK LEVELS 49
2.2.5. FINAL ASSESSMENT OF RISK LEVELS 56
2.3. Information and Data for Assessing Transportation Risks 56
2.3.1. AVAILABILITY AND COMPARABILITY OF
STATISTICAL DATA 57
2.3.2. COMPARING ALlERNATIVE ROU1ES VS. COMPARING
ALlERNATIVE TRANSPORTATION SYSTEMS 59
2.4. Classification of Regional Risk and Order of Priority 62
2.5. Hazards Resulting from the Transport of Dangerous Goods 65
2.5.1. TRANSPORT OF CHEMICALS 66
2.5.2. FREQUENCY ANALYSIS 70
2.5.3. FAILURE FREQUENCIES FOR RAILROAD TRANSPORT 71
2.5.4. BULK TRANSPORT MODES 71
2.6. Legislation Concerning Transportation of Dangerous Goods by
Rail, Road, and Air 74
2.6.1. INTERNATIONAL REGULATIONS 74
2.6.2. PACKAGING AND LABELING 75
2.6.3. SWISS REGULATIONS FOR DANGEROUS GOODS
TRANSPORTATION ON LAND 76
TRANSPORTATION BY AIR 77
TRANSPORTATION ON W ATERW AYS 78
2.6.6. REMARKS CONCERNING THE REVISION OF THE
REGULATIONS AND ORDINANCES AND THEIR
APPLICATION 79
2.6.7. TRANSPORTATION OF RADIOACTIVE SUBSTANCES 79
2.7. BUWAL's Guidelines for Assessing the Risks of Traffic Ways in
Switzerland 82
2.7.1. LEGAL BASIS TO THE SWISS "REGULATION
CONCERNING INCIDENTS" 82
2.7.2. MAIN TOPICS OF THE SWISS "REGULATION
CONCERNING INCIDENTS" 83
2.7.3. HANDBOOK III, TO THE "REGULATIONS CONCERN-
ING INCIDENTS, StFV - GUIDING PRINCIPLES FOR
TRAFFIC WAYS" 83
2.7.4. DUTIES OF THE EXECUTIVE AUTHORITIES 85
2.7.5. LIMITATION OF QUANTITATIVE RISK ANALYSIS IN
DANGEROUS GOODS TRANSPORTATION SYSTEMS 86
CONTENTS VII
2.7.6. REMARK CONCERNING THE BUWAL METHOD 87

Summary 87
References (Chapter 2) 88
CHAPTER 3: RISK ANALYSIS OF HAZARDOUS MATERIALS

TRANSPORTATION BY ROAD 91
3.1. Information and Knowledge about Truck Accidents and
Consequences 91
3.1.1. EXPERIENCE CONCERNING TRUCK ACCIDENTS 91
3.1.2. GASOLINE TRANSPORTS BY ROAD 94
3.1.3. FACTORS WHICH CONTRIBUTED TO GASOLINE AND-
LPG TRUCK RELEASES (CANADA EXPERIENCE) 95
3.2. Risk Assessment of the Transportation of Hazardous Materials
Through Road Tunnels 96
3.2.1. U.K. STUDY OF THE RISK OF ROAD TUNNELS 96
3.2.2. U.S. STUDY OF THE RISK OF ROAD TUNNELS 100
3.2.3. SWISS STUDY OF THE RISK OF ROAD TUNNELS 101
3.3. BUWAL's Methodology for Assessing the Risks of Hazardous
Materials Transportation by Road in Switzerland 101
3.3.1. SUBDIVISION OF THE ROAD TRACKS INTO ROAD
SEGMENTS 102
3.3.2. INFORMATION AND DATA CONCERNING THE
ENVIRONMENT 103
3.3.3. INFORMATION CONCERNING TRAFFIC RISE, TRAFFIC
STRUCTURE AND ACCIDENT HISTORY 105
3.3.4. ESTIMATION OF THE LIKELIHOOD OF AN INCIDENT
WITH SEVERE CONSEQUENCES TO PEOPLE OR
ENVIRONMENT 106
3.3.5. ASSUMPTIONS CONCERNING THE REPRESENTATIVE
INCIDENT SCENARIO FOR THE "POPULATION" 109
INCIDENT SCENARIO FOR "GROUNDWATER
CONTAMINATION" 113
SCENARIO FOR "SURFACE WATERS" 114
3.4. New Developments in Traffic and Vehicle Control 116
Summary 117
CHAPTER 4: RISK ASSESSMENT OF HAZARDOUS MATERIALS

TRANSPORTATION BY RAIL 121
4.1. U.S. Experience with Train Accidents 121
4.1.1. RAILACCIDENTRATES 121
4.1.2. BULK TRANSPORT BY TRAIN 125
4.2. British Experience with Train Accidents 128
VIII QRA OF HAZARDOUS MATERIALS TRANSPORT SYSTEMS
4.2.1. MAIN CAUSES OF SPILLAGE 128

4.2.2. ACCIDENT SPEED OF FREIGHT TRAINS 130
4.2.3. RESULTS FROM THE HSC-STUDY 130
4.3. Information Concerning the Swiss Federal Rail Network 134
4.3.1. CONFIGURATION OF THE SWISS RAIL NETWORK 134
4.3.2. SWISS RAILROAD STATISTICS 135
4.3.3. TRAFFIC STRUCTURE 136
4.4. BUWAL's Methodology for Assessing the Risks of Hazardous
Materials Transportation by Rail in Switzerland 136
4.4.1. TECHNICAL ASPECTS OF THE DETAILED RISK
ASSESSMENT 136
4.4.2. ESTIMATION OF THE LIKELIHOOD OF AN INCIDENT
WITH SEVERE CONSEQUENCES FOR THE
POPULATION OR THE ENVIRONMENT 140
INCIDENT SCENARIOS INVOLVING THE POPULATION 143
INCIDENT SCENARIO "GROUNDWATER RESOURCES" 144
INCIDENT SCENARIO "SURFACE WATERS" 146
4.5. Probability of Incidents on the Swiss Rail Network 147
4.5.1. SWISS RAILROAD ACCIDENT STATISTICS 148
4.5.2. ESTIMATION OF INCIDENTS FREQUENCY ON THE
SBB RAIL NETWORK 148
4.5.3. DEFINITION OF INDICATORS FOR ASSESSING THE
EFFECTS FROM INCIDENTS 149
4.5.4. DISTRffiUTION OF INCIDENT FREQUENCIES
ACCORDING TO RAIL TRACK TYPE 151
Summary 154
CHAPTERS: TRANSPORTATION OF DANGEROUS GOODS BY

SHIP 157
5.1. Risk Resulting from Ship and Barge Transport on Lakes and
Waterways 157
5.2. Techniques for Risk Assessment of Ships 158
5.2.1. OBJECTIVES 158
5.2.2. METHODOLOGY FOR ASSESSING THE RISK OF
INDIVIDUAL PORTS 159
5.2.3. FREQUENCY ESTIMATION FOR SHIP TRANSPORTS 163
5.2.4. PROBABILISTIC APPROACH 165
5.2.5. FAILURE CASES AND RELEASE SCENARIOS 171
5.2.6. POLLUTION OF WATERW AYS BY DANGEROUS
CARGO 173
5.3. Risk Management for Waterways and Ship Cargo 175
CONTENTS IX
5.4. Swiss Waterways and the Ports of Both Basle 175

5.4.1. HISTORICAL BACKGROUND AND ORGANIZATION 175
5.4.2. SHIP TRAFFIC IN THE PORTS OF BOTH BASLE 176
Summary 177
CHAPTER 6: PIPELINES AND STORAGE TANKS FOR GASES AND

PETROLEUM PRODUCTS 181
6.1. Transport Pipelines in Switzerland 181
6.2. Risk from Pipelines 187
6.2.1. EXPERIENCE AND SITUATION IN THE USA 187
6.2.2. PIPELINES INCIDENTS IN SWITZERLAND 192
6.2.3. HAZARD CONTROL 192
6.3. Storage Tanks 192
6.3.1. TYPES OF STORAGE 193
6.3.2. STORAGE OF SPECIAL PRODUCTS 197
Summary 199
CHAPTER 7: TRANSPORTATION OF DANGEROUS GOODS-

GENERAL INFORMATION AND STATISTICS 201
7.1. Information for Regional Planing 201
7.2. Dangerous Goods Movements in Switzerland 202
7.2.1. SOURCES OF STATISTICAL DATA 202
7.2.2. DANGEROUS GOODS CLASSES 204
7.2.3. TYPE AND QUAN;fITY OF GOODS TRANSPORTED IN
SWITZERLAND 206
7.2.4. STATISTICS CONCERNING PETROLEUM PRODUCTS 216
7.3. Traffic Accidents Statistics Involving Dangerous Goods 221
7.3.1. TRAFFIC ACCIDENTS STATISTICS (SWITZERLAND) 221
7.3.2. TRAFFIC ACCIDENTS STATISTICS (WEST GERMANY) 222
7.3.3. OTHER ACCIDENT STATISTICS INVOLVING
DANGEROUS GOODS TRANSPORTS 226
7.4. RaillRoad Transport Containers 228
7.4.1. INFORMATION CONCERNING SWITZERLAND 228
7.4.2. RELIABILITY OF TANK/COMPONENTS FAILURE 231
7.4.3. UNLOADING TERMINALS 233
Summary 233
CHAPTER 8: DATA BASES AND COMPUTER SUPPORT FOR RISK

ASSESSMENT 235
8.1. Data Bases for Hazardous Chemicals and Materials 235
8.1.1. DESCRIPTION OF SELECTED DATA BASES 236
X QRA OF HAZARDOUS MATERIALS TRANSPORT SYSTEMS
8.2. U.S. Transportation Statistics and Information on Commodity

Flow 242
8.2.1. TRANSPORTATION STATISTICS ANNUAL REPORT 242
8.2.2. NATIONAL TRANSPORTATION STATISTICS 242
8.3. Equipment ReUability Data Bases 243
8.4. Center for Exposure Assessment Modeling (CEAM) 244
8.5. Transportation Risk Management System 244
8.6. Spatial Data and Geographic Information Systems Within the
U.S. Bureau of Transportation Statistics (BTS) 245
8.7. Computer Support for Risk Assessment 249
8.7.1. COMPUTER PACKAGES FOR HAZARD ANALYSIS 249
8.7.2. TRANSPORTATION RISK AND COST ANALYSIS
MODELS 251
8.8. Special Literature and Risk Assessment Studies 254
8.8.1. GIS MAPPING AND SPATIAL ANALYSIS 255
8.8.2. GIS AND TRANSPORTATION 255
8.8.3. SPECIAL METHODS FOR MODELING AND
EVALUATING TRANSPORTATION RISKS 256
Summary 257
CHAFfER 9: TOWARDS INTEGRATED RISK ASSESSMENT AND

SAFETY MANAGEMENT AT REGIONAL LEVEL:
- TRANSPORTATION OF DANGEROUS GOODS 261
9.1. Hazard Assessment Considerations 261
9.1.1. INVESTIGATING AND ASSESSING THE RISK 261
9.1.2. ASSESSING THE CONSEQUENCES 262
9.2. The Potential Use of Comparative Assessment of Various
Transportation Systems 263
9.2.1. SE1TING BOUNDARIES FOR COMPARATIVE RISK
ASSESSMENT OF DIFFERENT TRANSPORTATION
TECHNOLOGIES FOR REGIONAL SAFETY
MANAGEMENT 264
9.2.2. ASSESSMENT OF ENVIRONMENTAL RISKS (IMPACTS) 265
9.2.3. COMPARATIVE HEALTH RISK ASSESSMENT 268
9.2.4. UNCERTAINTIES IN COMPARATIVE RISK
ASSESSMENT 269
9.2.5. UNCERTAINTY ESTIMATION IN ENVIRONMENTAL
REGIONAL RISK ASSESSMENT AND THE
TRANSPORTATION OF DANGEROUS GOODS 270
9.2.6. INTEGRATED ENVIRONMENTAL AND HEALTH RISK
ASSESSMENT IN LARGE INDUSTRIAL AREAS 273
9.2.7. INTEGRATION OF THE DIFFERENT ELEMENTS OF
RISK AND THE ROLE OF COMPARATIVE RISK
ASSESSMENT FOR SAFETY MANAGEMENT 275
CONTENTS XI
9.3. Some Methodological Issues in CRA for Severe Accidents 278

9.3.1. METHODS AND TOOLS OF COMPARATIVE RISK
ASSESSMENT FOR SEVERE ACCIDENTS 278
9.4. Basic Aspects in Decision Aiding Techniques 279
9.4.1. OVERVIEW OF DECISION AIDING TECHNIQUES 282
9.4.2. DECISION AIDING TECHNIQUES IN USE FOR
INTEGRATED RISK ASSESSMENT AND SAFETY
MANAGEMENT 283
9.4.3. SENSmVITY ANALYSIS 284
9.4.4. PRESENTATION OF THE RESULTS FROM APPLYING
A DECISION AIDING TECHNIQUE 285
9.4.5. ADVANCED TECHNIQUES FOR DAP 285
9.5. Loss Prevention and its Economic Aspects 287
9.5.1. CAUSEOFLOSSES 287
9.5.2. ECONOMICS OF LOSS PREVENTION 288
9.6. Risk Communication 290
Summary 293
CHAPTER 10: CASE STUDIES IN REVIEW AND MISCELLANEOUS

INFORMATION 295
10.1. The Vancouver Case Study 295
10.1.1. THE MODEL 295
10.1.2. FEATURES OF THE COMPUTERIZED MODEL 298
10.2. Transport Study for the Lyon Area 299
10.3. Assessment of Tran~portation Risk for the Canton "Basle·
Landscape'" 361
10.4. Transport Study for the Road Tunnel Baregg 307
10.5. Transport Study for the Canton Uri 311
10.6. The Project STRADA·DB 319
10.7. Companies Profiles and Miscellaneous Information 320
10.7.1. SOFTWARE ENGINEERING COMPANIES 320
10.7.2. MISCELLANEOUS INFORMATION 322
10.8. Swiss Cantonal Authorities Involved with Dangerous Goods
Transportation 324
10.8.1. ADDRESSES OF THE EXECUTING AUTHORITIES:
SWISS CONFEDERATION 324
10.8.2. ADDRESSES OF THE CENTRALIZED CANTONAL
INCIDENT REPORTING OFFICES FOR PROTECTION
AGAINST CATASTROPHES 326
Summary 327
SUBJECT INDEX 331
PREFACE
Industrial development is essential to improvement of the standard of living in all coun-

tries. In a given region, old and new plants, processes, and technologies have to
coexist Technological penetration and substitution processes are generally taking place;
they are entirely dynamic and this trend is going to stay like this.
People's health and the environment can be affected, directly or indirectly, by rou-
tine waste discharges or by accidents. A series of recent major industrial accidents and
the effect of poUution highlighted, once again, the need for better management of rou-
tine and accidental risks. Moreover, the existence of natural hazards complicate even
more the situation in any given region.
Managing the hazards of modern technological systems has become a key activity in
highly industrialized countries. Decision makers are often confronted with complex
issues concerning economic and social development, industrialization and associated
infrastructure needs, population and land use planning. Such issues have to be ad-
dressed in such a way that ensures that public health wiD not be disrupted or substan-
tially degraded.
While hazard managers and risk assessors have been successful at identifying haz-
ards and reducing overall risk exposure, economic growth and technological develop-
ment have led to a new risk situation characterized by:
- an increasing number and variety of hazards
- hazards giving rise to a broad range of partial and temporal risks
- public dissatisfaction with hazards managers and hazards' owners
- deregulation mechanisms
Due to the increasing complexity of technological systems and the higher geo-
graphical density of punctual hazard sources, new methodologies and a novel approach
to these problems are challenging risk managers and regional planers. Risks from these
new complex technological systems are inherently different from those addressed by the
risk managers of the 1960's and 70's. Recent awareness of environmental problems by a
large public has led worldwide to dissatisfaction and to the formation of all kinds of
pressure groups that do exert a strong political influence and are quite often not ready to
accept any compromise. As a result of such approach, interesting and vital research
projects were stopped (~wiss nuclear moratorium), plans are buried and decisions de-
layed for many years (i.e., repository for nuclear wastes). Nowadays it becomes increas-
ingly difficult to site new plant facilities which are perceived as risky or undesirable by
the local population. Projects of national interest cannot be launched or even achieved,
once started. On the other hand distrust has become so problematic, that at times, risk
managers are no more considered as an impartial person, or as reliable source of infor-
mation regarding risk protection aspects.
XN PREFACE
It was felt that existing hazard management techniques need to be supplemented

with concepts and methods that are integrative at a regional level. Integrated regional
risk assessment and safety/hazard management (lRRASM) represent a coordinated
strategy for risk reduction and safety/hazard management in a spatially-defined region
across a broad range of hazard sources (during normal operation and accidental situa-
tions) that includes synergistic effects.
Integrated Regional Risk Assessment and Sa/ety/Hazard Management (IRRASM)

IRRASM is a multi-disciplinary process: engineers. computer scientists and modelers play
a central role in the risk assessment stage. Social scientists can contribute with practical
advice to the embedding process concerning hazard sources and help communal organiza-
tions to deal with such problems. taking into account local economic conditions and politi-
cal reality. They can make important contributions towards a better understanding of how
the practice of regional risk management can meet the needs and expectations of
society/community. In this way other kinds of regional planing tasks (for emergency pre-
paredness. or remedial actions) or the implementing of utilities as part of the infrastructure
(such as canalization network, sewage plants. waste disposal units. power plants. drinking
water distribution. etc.) can be integrated more easily after reaching the consensus of
agreement of the population.
In view of the above mentioned, a joint project, PPR&S (Polyproject on "Risk and
Safety of Technical Systems) was launched by ETHZ (Swiss Federal Institute of
Technology, Ziirich), together with the participation of the following institutions:
• Paul Scherrer Institute, PSI (ViUigen, Switzerland)
• EAWAG (Diibendorf, Switzerland)
There arc a number of national and international efforts to deal with risk assessment
at the regional level. The "Poiyproject on Risk and Safety of Technical Systems". at
ETHZ, took the initiative to research on various aspects related to regional risk
assessment and safety management A series of basic questions were asked, and answers
had to be given; they had to comprehend the following main issues:
what is integrated area risk assessment and safety management
how to define a region/area for study
type of activities and targets at risk
the need for risk impact indicators
the need for a comprehensive methodology.
Regional risk assessment and safety management seems to be a medium which helps
to integrate people. issues and decisions in area risk assessment
PPR&S is the discursive description of knowledge in addition to the development of
interdisciplinary and practical methods for the application of risk management for a
large variety of technological systems (e.g .• rail and road transportation of dangerous
goods, chemical plants, nuclear power plants, biotechnology, landfill, etc.). The scope
of these applications is to be presented as a book series. It is intended to be an integrated
PREFACE xv
regional risk assessment and safety management guideline manual complemented by

specialized software, databases, and literature reviews. The Polyproject is aimed at the
development of a novel methodological framework, with due regard to the existing
conditions prevailing in Switzerland. Further goals of this project are namely:
i) Encouraging and promoting multi-disciplinary work at the Swiss Federal Institute of
Technology, and among other research institutions;
ii) Establish and confirm the technical competence of the ETHZ in the field of risk and
safety of technological systems;
iii) Encouraging scientific and academic contacts to other institutes, industry, govern-
mental agencies, and political institutions within Switzerland and internationally;
iv) Supporting the teaching aims in the interdisciplinary fields of risk analysis at ETHZ.
For the most part, the PPR&S has now developed from a local to a national interest
in disasters prevention and emergency planing activities. In the framework of the
PPR&S it was decided to collect and review basic technical information and topics
concerning the "Quantitative Risk Assessment of Hazardous Materials Transport Sys-
tems" and to publish the results in book form, as being part of a books' series presented
under the auspices of PPR&S. This series forms a whole and covers different aspects of
risk assessment, management, risk acceptance, as well as legal and societal aspects
thereof.
Assessing the risks of a region implies the use of a complex methodology dealing
with risks to health and to the environment, normal operation and accidental situations,
a large variety of industries, transportation of hazardous materials, impacts, regulations
and actors involved in the decision making process.
As opposed to other existing approaches (e.g., UN- interagency project on risk as-
sessment of large industrial complexes), the PPR&S project took the initiative to design
procedural guidelines for IARASM implementation by means of identifying tasks and
integrating them into a comprehensive and systematic approach. By contrast, other
existing guidelines take a problem solving oriented approach which is too global and
does not always assist systematically the analyst or the project manager.
This approach (i.e., task oriented approach) allows a systematic analysis of the
problem of regional risk assessment, offers flexibility and efficiency in the implement-
ation process, allows initiatives and ad hoc modeling and simulation. Integration of risk
can not be done through a single risk indicator. Integrated regional risk assessment
should be considered as a process whereby decision aiding techniques (ranging from
simple brainstorming, the Delphi method to Multicriteria Decision Analysis, and Deci-
sion and Knowledge Support Systems) should play an important role. Various tech-
niques have to complement expert judgment, public participation and risk communica-
tion. It is a need to balance hard approaches (models, calculations) versus soft ap-
proaches (acceptability) in regional risk analysis.
In the process of risk integration an important role should be played by the task of
comparative risk assessment. PPR&S made successful experiments in working with
such tools and approaches. The need for a comprehensive regional risk assessment
XVI PREFACE
methodology has been highlighted in different occasions during the time of the
Polyproject
Some of the lessons we have learned in the course of the PPR&S, when dealing with
above issues, are:
• When getting involved in a regional risk assessment do not take a simplistic
approach.
• On regional risk assessment, avoid excluding political or human interactions at
all stages. Develop a risk triplex, namely: "safety culture, environmental aware-
ness, and emergency culture".
• When running a research or a case study on regional risk assessment do not en-
tirely rely on a self-organizing effect within the project, or case study. A strong
interactive project management framework is needed from the beginning.
• Due to technical accidents or natural disasters, or their synergism, there is an
emerging need for national and international organized research and activities in
the above field Risk from normal operation or from accidents may have some
transboundary effects.
• Legal issues at the local level or national level have to be solved and harmonized
before any implementation of a comprehensive regional risk assessment meth-
odology can take place.
• There is a need for specialized databases; their systematic use will reduce the
uncertainty of the results.
• Recent advancements in information and telecommunication technologies (GIS,
ISDN), multimedia, virtual reality, neural networks, could play an important
contribution to the modeling of various risks.
• In regional risk assessment, all risks (local, regional, global) should be taken into
consideration.
• Safety culture, public participation and risk communication are relevant issues in
the overall landscape of the regional risk assessment process. Emergency culture,
preparedness, and planing is an integral part of regional safety management
As compared with similar projects in the world (e.g., the UN Inter-Agency on Risk
Assessment and Safety Management for Large Industrial Complexes), the present work
brought new answers to this interdisciplinary subject. WoIk done within the PPR&S is
complementary to the numerous activities developed recently in Switzerland.
Further information on the Polyproject and its publication series can be obtained
from:
Polyproject "Risk and Safety of Technical Systems"
ETH-Center Phone: +(41) 1 6322356
CH-8092 ZUrich Fax: +(41) 1 632 1094
Switzerland
FOREWORD
In recent years, the community has become increasingly aware of the risks of locating
hazardous industries near heavily populated environmentally sensitive areas. This new
awareness could lead to a novel approach to safety planing for hazardous industries,
looking at the problem from the point of view of integrated regional risk assessment,
which should include beside the risks arising from natural events (like earthquake,
flood, forest fires, etc.) also the risks arising from processing plants, storage and trans-
portation of dangerous goods.
The purpose of this volume is to highlight the main procedures for assessing the re-
gional risks resulting from dangerous goods storage, and transportation by means of
different technical systems (i.e., road, rail, ship, and pipeline).
The information contained in this book is based on a wide range of references and
studies. The main procedural steps involved in quantitative risk analysis for transporta-
tion systems are supported by relevant methods of risk assessment recognized on an
international level. The present book gives an overview of the criteria and guidelines
applicable to the implementation of risk assessment and management at different stages.
Chapter 1 describes the environmental and safety factors to consider, when perform-
ing a transportation risk analysis for a region. Chapters 2 presents risk definitions and
the methodology for analyzing transportation risks in a complex area. Chapter 3 pres-
ents general information about truck accidents and thei!" consequences, and reviews the
risk presented by road tunnels. The BUW AL methodology for assessing the risks of
hazardous materials transportation by road is explained in detail. Finally new develop-
ments in traffic and vehicle control are discussed. Chapter 4 deals with transportation
of hazardous materials by rail. The U.S and British experience is reviewed. information
concerning the Swiss Federal Rail Network is presented, then the BUWAL methodol-
ogy for assessing the risk of rail transportation is described and the findings of a risk as-
sessment study by the SBB are finally summarized. Chapter 5 is more concerned with
the assessment of transportation risks on water ways. Chapter 6 furnishes a descrip-
tion of the transport pipelines for natural gas and petroleum products and describes the
situation of Switzerland. Some information on storage and handling as well as informa-
tion concerning codes of practice and international regulations can be found at this
place. Chapter 7 presents a compilation of statistical data related to accidents and dan-
gerous goods' movements. This information is related more specifically to the case of
Switzerland Chapter 8 is devoted to the description of data bases and computer sup-
port for risk assessment. The content of several data bases and programs is presented
and the source of availability mentioned. Special literature and topics concerning GIS
applications, and other risk assessment methods are reported at this place. Chapter 9
deals with integrated approaches for regional risk assessment and safety management
with special emphasis to the transportation of hazardous materials. Chapter 10 presents
several relevant case studies and miscellaneous information.
XVIII FOREWORD
In Chapters 3 and 4, dealing with rail and road transportation, we have summarized
the method for assessing transportation risks, which is recommended by the Swiss Fed-
eral Office for Environment, Forestry and Landscape, BUWAL.
Due to the time and manpower imparted to the preparation of this book within the
framework of PPR&S, it was not possible to assess in detail the risks presented by the
rail and road transportation of hazardous materials in Switzerland Much to our regret it
must be stressed that the Swiss statistics concerning road tankers' and trucks' accidents
are not suitable for probabilistic risk assessment Due to this situation we have re-
nounced to assess globally the transportation risk for Switzerland on the basis of a prob-
abilistic approach. However the methodological approach to consider has been de-
scribed with much details. The task of the authors was to review the state of the art
concerning the analysis and assessment of transportation risks for dangerous goods,
with a special issue concerning Switzerland. In that respect we hope that our contribu-
tion will be of some value to the reader interested by such topics.
Environmental impacts from accidental releases ofpoUutants into air, water, and soil
were mainly treated in Volume II of Integrated Regional Risk Assessment, by the same
authors, published recently by Kluwer Academic Publishers, Dordrecht, The Nether-
lands.
It should be noted that the risks of nuclear fuel and radioactive wastes transportation
have not been treated here at length. This is because nuclear risk assessments are cur-
rently carried out at a higher level than that used for other facilities (e.g., in Switzerland
by the HSK, Villigen) and the information would be available for use in integrated risk
assessment at community level.
This book shall be valuable to students, engineers, and scientists in charge of dew'l-
oping new methodologies for !.Iazard analysis and risk assessment for transportation
systems; practitioners active in the field of environmental protection; local or govem-
mental Authorities in charge of implementing environmental risk impact assessment
procedures and guidelines.
ACKNOWLEDGEMENTS
The authors wish to express their sincere gratitude to Professor Wolfgang Kroger,
Chairman of the Executive Committee of the "Polyproject, Risk and Safety of Technical
Systems (PPR&S)", ETHZ - Swiss Federal Institute of Technology, Ziirich, for his
guidance and critical reviews during the different phases of this work.
We are also greatly indebted to all people, who through their support and many
valuable suggestions for corrections and improvements of the manuscripts, helped us to
finalize this work. We wish also to acknowledge more specifically the following indi-
viduals:
Prof. R. Hatter, Vice-President Research, ETH ZUrich
S. Chakraborty, HSK, Villigen, Switzerland
Dr. Hans-Jorg Seiler, Project Manager for the PPR&S
Prof. J. Schneider, Institut fUr BaustatiklKonstruktion, ETH, ZUrich
Mr. H.A. Men, EBP, Basler Ernst & Partner Ingenieuruntemehmen, ZUrich
Prof. Dr. B. Bohlen,former DirectorofBUWAL, Bern
H.R. Wasmer, Deputy Director, EA WAG, Dabendorf
Prof. K. Hungerbuehler, Institut fUr Technische Chemie, ETHZ
Dr. H.P. Alder, Staff Nuclear Energy and Safety Research Department, PSI, Villigen
Dr. R. Kieselbach, EMPA, Dabendorf
Dr. H. Kiinzi, Konzernsicherheit und Umweltschutz, Hoffmann-La Roche AG, Basel
Mr. K. Cassidy, Head Major Hazard Assessment Unit, Health and Safety Executive,
London.
One of us (A.G.) would like to express special consideration and high appreciation
to Mrs. Fran~ise Bordier for her exquisite support and distinguished encouragement in
his professional activity while in Switzerland. Finally, in the preparation of this book,
the authors are greatly indebted to Mrs. I. Kusar (PSI), who skillfully prepared the
drawings and pictures for illustrating the manuscript.
LIST OF ABBREVIATIONS
API : American Petroleum Institute

BUWAL : Bundesamt fUr Umwelt, Wald und Landschaft, Bern (Swiss Federal
Office for Environment, Forestry and Landscape)
CEAM : Center for Exposure Assessment Modeling
CEC : Commission of the European Communities
CRWM : Civilian Radioactive Waste Management
DOE : u.S. Department of Energy
DOT : U.S. Department of Transportation
EAWAG : Eidgenossische Anstalt rur Wasserversorgung, Abwasserreinigung und
Gew3sserschutz, Dubendorf, Schweiz (Swiss Federal Laboratories for
Water Supply, Waste Water Treatment, and Water Pollution Prevention)
EC : European Communities
EGES : Expert Group for Energy Scenarios, Bern (Switzerland)
EMPA Eidgenossische Materialprtifungs- und Forschungsanstalt (Swiss Federal
Laboratories for Materials Testing and Research)
EPA : Environmental Protection Agency (USA)
ETHZ : Eidgenossische Technische Hochschule, Zurich (Swiss Federal Institute
of Technology)
FEMA : Federal Emergency Management Administration (USA)
GIS : Geographical Information System
HEC : Hydrologic Engineering Center, California (USA)
HMIX : Hazardous Materials Information Exchange (USA)
HSC : Health and Safety Commission (UK)
HSE : Health and Safety Executive (UK)
HSK : Hauptabteilung fUr die Sicherheit der Kernanlagen, Villigen, Schweiz
(Swiss Federal Nuclear Safety Inspectorate)
IARC : International Agency for Research on Cancer
IGWMC : International Groundwater Modeling Center (USA)
NRC : National Research Council
OECD : Organization for Economic Co-operation and Development
UNEP : United Nations Environment Program
u.S. EPA: Environmental Protection Agency (USA)
WHO : World Health Organization
CHAPTER 1
ANALYSIS AND ASSESSMENT OF TRANSPORTATION RISK

• ENVIRONMENTAL AND SAFETY FACTORS
1.1. A Methodological Overview
It was not the purpose of the present book to analyze in depth the risk of transportation
of dangerous goods in Switzerland, but instead, in the framework of the Poiyproject
Risk and Safety of Technical Systems (PPR&S), we concentrated mainly on applied
risk analysis techniques and special attention was also given to highlighting the
methods and to complement them with examples and cases studies, to present general
infonnation and, as far as available, international as well as Swiss national statistical
data on accident frequencies, accident speed, spill frequencies, and other matter related
to the risk analysis process of hazardous materials transportation systems, involving
road, rail, ship and pipelines, to be viewed in the context of integrated regional risk
analysis.
In the first two chapters we present important definitions concerning risk, and
explain in a general way the methodology to apply for assessing the risk of hazardous
materials transportation. The Chapters 3 through 6 are treating more in detail specific
topics of road, rail, ship and pipeline transport systems. The reader will find various
st;ltistics and infonnation on accident, releases of hazardous materials, and conse-
quences of fire and explosion. Chapter 7 is more concerned with infonnation and
various statistics on dangerous goods movements in Switzerland. Chapter 8 describes
data bases of interest, and in Chapter 9 we also present some concepts of integrated risk
assessment and safety management at regional level related to hazardous materials
transportation. Finally in Chapter 10 we present some case studies showing the
methodology followed by different companies or authorities in Switzerland and in other
countries, for assessing transportation risks of hazardous materials. These case studies
present further methodological details and thus complement the topics covered in the
other chapters.
In this book we also describe the methodology recommended by the Swiss Federal
Office for Environment, Forestry and Landscape (BUWAL)I for assessing the risk of
dangerous goods transportation by rail and road in Switzerland. The principles of the
Swiss methodology can easily be applied to countries other than Switzerland, under the
condition that the coefficients used in the equations be replaced with coefficients
reflecting the conditions prevailing in the country under consideration, because the
accident statistic and traffic conditions might differ considerably.
The transport and distribution of hazardous materials, such as petroleum products,
liquefied petroleum gases, chlorine gas, pesticides, chemicals/petrochemicals and radio-
active materials, inevitably involve the potential for incidents and accidents which may
2 CHAPTER 1
result in death or injury to people, property damage or damage to the biophysical

environment through the effects of fire, explosion or toxicity. An increasing number of
transportation accidents involving hazardous materials have occurred worldwide. Such
accidents with their resultant effects on people and the environment have increased
awareness in government, industry and the community at large and resulted in a re-
think in the risk assessment process for hazardous materials transportation. In that
context, it is now recognized that the safety planning of transportation routes,
accounting for the type and nature of surrounding land uses, is an integral component of
the safety management of hazardous materials transportation. Delineating hazardous
materials transportation routes is, as such, a significant and essential complementary
measure to technical and operational safety and environmental controls on the
hazar40us materials containers and associated regulatory processes. It is relevant to note
that fixed installations are more amenable to local, organizational and operational
hazard controls. Transportation systems are dynamic systems with additional external
variables (e.g., drivers, traffic conditions, etc.) difficult to bring into one overall control
system.
This chapter will focus more particularly on the analysis and assessment of trans-
portation routes (i.e., roads) for the carriage of hazardous materials. The integrated risk
assessment approach to the safety of hazardous substance transportation necessitates
consideration of three main elements in an integrated manner:
a) Transportation risk and environmental land use safety factors, including the
identification and quantification of risks to people, property and the environment
from the transport of hazardous material, particularly as they relate to effects on
land uses and environmental 'ecosystems along the transportation routes. These
are environmental and risk factors.
b) Capability of the existing road netwurk and cumulative traffic implications,
including overall traffic movement, congestion and level of service on used or
potential routes, accident rates, route conditions. These are traffic related factors.
c) Economic distribution considerations and operator's requirements for practical
transportation economics, including considerations of travel distance and time,
and the transportation costs of alternative route systems.
An integrated assessment of the safety adequacy of an existing hazardous materials
transportation route or the formulation of alternative routes for the safety management
of such transportation necessitate the quantification and weighing of all three elements
indicated above. Although a brief description of elements (b) and (c) will be provided,
the focus of the chapter is on the risk and environmental considerations of hazardous
material transportation in line with the main focus of this guideline document.
The assessment of the safety suitability of an existing transport network for the
transport of hazardous material and the formulation of routes for the safe transport of
such material are therefore major objectives.
There are three main applications for the information, tools and techniques outlined
in this chapter (see box):
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FACTORS 3
Analysis and Assessment of Transportation Routes

• Identification, analysis and assessment of the environmental and safety land use
implications (as well as traffic and economic implications) of existing routes and trans-
portation of hazardous materials on a area scale. The output being a quantification of exis-
ting risk from the transportation of hazardous material and assessment of the adequacy and
appropriateness of existing routes for the transportation of such material.
• The formulation and designation of hazardous material transportation routes as an
integral component of the environmental and safety management of such transportation,
including the exclusion of routes with the highest risk to people and the environment
(Section 1.4 deals with the management aspects of dangerous goods transportation).
• Providing the basis for the assessment of both the individual and cumulative environ-
mental and safety implications of a development proposal which generates or receives
hazardous material.
It should be noted that although this chapter deals mainly with the risks of
hazardous materials being transported, when considering the overall system, the risks of
choosing a particular mode of transport may well be dominated by the risks involved at
the loading/unloading facility. These risks can be calculated either as part of the fixed
facility assessment or as part of the transport assessment - whichever is the more
appropriate for the specific situation. It should also be noted that, for most forms of
transport of hazardous materials, there exists a comprehensive and extensive list of
international transport regUlations.
The results of a transport risk assessment are very specific to that particular assessment
and care must be taken to avoid drawing conclusions from a few studies and then
generalizing from these, i.e., rail may not always be safer than road, pipelines muy nat
always be safer than raj{ etc.
This section describes the procedures for analyzing and comparing alternative routes
for the transportation of hazardous materials on the basis of land use and environmental
safety. It is not intended to provide in-depth documentation of the assumptions and
processes implicit in the methodology. Rather, the purpose is to highlight the most
relevant procedural information and a concise description of the criteria that may be
applied for hazardous materials routing.
Factors that influence routing decisions, from an environmental safety viewpoint,
may be grouped into three inter-related categories (see box):
• Mandatory factors such as legal and physical constraints (e.g., topography of the area).
• Environmental and land use risk (including various hazards and their associated risks).
• Subjective factors that reflect community priorities and values which may not be easily
quantified (e.g., population distribution, special land uses, routing, and emergency
response).
4 CHAPTER 1
Consideration of each of the above factor may on its own or in combination

preclude the use of any particular route for the transportation of hazardous material or
favor an alternative route.
1.1.1. GATHERING INFORMATION ON THE REGION OF INTEREST
Environmental Resource Mapping. It is of good practice to identify and map the

environmental resources of the region to be studied (fish habitats, bird habitats, and
relevant fauna). Major habitats should be rated according to their sensitivity to
dangerous goods spills.
Demographic and Geographic Data Base. A demographic data base should be

compiled for the region and is to be used for risk assessment studies. The data base
may, for instance, contain information on population density and property values, which
are based on national statistics and additional information from the communities
belonging to the region. A thematic data map may be overlaid on a series of digital
cartographic bases. The display should be organized in layers (themes) in such a way as
to facilitate comparison of data sets by overlay techniques. Geographical information
systems like GISIREGIS and ARCINFO have special features for corridor analysis (see
Chapter 8.8 concerning applications dedicated to transportation).
Railroad Networks. Railroad transportation of dangerous goods has to be assessed as

well as road transportation in case. of an evaluation of the global risks for a given
region. A general description of the rail networks, routings and transshipment facilities
is ne<>.ded. Data on freight volume of dangerous goods must be collected (on a
confidential basis) and shall include actual quantities of commodities moved over
particular route segments and the number of carload movements as well as the type of
railway equipment used. Tank cars are normally divided into pressure and non-pressure
categories. Pressurized tank cars are capable of transporting pressurized, refrigerated
liquids. Non-pressure or general service tank cars are used for transporting all kind of
goods from corrosive materials, to oxidizing substances, to poisonous and flammable
liquids.
Road Transport. The transportation of dangerous goods may be done by several

different companies. Origins and destination for the products moved by these firms
range from local manufacturers and consumers to other inter-state (French regional
"Departments", Swiss "Cantons", German "Buodesliinder", etc.), national and inter-
national transit traffic.
Two methods may be used to identify the types and quantities of "high hazard"
dangerous goods transported within the region.
a) Using a questionnaire: The ftrst method is to survey transporters of truck load
quantities of dangerous goods (gasoline, LPG, chlorine, ammonia, and other
chemicals, or explosives, etc.) these companies transport, by submitting a question-
naire to them. This questionnaire should provide indications as to the types and
quantities of dangerous goods these companies transport and the route they use
when moving these commodities.
b) Using statistics: The second method is to use statistics based on a trucking

survey conducted, if possible, annually to measure commodity origin and
destination (like for instance the survey of the "Swiss Alps transit through the
Gotthard tunnel").
Case histories and statistical data on reported rail and road transport accidents are
needed for regional risk assessment.
Road Accident Rate. The probability of a dangerous goods accident can be estimated
in two steps. First, the likelihood of an accident for any accident must be extracted from
historical data records for the road network, using national statistics or when missing
using the reported statistics of countries, where similar traffic conditions are prevailing.
The accidents must be matched with the volume of vehicle traffic, expressed as the
Average Annual Daily Traffic (AADT). For highways and other national roadways, a
representative accident rate can be derived from their physical characteristics, such as
number of intersections and average traffic volume. Accident rates can be computed for
a given road length (called a link) according to the formula:
A CCI·dent rate Accidents per year

=----....:....--''--- (1.1)
Vehicle - km per year
. Accidents
ACCIdent rate =-----
Vehicle- km
(1.2)
Generally due to a lack of good statistics (based on accident categories), accidents

could be assumed to include any incident that results in injury, death or property
damage exceeding a certain amount in $, or other currency.
Pedestrian accidents should be excluded. The vehicles to be assumed in the accident
rate derivation should include all passenger, transit and commercial vehicles. A typical
value for the road accident rate is 3.0 x 1(}-6 accidentslvehicle-km. The inference of this
approach is that dangerous goods truck accidents can be predicted, at least to some
extent, from the historic accident rate for a given route. This implies the combination of
physical features (such as intersections, road width, and traffic speed) to delermine the
exposure of a single vehicle to an accident.
While this approach has been widely adopted by other communities, and is useful in
comparing one route with another, it carries inherent problems.
It has been recognized, that among the factors that contribute to dangerous goods vehicle
accident, humon error ranks highest. The most common causes of vehicle accident are due
to excessive speed, following to close the preceding vehicle, non observance of rest-time
leading to driver overfatigue, and failure to observe traffic warnings.
6 CHAPI'ER 1
Rail Accident Rates. The selected method relies on available dangerous goods accident
and traffic infonnation to calculate an accident rate for each of the given rail links.
In Switzerland, statistics on rail accidents are prepared on an internal basis by the
Swiss Federal Railroad (SBB/CFF). Historic accidents for cars carrying dangerous
goods can be identified for a given period and divided by the total number of dangerous
goods (DG) car movements over the same period to yield:
DG accidents per year Accidents
----------~~~---=----------
DG car movements per year Car movements
li!!K: 00. here. means dangerous goods.
The inference is that the" greater the accident rate for a given link, the greater the
likelihood of a dangerous goods rail car being involved in a crossing accident, collision
or derailment A typical value for the rail accident rate is 6.0 x ](}6 accidentlcar-km.
This approach assumes that the historic accident rate, recorded by link, predicts the
future accidents.
Emergency Response
Regional and national contingency plans for emergency preparedness and response
must be taken into consideration, when preparing the risk assessment study.
The Equivalent Safety Concept and the Population Vulnerability Model

Two tools to determine the hazard presented by a large quantity of hazardous materials are
the Equivalent Safety Concept and the Population Vulnerability Modef
• The Equivalent Safety Concept is a non computer technique that develops indexes for
cargo hazards. for vessel design. and port safety. These are used to assist in a
judgmental de~·lSion of authorization of vessel transit.
• The Population Vulnerability Model is a computer simulation of a cargo spill integrated
with census data. The damage to life and propeny are calculated using the census data
and the cargo properties to determine the number of deaths and injuries to personnel and
dollar losses from the cargo release. The results and relative hazards are calculated
using these techniques.
1.1.2. SELECTING DANGEROUS GOODS FOR REGIONAL RISK ASSESSMENT
For the assessment of transportation risk, the objective is to first develop a

consensus on the process and criteria to be used to define the high hazard goods related
to a geographical region, and then to apply the process and criteria to the dangerous
goods listed in the regulations referenced (RID3IRSD4 ; ADR sISDR6 ).
Risk Assessment
For assessing road and rail risks, the study should include the following areas:
a) Goods in Movement. For transport en route, the population at risk is twofold: (i)
people living near the road or railway, and (ii) the other users - passengers on
passenger trains, or users of other road vehicles which might become involved in
a major accident
b) Temporary Stopover Points. These include railway marshaling yards, hucke-
pack loadinglunloading yards, lorry parking areas at motorway service points,
custom yards, etc.
c) Goods Handling. The loading/unloading process itself for trucks and road-
tankers.
Substances to Investigate
The study should in a first phase be limited to those substances which are transported in
large quantities and which have the potential to cause major accident risks from fue,
explosion and toxic release. Such hazardous substances belong to the following cate-
gories:
a) explosives
b) flammable gases
c) toxic gases
d) highly flammable liquids (with flash point < 21°C)
e) flammable liquids (with flash point between 21 °C and 55°C)
1) flammable solids (including substances liable to spontaneous combustion)
g) substances which become dangerous when wet
h) oxidizing substances including liquid oxygen
i) toxic liquids.
Thus among the substances listed above, priority should, in a first step, be given to
three groups of substances:
- those which appear to be potentially the most hazardous (for instance,
specifically liquefied toxic and flammable gases, with liquefied petroleum gas
(LPG), chlorine and ammonia)
- those moved in the greatest quantities by road or rail, according to goods
movement statistics (e.g., flammable liquids lique motor spirit, domestic fuel oil)
with a potential to take fIre or to contaminate surface and groundwaters.
- various explosives and ammunitions
Other substances can and do present risks (e.g., oleum, or fuming sulfuric acid).
However molten solids. acids. liquefied inert gases and caustic substances will usually
either be transported in small quantities, or have a limited range of action for putting
only few people at risk at anyone time.
Loading and unloading can be a source of potential risk. Loading is usually done at
sites which come within suitable regulatory regimes. It is therefore the unloading
(mainly at sub-notifiable sites) by road tankers of motor spirit. LPG and ammonia,
which should be studied in detail.
8 CHAPTERl
1.1.3. THE NATURE OF THE MAJOR RISK TO BE ASSESSED
Harbors, unloading facilities, stopover points and marshaling yards present risks similar
in nature to those of fixed installations: - a (temporary) concentration of dangerous
substances at the vicinity of a local population surrounding it. In this case both
individual and societal risks should be estimated.
For road and rail en-route risks, it is the societal risk that becomes dominant because
the population at risk is constantly changing during the course of the day and the
individual may be only exposed to the hazard for a short time. However the total
number of individuals at anyone place potentially at risk from a release of a quantity of
hazardous substance may be larger than at a factory where such substance is produced
or used.
In the case of rail and road transport, the calculation of the societal risk is complex
because a route may expose a population in many towns and elsewhere, hence the risk
is spread over a wide area. There is also the group in transit of other road users, or
railway passengers, who might become involved in a major incident. A new concept of
"route societal risk" must be applied to consider the aggregated societal risk presented
by the transport of substances from one place to another.
An estimate of national risk from the carriage of a hazardous substance by rail or
road can be calculated from the route societal risk. scaling up by the volume of trade in
the given substance and length of route.
The accuracy of this scaling depends on the extent to which the routes chosen are
characteristic enough and truly representative of the conditions prevailing for the rest of
the country.
The risks from road transport of chlorine and ammonia are much lower than for
gasoline or LPG, but because ot the long-range effects if an accident of this kind shOUld
happen, there would be a very .. mall risk of an incident causing a very large number <1f
fatalities.
What concerns lorry stopover points, usually short-term parking is taking place in a
"safe place" with the vehicle usually in sight and within easy reach of the driver or a
competent person. The risk from chlorine and ammonia at a distance of 50 meters is
estimated to belO·6 (one in a million) per year of receiving a dangerous dose. This risk
at 50 meters would be consistent with the criteria for housing developments in the
vicinity of notifiable installations, though schools and nursing homes would need to be
built further away. For LPG the risk is somewhat higher. A distance of at least 120
meters would be recommended for certain kinds of housing developments and 170
meters for more vulnerable populations.
What concerns the loading/unloading of vehicles, there are a large number of
deliveries: for motor gasoline delivery is to numerous filling stations - usually urban,
but also on motorways and highways; For LPG, there is an expanding market and the
deliveries are made to filling stations respectively to a variety of sites, ranging from the
domestic to the industrial. Spills on delivery of motor gasoline is not infrequent, but
cases of ignition are seldom. Vent fire (i.e., ignition of the small quantity of vapor
necessarily escaping from vent) are more likely. The situation is different for LPG.
There the risks arise from errors in the filling process, or possible pull-away leaks
(driver pulling away without disconnecting the hose). Ignition thereafter can and does
occur. It can result in a torch flame followed by a BLEVE. International experience
confinns the possibility of serious consequences. Less than half of the risk comes from
the unloading part of the transport process, the remainder comes from the installation
itself.
For ammonia the risk are somewhat larger, though the transport risk are only a small
proportion of the total risk from the site.
Hazard Classification. There are different classification scheme which may relate to
the substance itself or to qualitative differences (for toxic material) or to degrees of
hazard (for explosives). An alternative approach is to base the classification on the
hazard itself. Another important classification relates to the compatibility of chemicals
for adjacent bulk loadings. The hazard associated with compatibility are principally
evolution of gas and/or heat and evolution of hazardous materials. Such substances can
be flammable or corrosive.
Hazard Identification. The process of assigning a chemical to a hazard category

involves the identification and evaluation of the hazard. It also helps to identify suitable
precautions. There are a number of systems of classification of materials devised for or
relevant to transport.
TABLE 1.1. List of hazardous products (According to RID)
Class Product Hazardous Product

Type
2 Gases Chlorine, Hydrous bromide, hydrogen chloride, hydro-
gen fluoride, hydrogen sulfide, phosgene, sulfur
dioxide.
3 Flammable Carbon disulfide
liquids
6 Poisons Acetone cyanhydrin, acetonitrile, acrylonitrile, allyl
alcohol, allyl chloride, aniline, epichlorhydrin, lead
alkyls, organo-phosphorous compounds
8 Corrosives Bromine, fluoboric acid, hydrazine, liquid acid halides
and chlorides which gives off acid fumes in contact
with moist air (SbCI,)
The RID list (TABLE 1.1) indicates some of the chemicals which are regarded as
particularly hazardous in West Germany.
For certain chemicals which are transported by road special measures are required.
These include:
• Substances liable to polymerization (ethylene, butadiene)
• Substances carried fused
• Hydrogen peroxide
10 CHAPTER I
• Organic peroxides
• Sulfur trioxide
• Bromine
• Lead a1kyls.
In order to identify hazardous cargoes make use the UN list of hazardous materials
as basic check list. Special attention should be given to chlorine, ammonia, LPG
(liquefied propanelbutane gas), and other liquefied flammable gases; toxic gases;
flammable liquids and gases.
Method for Selecting Dangerous Goods in the Study Area

Step 1: Identify 9-10 classes of dangerous goods based on the haZJlrd and risk they
represent (e.g., using the Dow Criteria List, etc.).
Step 2: Eliminate from the list of over x-thousands goods, those dangerous goods which
are not transported into, from or through the given region of interest.
Step 3: Use the toxicity indices' list (TERA or equivalent list}/a) 7 those indices are
partially based on human, animal and aquatic toxicity to find out the more toxic goods, and
the criteria set in the Swiss BUWAL Handbook ("Handbuch I zur StOrjallverordnung StFV"
for further reducing the number of goods to obtain ''priority dangerous goods". The
reduced list should now include those products which are potentially dangerous to the
environment as well as the Public safety.
Step 4: Reduction of the priority list to a smaller number of ''potentially haZJlrdous"
dangerous goods, by eliminating those dangerous goods transported in small volumes when
compared to others in the same class. The volume criterion should be used because of the
high correlation between volume transported and probability of spill. Products can also be
eliminated from the list if their physical, chemical or biological characteristics or potential
impact on public health and the environment are similar to and/or less significant than a
product which remains on the list. Products must be retained on the list if the volume
criterion does not apply or if their is a significant level of public concern about a particular
product.
Step 5: Reduction of the ''potentially hazardaus" dangerous goods (long list) to result in a
"short list" to enable a greater degree offocusingfor inventory, hazard and risk assessment
purposes.
Figure 1.1. Method for seJecting dangerous goods in the study area
Identify the main modes of routes of transportation, and if possible identify also the
main origins and destination. Road, rail barge, ship, pipelines, as well as main transfer
facilities should be considered in the region of interest. Usually the origin and
destination of goods in transit from a foreign country to another one are known. The
check points are the main custom offices at the frontier. In Switzerland these are the
custom .offices of Basle, Chiasso, Geneva, etc. There the situation is even more
complicated for goods transported by ship on the Rhine river, because these are
(a)The Substance Hazard Index (SID) distinguishes chemical substances by their vapor pressures and toxicity
ratings (API, 1990). Use toxicity, flammability and reactivity to determine whether a chemical substance is
hazardous.
complicated for goods transported by ship on the Rhine river, because these are
redistributed at their arrival in Basle between rail and road-tanker/trucks. No official
statistics exists for dangerous goods transports within Switzerland, beside the statistics
based on schedule of dues and not on dangerous goods classes that are established by
custom officers.
CriteritJ list for Dangerous Goods and Products

The Swiss Ordinance on "Protection Against Incidents" ("Verortinun,g iiber den Schutz vor
Storjiillen", {Stor/allverortinung, StFV}, vom 27.Februar 1991) contains prescriptions and
a list on the threshold quantities of dangerous goods and toxicity criteria, as well as a
classification regarding fire and explosion properties, that must be taken into
consideration, when establishing a report for risk assessment. In principle the criteria
defined by the European Community form the basis for such lists.
1.1.4. MANDATORY ROUTING FACTORS
Physical Mandatory Factors: These may preclude a routing alternative and include
weight limitations on bridges, height restrictions on overpasses, inadequate shoulders
for breakdowns, extensive construction activities or inadequate parking and turning
spaces.
Laws and Regulations. They may apply to any routing alternative in prohibiting the
transport of hazardous materials through certain routes or structures (e.g., tunnels,
bridges). Local, state and national transport authorities should be consulted in all cases.
Such prohibited routes are obvious frrst cut alternatives to be eliminated.
In the USA, the DOT Research and Special Programs Administration RSPA guide
for routing analysis (RSPA8, 1988) allows routes to be excluded for physical or legal
reasons. Factors that are not easily quantified, such as the presence of schools or
hospitals that are not easily evacuated, or the presence of a reservoir, can be considered
subjectively. A recent notice of proposed rulernaking (FHW A, 1992) lists 13 elements
that should be considered in establishing highway routine requirements:
Population density
Type of highway
Type and quantities of hazardous materials
Emergency response capabilities
Result of consultation with affected persons
Exposure and other risk factors (such as distance to hospitals)
Terrain considerations
Continuity of routes
Alternate route
Effects on commerce
Delays in transportation
12 CHAPTER 1
Climatic conditions
Congestion.
Five criteria for routing analysis have been evaluated by Abkowitz et al. (1992).
these include: (a) minimizing shipping distance, (b) minimizing travel time, (c)
minimizing release-causing accident likelihood, (d) minimizing population exposure,
and (e) minimizing the product of accident likelihood and population. Of these, the first
two criteria minimize economic costs, and the latter three maximize safety. It was found
that the routes that minimize risk may be so sinuous that they can be economically
unfeasible, or at least impractical. Abkowitz et al. recommend that a routing analysis
consider combinations of factors and use different weighing factors to evaluate trade-
offs, such as travel time vs. minimum risk.
The risk, R j , for accident scenario i is a function of the scenario frequency and, Fj,
and the scenario consequence, Cj:
(1.3)
The product of the two terms is a relative risk indicator. Generally speaking, simple
risk indicators used for routing analysis can be considered as reduced forms of the more
complete quantitative risk equation, that will include other terms like probability of
leak, probability of given meteorological conditions, etc .. They do not provide as much
information regarding the safety of dangerous goods transport as do more complete
quantitative risk analyses. However, simple risk indicators can be useful in some
decisions involving two or more alternatives.
1.15. ENVIRONMENTAL AND LAND USE SAFETY FACTORS
The overall environmental and land use safety criteria for route selection is that the
route which has the lowest risk value to surrounding people, property and the natural
environment should be selected. In this context, risk is determined in terms of the
cumulative combination of the probability of accidents and the consequences of such
accidents. These two elements of risk are dependent on the extent of population
exposed and number of properties or extent of natural environment ecosystems and the
accident rates. In general, routes with the smallest adjacent population as well as
accident rates, will have the lowest risk values.
Designating Routes for the Tmnsportotion of Hazardous Materials

In designating routes for the transportation of hazardous materials, the risk values in
absolute terms are of limited practical use. It is the relative difference in the risk values that
should mainly be considered when differentiating between the different route alternatives.
If sufficient differences exist between the risks of alternative routes (e.g., 25% or more), it
may be possible to designate the preferred hazardous materials route on the basis of the
mandatory factors and the risk calculations.
Estimations of the Consequences of Hazardous Materials Transportation

The subject of "consequence assessment" has been treated with more details in another
book by the same authors. 9 The estimation of the consequences of accidents from the
transportation of hazardous materials necessitates the data mentioned in the box.
Data Needed for Estimating the Consequences of Hazardous Materials Tmnsportadon

• The nature of materials being transported; (this should also include hazardous waste).
• The storage/transportation conditions (e.g., temperature, pressure).
• The quantity of the load.
• The nature of the transportation tanker(s) including configuration of major characteris-
tics.
• Prevailing meteorological conditions applicable to the road network under consideration
(including wind speed, direction and where possible atmospheric stability).
• Topographical characteristics of the general area-both natural and man-made.
• Land use survey of the surrounding areas along the transportation routes, including the
type and nature of land use (residential, commercial, schools, hospitals, etc.) and the
residential/population density associated with each type of land use.
Accident scenarios usually includ:

leakage of the tanker's contents
pool fires
tanker flre
explosion or release of toxic substances into the environment.
The consequences of each accident ·scenarios are computed, usually in terms of heat
flux, explosion overpressure and toxic exposure using consequence modeling tools.
Based on such estimates and the population densities for land use adjoining each route,
the number of people affected by the postulated incidents, in terms of injuries or
fatalities can be determined. The effects on the environment can be determined in a
similar manner, using the models described in another book by the same authors9•
TABLE 1.2 outlines internationally suggested potential areas of impact based upon
the maximum recommended evacuation distance. The values indicated are conservative:
greater rather than smaller potential impact distances are used. The analyst may
therefore use either the locally derived impact distances (more accurate) or the values in
TABLE 1.2 as a general guide. Whatever distance is chosen must be consistently
applied in each alternative analysis for an objective evaluation.
It may also be insufficient to determine just the average risk over the whole route as it may
be necessary to determine the local risks adjacent to the more hazardous locations or the
more sensitive surroundings.
14 CHAYI'ER 1
TABLE 1.2. Potential impact area for different classes of hazardous materials
Class of Hazardous Material ImpadArea

Combustible Liquid 0.8 kIn all directions
Flammable Liquid 0.8 kIn all directions
Flammable Solids 0.8 kIn all directions
Oxidizers 0.8 kIn all directions
Non-FIammable Compressed Gas Downwind 2.1 kIn wide x 3.2 kIn long
Flammable Compressed Gas 0.8 kIn all directions
PoisonIToxic Downwind 0.3 kIn x 0.5 kIn long
Explosives 0.8 kIn all directions
Corrosive Downwind 0.8 kIn long x 1.1 kIn wide
Estimations of the Probability of Transportation Accidents

The probability of a hazardous materials accident is the likelihood or chance that a
vehicle carrying hazardous materials will be involved in an accident. To calculate this
probability, the analyst derives the accident rate applicable to the load and the route
segment and then must adjust this accident rate to reflect the amount of exposure or
vehicle experience (Figure 1.2).
~: Determine the accident rates on a particular route. Ideally, the most reliable data
concerning accident rates would be those associated specifically with hazardous materials
transpartation tankers in terms of number of hazardaus mIlterials accidents per tankers. If
such infortnlltion is available then it should be used directly into probability estimlltions.
In 1tIIl1Iy cases, however, such infortnlltion is not readily available. It is usually necessary
therefore to rely on accident rates statistics for all vehicles and then to adjust these to
reflect the smlliler share of hazardous mIlterials in the traffic stream. The first step is
usually to obtain statistics from historical records of the total rates of accidents from all
vehicles, usually in terms ofAccidentslvehicles-km
~: Calculate the probability of an accident for any vehicle based on vehicle exposure.
The probability of any vehicle being involved in an accident of a specific segment is
calculated by multiplying the segment accident rate (from step I above), with the road
segment length (or amount of exposur,:). This probability is in terms ofAccidentslvehicle.
Step 3: Factor the probability statement for any vehicle to reflect the incidence of
hazardous mIlterials vehicles in the traffic stream. This is done by multiplying the proba-
bility figure from step 2 by the hazardous mIlterials accident factor (being the ratio of
hazardous mIlterials transport accidents I all vehicles transport accidents). This probability
is in terms of Hazardous Materials Accidents I vehicles.
Figure 1.2. Suggested sequence of steps for estimating the probability of transportation accidents.
The above three steps (Figure 1.2) may be summarized as follows:

All vehicles accidents Hazardous material accidents

PHMA = *I.ength of link (km)* (1.4)
Vehicle-kIn All vehicles accidents
where
PHMA : Probability of hazardous material accident
Note: Vehicle-Ian above refers 10 the total number of kilometers traveled by all vehicles for which
accident statistics are available. The length of link is identical with length of road segment
In cases where the values for hazardous material accidents could be obtained from
available statistics. then the following steps could be directly applied:
• Obtain the accidents statistics applicable to hazardous material tankers and convert to
Hazardous Materials Accidents/vehicle-km (Le .• per total number of km traveled by all
hazardous material tankers to which statistics apply);
• Obtain probability of hazardous material accident:
- Hazardous materials accidents I.e th f link (km)

PHMA - * ng 0 (1.5)
Vehicle-kIn
where
PHMA : Probability of hazardous material accident
Note: It is necessary in some justifiable circumstances 10 further introduce a correction factor that
reflect physical characteristics of the particular route segment which may increase the probability of
an accident on that particular route. The above equation should incorporate an allowance for this
factvf (F).
Risk Computations
The potential consequences (population and/or property) and accident probabilities for
each route segment are multiplied together to calculate the segment risk. The cumula-
tive summation across all route segments produces the total risk for the route. The
accident probabilities derived by the method'presented above should further be conver-
ted to a likelihood or probability of impact.
The probability of a release and a hazardous event occurring is computed using tools
such as event and fault tree analysis to incorporate factors such as:
• whether the load will be dislodged as the result of an accident
• extent of such load loss ensuing spillage
• effectiveness of any containment/emergency procedures
• likelihood of the spill or release reaching environmentally sensitive areas or
having an effect on people. buildings. etc.
Event Probability Estimation. Careful definition of the events' sequence to be quanti-

fied is an important stage in an analysis. In the case where loss of containment is the
major concern, the breakdown of event probability against size of release is necessary to
couple probability and consequences. Similar releases are usually grouped together in
16 CHAPTER 1
order to achieve the coupling of events with the appropriate consequences in the overall
risk analysis
There are two basic approaches of estimating event probabilities: the first is direct
use of statistical data on failure of "whole systems or sub systems". This is sometimes
called the "historical approach". The second is to break down the event in its
contributory factors and causes. This process is pursued up to the point where data is
available, or can be obtained, for the contributory sub-events. The overall probability is
synthesized from the data for these various sub-events. The analytical approach will
inevitably be adopted where the historical data is inadequate or simply not available. It
may also be adopted where there is a cause for expecting a particular system or
component to have a sufficiently different failure probability from that indicated by
historical data.
Sometimes the analytical approach will not succeed in breaking the event down to
the point where data is available for all sub-events. In this case judgmental inputs may
be required and this has produced suspicion that the technique is not scientifically
rigorous. However, the philosophy here is consistent with the basic approach of
breaking down the judgment of how likely the event is into smaller areas of more
specific judgment which are within the realms of practical experience. The overall
outcome is then less sensitive to anyone particular judgmental input. An important
consideration in the analytical approach is how far to continue the breakdown of the
main event.
An advantage associated with the use of historical event data is that, where the
accumulated experience is relevant and statistically meaningful, the assessment will not
omit any of the significant routes leading to the event. The data already encompasses all
relevant contributory aspects including the reliability of equipment, human factors,
operational methods, quality of construction, inspection, maintenance, operation,
environment, etc. However, routes which may not be relevant to a specific case under
study will be included, resulting in an over-estimate - usually referred to as a
"conservative" estimate - of the chance of the event. The sample may also include older
equipment designed to lower standards.
Often the historical data is of such quality that, while undoubtedly relevant, it is not
considered adequate. In this case synthesis must be used, and the predicted event
probability should be tested against whatever experience exists, to judge whether the
different approached produce compatible predictions.
The chance of occurrence of an event is generally referred to by the term
"probability". This quantity may indeed be a probability and will therefore be dimen-
sionless. However, in many circumstances it will be expressed as a frequency of occur-
rence over a specified time interval, for example a year or a system lifetime. It may be
possible to relate such a frequency to finite number of occurrences over a sufficiently
long time, for example a large number of events over any relevant time scale they are
taken to be indicative of a degree of belief that an event will occur during a more
limited period. This distinction between "frequency" and "Bayesian" interpretations of
probability is generally considered academic in the context of risk analysis. However,
the distinction between dimensionless probability values and frequency values is
important where these numbers are to be used in combination. Because it is not always
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FACI'ORS 11
possible to be deftnite about which way a particular value will be expressed, some
authors use the tenn "likelihood" to express the chance of something happening, leaving
it open as 10 whether a "probability" or a "frequency" sho uld be used depending on the
circumstances.
Types of Data for Event Probability Estimation

There are several types of data which can be used in quantifying event probabilities:
- Accident information
- Event data
- Reliability data concerning components of a system
- Data on human error.
A variety of data bases have been set up to slOre and process this type of information
(see Chapter 8), a successful example being MHIDAS operated by SRDRISE (UK
Health and Safety Executive I Safety Reliability Directorate "Major Incident Data
Service"). High quality even! data is very e xpensive to collect and ofteo a long time
elapses before a statistically viable sample has accumulated.
Publications ale another source of information. A problem wi th inserting this type of
informatio n into a data base is that distortion can be introduced by the data bank coding
system when data has not been collected specifically for the data base in questioo. Data
collection scheme must be arranged with due consideration for the nature of the events
or failure concerned. A different approac h is necessary for collection of information on
relatively frequent occurrences compared with rate events, when an extremely large
amount of experience may be necessary for the data to be statistically meaningful.
It must be remembered that some types of human errors (e.g., inspection, calibra-
tion, maintenance, etc.) are implic itly included in the reliability data for mech:uucaV·
electrical components. Other types of human error are not included in compo.... nent
reliability data, although they would be included in accidents data. The human error rate
observed has been found 10 be dependen t on many factors, such as the level of stress,
familiarity with an operation, or procedure, complexity of the task. training, and
situation-specific features. Several attempts have been made 10 gather relevant human
error data for a variety of taSks under various conditions and levels of stress.
Event Tru
Event trees follow initial causes through to possible outcomes. 1bey can be used to deal
with independent and coincident events, but ale particularly powerful for portraying
event sequences. Event trees are usually applied to binary state systems but can be used
for multi-<lutcome states. Although event trees are often used in their own right and
provide easy CJltension 10 q uantificatio n they are also useful in identifying sub-systems
which require faul t troe analysis. II is difficull 10 represent interactions between event
states and a separate troe will be required for each initiating event - the relationship
between different trees must then be considered very carefully. Outcomes ale related to
the specific cause being analyzed and could arise from other causes, which would not
be shown.
18 CHAPTER 1
In recent years a number of computer codes have been developed for event tree
analysis. TABLE 1.3 lists some of the more readily available computer codes.
TABLE 1.3. Computer codes available for event tree analysis
Computer Code Availability Source
ETA II Science Applications (USA)!O

RISKMAN Pickard, Lowe & Garrick (USA)"
SUPER Westinghouse Risk Management (USA)
The Figures 1.3 through Figure 1.5 present below different exemples of event trees:
a) for chlorine truck accident
b) for petrol road tanker accident, and
c) for a LPG tanker accident.
I IoccidenI wit! Truck Car!yirG CIbk1e I

I ~Criod I I ~ ~ Criod I I ~ tCriod I
33kg 12 500 ~~ker
PrubabIIy 0.14 0.71 0.13 0.02
of load I I I I
o·r or
PrubabIIy thIt
CyIindIIn'!lrwn (0.1) (0.1) (0.1) (0.1)
0·r 3
r
ill DiIIodged O'j
~ T~1~ 9 ~O~ T~~~ T~ Om 9

PrubabIIy (0.1) (0.1) (0.1) (0.1)
mm mm
=or (T1) (Or) (Or

of.... Hole Hole Hole Hole Hole Hole Hole Hole Hole Hole Hole Hole
rmm rmm
~of
(Of) (Or) (Or (T1) (Or) (019) (Of> (Or> (019)
~- ~~~ ~~~ ~~J; ~~~

pr 0.013 0.052 0.133 0.113 0.411 0.928 0.152 0.872 1.313 0.079 0.413 0.224
SevertIy Index pr = 4.7
Figure 1.3. Event tree for a chlorine truck accident.

TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FAcroRS 19
Petrol Tanker Accident
Probability 0.5 0.5
No
(0.8) (0.2)
I I
Probability
0.4 0.1
Sizlt of Pool Fire
1 000 5000 10 000 >20 000
(0.25) (0.2) (0.2) (0.2) (0.15)

I I I I
Probability 0.9 0.025 0.02 0.02 0.02 0.015
Fatality
Radius(m) ~ b~~~~
pr o 1.3 1.8 2.8 4.0 5.3
Severity Index pr =14.9

Figure 1.4. Event tree for a petrol tanker accident.
20 CHAPTER 1
LPG Tanker Accident
I I I I
Probability 0.95 0.045 0.0035 0.0015
Small Release
Conto Small Initial Large Total Tank
<100 kg
Leak Leak Failure
Or No Damaae
Pipe Failure
/Hole
h 0000':"
(0.95) (0.05)
Probability O. 00f25 O. 175
ReiiefVaIve Total Tanker
I
(0.9)
I
Operates
-----' (0.1)
Release
L...--"""T""-----l
Probability .0029925 O. 3325 0.001 125
Ignition ~
Immediate
Ignition
Ignition
(0.5) (0.2)
Probability - 0.000 25
[= Explosion
(0.5) (0.5) (0.7) (0.3) (O.B)

Probability
Figure 1.5. Event tree for a LPG tanker accident (i.e., liquid propane gas).
Fault Trees
Fault trees are the best known and most widely used technique for developing failure
logic. The basic process adopted is to select an undesired "top-event" (e.g., leakage,
fire, or explosion), and trace it back to the possible causes which can be component
failures, human errors or any other pertinent events that can lead to the top event. This
procedure should be followed methodically, identifying the immediate precursors to the
top event, then the immediate precursors to these sub-events and so on.
A fault tree is comprised of a complex of entities known as "gates" which, when the
logical combination of the input conditions is satisfied, produce a specified output
which is propagated. The construction of a fault tree follows a defined methodology,
but there is, as yet, no universally adopted set of symbols, which is unfortunate as fault
trees provide a vivid method of communication.
The fault tree is not a model of all possible system failures, or all possible causes for
system failure, as it only includes events which contribute to its top event. A fault tree
does not show sequences of events, although some complex gates can be "used to
illustrate conditions on the sequence of input events. This can be confusing and often
leads to difficulties in evaluating the tree. A fault tree essentially shows system states.
When compiled rigorously in "top-down" procedure, certain sub-events may appear
more than once in the fault tree. This is not anomalous and does not prevent the tree
being quantified. Attempts to draw the tree without repetition of events in the tree
usually lead to the "top-down" approach not being rigorously followed, increasing the
chance of omissions. An important feature that the analyst must consider is the
possibility of failure of independent items by a single cause.
Over the past few years a number of computer codes have been developed for fault
tree analysis. TABLE 1.4 lists some of the more readily available computer codes.
TABLE 1.4. Computer codes available for fault tree analysis
Computer Code Availability Source
Rikke R. Taylor (Denmark:)

CAT G. Apostolakis et al.
Fault Propagation F.P. Lees (U.K.)
Digraph S. Lapp and G. Powers
IRRAS-PC EG&G (Idaho, USA)
CAFTA-PC Science Applications Inti. Corp. (USA)"
TREDRA J.B.F. Associates (Tennessee, USA)"
GRAFTER Westinghouse (USA)
BRAVO J.B.F. Associates (Tennessee, USA)"
Fault tree analysis provides an extremely powerful tool which is capable of

handling most forms of combinational events. It provides a good basis for quanti-
fication and is particularly useful where a small number of major outcomes are of
concern, as is usually the case in hazard analysis. Very large trees can result, with a
separate tree required for each top-event - relationships between different trees then
need to be considered carefully. Only the outcome under consideration is shown - other
outcomes from the causes in the tree will not be shown. Transition routed between
states are not represented and the technique generally deals with binary states: partial
failures and multiple failure modes can cause difficulties.
A method for the computation of transportation risks along different route segments,
based on the cumulative combination of the consequences and probabilities of accidents
is suggested in the plate. For each type of cargo there is a series of accident scenarios,
22 CHAPTER I
each of which can lead to fatalities within a different radius for each scenario. The road
can be divided into a number of segments, each of which will be related to a corre-
sponding population density. The following procedural steps are appropriate (Figure
1.6).
If one considers the movements of a road tanker carrying hazardous materials along a
route, for each sub-segment i of the route, there is a probability Pai for the tanker being
involved in a accident.
For each accident there is a number of possible accident scenarios Sj, each of which may
be considered to be fatal to individuals present within radius rj.
The number of people N present at the location of the accident, which may be affected,
depends on the density of the population D,
N=1CrfDj
Thus if one considers the passage of a tanker on route segment i, the probability of
someone being killed for scenario j is given by:
Pfataljty = Pai l'.j1C rf D j
The probability of someone being killed, due to the passage of the tanker on the sub-
segment i, is the sum of the probabilities for all possible accident scenarios:
Pfatality = L j P,.;l)Sj1C rf Dj
Fatality Probability = PaiDj L 1C rf l'.j
For anyone type of load, the term 1C rf l'.j is a constant which is independent of the route.
This term can be called the Severity Index, Sf, for the load in question.
Thus the probable number of fatalities from the passage of a tanker carrying load L along
sub-segment i is:
PaiDj(SI)L
and for the entire length of the route, it is equal to (SI)L LPaiDj . For anyone given
load, it is possible to compare the relative safety of two alternative routes by comparing
the term LP,.jDj , i.e., the population density, D;, along the route times the probability,
Pai, of an accident.
Figure 1.6. Methodology for risk comparison of alternative transportation routes.
Cause-Consequence Analysis
Cause-consequence analysis is a technique which combines the ability of fault trees to
show the way various factors may combine to cause a hazardous event with the ability
of event trees to show the possible outcomes. Sequences and therefore time delays can
be illustrated in the consequence part of the diagram. The technique has considerable
potential for illustrating the relationships from initiating events through to end
outcomes. It can be ·used fairly directly for quantification, but the diagrams can become
extremely cumbersome. Because of this, cause-consequence analysis is not as widely
used as the first two techniques for complex systems, possibly because fault and event
trees are easier to follow and so tend to be preferred for presentation of the separate
parts of the analysis.
Calculatio" of the populatio" risk for a give" IrtInsportalio" rou"

Step 1: For each substance (classified by load category) transported, establish the range
of hazardous events scenarios, the probability of each event and the radius of fatality (or
injury) effect from each event. Figures 1.3 through 1.5 are examples of event trees that may
be followed. The depth, extend and number of hazardous scenarios will depend on the
completeness of the analysis study. It may be possible for simplified case analysis to
postulate two or three hazardous accident scenarios and assign probabilities and estimate
the radius offatality (or injury) effects from each event.
~: From the above, estimate the severity index for each category of hazardous load:
• Calculate p 7t rj for each hazardous scenario, where p represents the probability of a
given scenario occurring with a tanker.
• Calculate the severity index (SI) by summing-up 1: p 7t rj for all postulated scenarios.
~: Multiply the severity indexfor each load category,1:p 7t rj' by the probabilityofa
hazardous material accident as determined in the previous section of this chapter.
Step 4: Multiply the result of step 3 by the population density along each of the
transportation route under consideration. This is the population risk for the route(s) under
consideration for the hazardous load.
Note: The population density (number of peopleAarr) may be obtained by calcula-
tions or from population statistics for different categories of land uses.
~: Compare the population risksfor the different route alternatives.
Figure 1.7. Procedure for calculating the population risk of transportation routes.
1.1.6. SUBJECTIVE FACTORS
Subjective routing factors in the selection (or elimination) of routes for the transport of
hazardous materials usually include:
• The location along the roadway (or in its vicinity) of sensitive land uses such as
hospitals, schools, old age person housing, churches or items of heritage or cultural
significance; or the location of sensitive ecological systems and natural landscape such
as park reservations, wetlands.
• Emergency and evacuation planning and infrastructure, including: the availability
of formalized emergency and evacuation procedures and plans; the location of emer-
gency response teams and their ability to respond to hazardous material release; access
and ease of emergency evacuation.
Subjective factors should reflect community priorities and values and should
preferably arrived at through community discussion and consensus. These factors are
particularly relevant in the assessment process when not one alternative is clearly
superior to the others. As such, whether or not the analyst chooses to select and apply
subjective factors which will depend upon the outcome of the risk calculations and how
conclusive the findings are.
24 CHAPTER 1
1.1.7. IMPLEMENTATION GUIDANCE
The guidance work sheets (Figure 1.8 and Figure 1.9) may be used as a guide in the
computation of the land use safety factors for assessment purposes.
Work Sheet for Compiling Route Characteristivcs and Mandatory Factors

A) General route characteristics
• Alternative No: .................................... Length: .................................... .Ian
• Origin: ....................... ....................... ... Destination: .................................................... .
• General description: ........................................................................................................ .
• Type of hazardous materials transported: .......................................................................
B) Mandatory factors
• Are there any physical constraints
(explain): ........................................................................................................................... .
(Explain): ...........................................................................................................................
• Explain any of the following subjective factors, as applicable:

- Special population: ......................................................................................................... .
- Special properties: .......................................................................................................... .
- Emergency response capabilities: .................................................................................. .
- Other factors: .................................................................................................................. .
Figure I.B. Example of a work sheet for compiling route characteristivcs and mandatory factors.
Calculation Sheet for Risk Estimation

Segment P HMA Population Density Hazard Index Population Risk
No
x .....•.................. x ....................... .
x ....•................... x ........................
x ....................... . x .••.................••.•
x ........................ x ...•...............•.•..
x ....................... . x ........................
x ........................ x ......................•.
x ....................... . x ........................
x •...••.............•.••• x ....................... .
TOTAL: ......... .
Figure 1.9. Example of a work sheets for computing land use safety factors
(Note: PHMA means the probability of a hazardous material accident).
The work sheets (Figure 1.8 and Figure 1.9) summarize the factors of influence that
must be considered when dealing with regional risk assessment and safety management
related to the transportation of dangerous goods.
1.2. Traffic and Operational Factors for Road and Ran Transport
The following traffic factors influencing the transportation of dangerous goods reflect
the ability of a route to effectively and safely move the traffic flows using it:
• Traffic volume and composition

• Carriage way level of service
• Structural and geometric adequacy of roads
• Number of traffic signals
• Travel time
• Availability of alternative emergency routes
An overview outline of each of these factors is provided thereafter.
1.2.1. TRAFFIC VOLUME AND COMPOSmON
The composition of vehicles by size and type is required to assess the road structural
adequacy as well as its operating level of service.
Traffic volume and composition along various sections and segments of the road
network may be obtained from published statistical information but preferably through
field screening surveys. Traffic volume may be expressed in terms of: annual average
daily traffic (AADT), and hourly traffic volume (average and peak). The directional
distribution of traffic should also be obtained. This information together with hourly
intersection counts can be used to estimate the peak directional hourly volumes along
all road sections within the study area.
Classification counts to establish the type of vehicles would differentiate, as a guide,
between: Light vehicles and Heavy vehicles (both rigid tankers and articulated tankers).
1.2.2. CARRIAGE-WAY LEVEL OF SERVICE
The level of service for a road section indicates the capability of roads for moving the
type and volume of traffic using it.
One definition of "Level of Service" is: qualitative measure describing operational

conditions within a traffic stream, and their perception by motorists and/or passengers.
It describes these conditions in term of several factors such as speed and travel time,
traffic interruptions, safety, driving comfort. A possible designation of the level of
service is from A to F, with level of service A representing the best operating conditions
(i.e., free flow), and F level of service the worst (Le., forced or break down flow).
26 CHAPTER 1
TABLE 1.5. Example of one-way traffic volumes (PCU) for urban roads at different level of service
(for interrupted flow conditions)
Type of Road Level of Service

Carriage-way
A B C D E F
--------------------------- PCU -----------------------------
2 Lane, 540 630 720 810 900 F
undivided
4 Lane, 900 1050 1200 1350 1500 0
undivided
4 Lane, 1080 1260 1440 1620 1800 R
undivided with
c1earway
4 Lane, divided 1140 1330 1520 1710 1900 CEO
with c1earway
6 Lane 1440 1680 1920 2160 2400 FL
undivided
6 Lane divided 1740 2030 2320 2610 2900 OW
with c1earway
PCU =Passenger car unit, i.e., heavy vehicle volumes are converted into passenger car equivalent.
TABLE 1.5 above shows suggested one-way hourly volumes for interrupted traffic
flow at different level of service.
A "road service volume" is defined as the maximum hourly rate at which vehicles can
reasonably be expected to traverse a point or uniform section of a lane or roadway during
a given time period under the prevailing roadway, traffic and control conditions while
maintaining a designated level of service.
It is suggested that for arterial/sub-arterial roads used for hazardous material

transportation, a level of service C would not be ideally exceeded with an utmost level
of service D in urban situations. Traffic volume estimated as per section 2.2.1 may
therefore be used to estimate the appropriate level of service of each road under consi-
deration.
1.2.3. STRUCTURAL AND GEOMETRIC ADEQUACY OF ROADS
The structural and geometric adequacy of the routes under consideration (to cater for
heavy vehicles carrying hazardous material) should be assessed. Routes with good
geometry (e.g., wider carriage-way with minimum horizontal and vertical curves) and
good line of sight should be selected in preference to routes of lesser quality. In
situation, where for other reasons, a route in the latter category was selected then it
should be upgraded to provide better geometry and reconstruct, if necessary, the
pavement to cater for increased volumes of heavy vehicles.
1.2.4. NUMBER OF TRAFFIC SIGNALS
The number of traffic signals is often used as a measure of delays along a route section.
A route with smaller number of signals would most likely be chosen as it would have
the potential for less delays.
1.2.5. TRAVEL TIME AND TRAVEL SPEED
Travel time for vehicles using a route indicate the congestion points as well as reflect
the level of congestion. Travel time information are usually available from transport
authorities or may have to be undertaken by way of field surveys. NAASRA has
suggested the following average vehicle travel speed for different level of service, as
shown below:
Average Vehicle Travel Speed ,

• Travel speeds in the range 25 kmlhr - 30 kmlhr corresponds to levels of service C-D
bordering the range of suitability for route selection .
• Routes with higher travel speeds are selected in preference to those with lower speeds.
1.2.6. AVAILABILITY OF ALTERNATIVE EMERGENCY ROUTE
In case of an emergency which would require the closure of a route designated for the
transport of hazardous material, an alternative route should be available.
TABLE 1.6. Example of travel speed and flows for different categories of level of
service
Level of Type of Flows Av. Overall Travel

service Speed (km/h)
A Free flow (almost no delay) >=50
B Stable flow (slight delay) >=40
C Stable flow (acceptable delay) >=30
D Approaching unsuitable flow >=25
E Unsuitable flow (congestion) at 25 approx.
F ---- S25
28 CHAPTER 1
1.3. Procedures for the Development of Estimated Truck Accident Rates
Recommendations have been made to the US Federal Highway Administration to

modify the equation used for representing the probability of a hazardous material
release given an accident involving a truck. IS, 16 The revised equation is given below:
(1.6)
where
PRj = Probability of a toxic release for route segment i due to a truck accident
TARi = Truck accident rate (accidents per vhc-km) for route segment i
P(RlA)i = Probability of toxic release due to a truck accident
Li = Length (kIn) of route segment i
The truck accident rate data used as default values in risk assessment and routing
studies should reflect the influence of highway geometry and traffic variables that have
a demonstrated relationship to truck safety. Two relevant variables are the highway type
(two-lane highway, freeway, etc.) and area type (urban/rural). It would be desirable to
also reflect the influence of other geometric factors such as lane width, shoulder width,
curves, grades, and intersections. In addition to truck accident rates, the distribution of
truck accident types also varies with highway and area type:
• Rural highways and urban freeways tend to have a larger proportion of single-vehicle
non-collision accidents, while
• Lower-speed urban highways tend to have a higher proportion of multiple-vehicle
collisions.
In general the probability of release given an accident involving trucks carrying
dangerous goods is much higher in single-vehicle non-collision accidents than in single-
or multiple-vehicle collision accidents. Thus the probability of a release given an
accident is also expected to vary for different highway and area types. Each accident-
involving vehicles should be treated as a separate observation (Le., an accident
involving two trucks should be counted as two accident involvement.
The data needed for analysis are:
Highway / Road Geometric Data. They should include: number of lanes, lane sepa-
ration, one-way/two-way, urban/rural. In principle different categories can be formed:
Highways
• multi-lane divided highways
• three-lane undivided
National and DepartmentaVCantonal Roads
• two-lane undivided (Departmental, and Cantonal Roads), with allowed
overtaking
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FAcroRS 29
• two-lane undivided (Departmental, and Cantonal Roads), with forbidden

overtaking
• three-lane undivided, with alternating overtaking track
Urban Freeways and Roads
• multi-lane divided highways
• multi-lane divided streets
• multi-lane undivided streets
• one-way streets
The recommended accident type categories to be used for classification and repor-
ting are shown in TABLE 1.7.
TABLE 1.7. Acx:ident type categories
A. Single-vehide Accidents
Non-c:ollision acddents Collision acddents
- run-off-road - collision with fIXed

obstacle
- Jackknife - collision with a parked
vehicle
- overturn - collision with a train
- fife - collision with a non-
motorist
- cargo spillage - other type of collision
- cargo shifting
- other non-collision
B. Multiple-vehide Accidents
- collision with a passenger truck

- collision with another truck
- collision with other vehicle types
In general, the reporting and description of the circumstances under which an

accident has occurred should imperatively be unified at the Federal or Cantonal level.
At present time, in Switzerland, every cantonal police office is using a different report-
ing protocol, if any, and usually too much leeway is left to the police officer for judging
of the content of his own report. It must be noted that many accidents result in material
damage only, and do not involve injuries to passengers or road users, nor have any
other consequence for the environment. In such cases it is not mandatory to even
announce the accident to a police office, and this type of accident will not be recorded
in the statistics of the cantonal police.
The reporting of accident by the police should in any case mention the type of
tanker or truck involved, and whether the tanker or truck was carrying a hazardous
cargo. It is important to record the kind of cargo transported (i.e., petrol, jet kerosene,
30 CHAPTER I
heating oil, other toxic liquids or solid products) according to the official dangerous
products list Tank trucks may be full or empty at the moment of an accident. Further-
more the report should also contain an indication whether the hazardous cargo was
released or not, and the kind of consequences resulting from the accident. In cases
resulting in pollution of the environment, as a consequence of leak, fire, or explosion,
the appreciation of the danger to the environment should include a statement concerning
the kind of pollution (i.e., air, soil, and/or water pollution), and extent of contaminated
area.
Truck Volume and Accident Data. Traffic volume data: These usually include data on
Annual Average Daily Traffic (ADDT) and may also include either the average daily
truck volume, or the percentage of trucks in the traffic stream.
Average Truck Accident Rates: In selecting accidents for inclusion in the analysis, it
is important to use the same definition of a truck (type) that was used in obtaining the
truck volume. Since detailed data for Switzerland are not always available it will be
necessary to use accident data concerning the whole truck volume in the country
(indifferent of their type) or to use data for the totality of the motorized traffic. The
Swiss Federal Office for Environmental Protection, Forests and Landscape (BUWAL)!
has issued a "Handbook No. III to the incident regulations for assessing the risks of
traffic ways", that complementS the regUlation and ordinance on incidents.!7 In fact the
Swiss authorities are recommending their own method for estimating truck accident
rates. The Swiss methodology (i.e., BUWAL "Handbook No. nr') has been summa-
rized and will be presented in the chapters on rail and road).
The average truck accident rate for each highway/road class can be computed as the
ratio of total truck accidents to the total vehicle-kilometers truck travel for that
highway/road class. The corresponding formula is given as:
A ..
TAR. = __1)_
(1.7)
J ATM;
where
TARj = Average truck accident rate for highway class j
AU = Number of truck accidents in one year on route segment i for
highway/road class j
ATM; = Annual vehicle mileage (kilometers) of truck travel on road i.
Accident Severity
The . data should be compiled to show the truck accident severity distributions by
highway/road/street class. The percentage of fatal and injury accidents should be
indicated.
Probability of Accidental Release

The distribution of accident types by highway/road class and the release probabilitit;,
for different accident types can be mUltiplied together to estimate the average release
probability for accidents on each highway/road class. The release probability for a
particular highway class is computed as the sum for all accident types of the proportion
of each type of accident, times the probability of release given an accident falling into
the category in question. The probability of release is the highest for the case of a
collision with a train, in case of overturn and run-off-road.
Canadian Review of Dangerous Goods Accidents. An analysis l8 for the region of

Alberta, Canada, has reviewed 181 regulated dangerous goods road incidents and
revealed that spills accounted for 53.6% of the incidents and rollover for 32.1%. Spills
occurred during unloading operations (21.6%) and during loading operations (14.4%).
The remaining 64% of the spills reported for the year 1991 occurred during
transportation. The causes were punctures to the container during manipulation, or were
the result of unsafe loading practice. Human error could be attributed to 86 road
accidents (47.5%) and equipment failure occurred in 39 cases (21.5%).
Among the human errors encountered, we fmd: discharge valve remained open
during transport, overfilling of the tank, flliing into wrong tank, wrong tank capacity
resulting in overflow spill, wrong or improper hose connections, etc. Other spills
occurred as a result of bad manipulation practice, piercing drums with a forklift tine, or
otherwise damaging the transport tank. Road conditions and poor driving techniques
(speed or sign violation, fatigue, inattention, etc.) are factors frequently contributing to
a rollover.
U.S Study of Accident Types. Another US study investigated whether the type of
accidents caused by trucking firms that haul hazardous materials differ significantly
from firms who do not haul these goods. The data were obtained from the US Federal
Highway Administration and covers the period 1986-1991. The raw data (13,498
reported accidents) were interpreted using a Poisson formulatIon in which the number
of accidents is the dependent variable, the explanatory variables are multipliable and
one takes the exponent of a coefficient in order to interpret it. The Poisson regression is
by definition non linear and fits an exponential curve to data. The method offers distinct
advantages when applied with count data (accident). 19, 20, 21, 22 For this study, a
regression with a constant and the log of the fleet kilometers was used as appropriate
relationship. A major advantage of the Poisson regression is that it can deal effectively
with datasets where a large proportion of the observations on the dependent variable,
i.e., accidents, take the value zero.
Experience with accident reporting has also shown that there are serious
inconsistencies and under-reporting of damage-only accidents. Hazardous materials
carriers in the U.S. are five time larger than carriers of non-hazardous goods in terms of
an annual fleet kilometers. In essence it was found that among hazardous materials
carriers, accident rates decline with firm size, but firms carrying hazardous materials
exclusively have an accident rate 11 % higher than the other firms and a rate of fatalities
and serious injuries that is 22% higher. Firms that carry hazardous materials in
combination with general freight, have an accident rate that is 18% higher and a
fatality/injury rate that is 24% higher. Private carrier are safer than for-hire carriers.
Haulers of gases in packages and liquids in tanks have the highest accident rates.
Carriers of hazardous wastes have the lowest accident rates. Long distance operations
32 CHAPTER I
(more than 160 km) are associated with accident rates 22% higher and fatalities and
injuries rates 53% higher than that of firms that are exclusively involved in short
distance operations. The accident on long-distance trip tend to be more serious and
result in a higher rate of accidents that involve fatalities and injuries, as well as more
property damage.
1.4. Risk Management for Dangerous Goods
Risk Management is defined as: "The impartial adjustment of conflicting claims in

selecting means to reduce the possibility of loss, injury or death". (according to a
definition from Webster Dictionary).
In order to manage or impartially choose between risks, it is necessary to place a
value on risk. A value is placed on the probability of loss of life.
The management of dangerous goods movements can be conceptualized and does
include transport means, time of shipment, safety plans, emergency response, etc. The
means of controlling risks are evaluated and compared to the cost involved. After
analysis of all factors of concern, a safety plan for risks can be established. This plan is
usually implemented partially. and will be reviewed in the event of accidents taking
place. A common finding in reviews of major disasters is that the existing safety plan
did not work and had it worked the disaster would not have occurred.
There are two approaches for managing the risk of dangerous goods movements:
a) After a disaster has taken place, inquiries and reviews of the accident and its
causes and the damages occasioned will take place. New regulation or
enforcement will be implemented in response to the disaster. Then with time and
experience, these measures are evaluated by considering the improvement in the
rate of incidents or the number of infractions, etc.
b) The second approach involves the government audit and approval for a total risk
management process and safety plan to be established by the industry or the
transport companies. The plan is evaluated on the likely achievement of
acceptable risk levels etc.
The chemical industry in North America has developed a Community Awareness

and Emergency Response (CAER) program. This program has two major tasks:
• To provide the public with information on chemicals manufactured or used at the
local plant and describe the emergency plan in place to protect them.
• To improve community emergency response planning by combining the
chemical plant emergency plan with other local emergency planing to produce a
more effective integrated community emergency plan.
The following tasks are important:
I. Identify emergency response industries and their resources
2. Evaluate risks and hazards
Vehicle Flow
(amount of traffic exposure)
Accident occurrence
..
(probability)
Tank Heating UDder Fire
..
(fire likelihood)
Safety Value Venting

(discharge preaaure and rate)
Tank Rupture
(timing I rupture preuure)
Vapor Cloud Ignition

(probability)
Overpresaure Propagation
..
(pressure decay)
Damage Occ\IITeDCC
..
(damage threaholds)
Damage Evaluation
(monetary damage cost)
Impact I Service Intenuption
Figure 1.10. Consequence pathway for a

spill accident involving a release of propane
gas.
3. Review and improve existing plans

4. Integrate plans into an overall community emergency plan
5. Obtain fonnal approvals from participating local governments
6. After the plan has been approved. it is necessary to arrange for training all
participants, testing the plan with drills and exercises (simulating real life
conditions) and revising it as necessary.
Such programs are necessary and useful to address rising public concern about
chemical hazards in the community and they contribute to improve the emergency
response and the good relationship between industry and the community under the
34 CHAPTER 1
hospices of the local government. A similar approach can be applied to the transport-
ation. storage and handling of dangerous goods.
1.4.1. A PATRWA Y APPROACH APPLYING TO THE TRANSPORTATION OF

DANGEROUS GOODS
Now considering more specifically the transportation of dangerous goods. a pathway

approach has been proposed by Van Aerde and Shortreed,23 (1989). A pathway
approach represents a simple but systematic procedure to establish a framework for
analyzing important inter-relationships and their relative significance. A spill conse-
quence pathway following a transportation accident may be established to identify the
therm~ynamic and physical relationships involved and to establish the criteria for an
analytical model of this sequence of events (see Figure 1.10).
A pathway approach allows an interdisciplinary input to risk analysis from a variety
of. sources of expertise and provides a focus for discussion and understanding of
interconnections between activities.
1.4.2. TRANSPORT OPERATIONAL COSTS AND OPERATOR'S

REQUIREMENTS
An important criterion in the assessment and selection of a route network for the
transportation of hazardous material is the relative cost of delays and travel time. The
analysis of this information would enable the determination of the economic
implications of particular routes for the transport of hazardous material and the
transport operator's requirements for practical transport economies.
Transport costs: They fail into two basis categories. i.e.• fixed costs and variable
costs (usually referred to as operating costs). Generally, the former costs do not vary
significantly with the vehicle-kilometers traveled. If the tanker carrying hazardous
materials need to change to a route of a longer or shorter distance. (only) the operating
costs will vary and be higher or lower respectively. In many cases, both operators' cost
requirements and operators' convenience result in the use of the shorter route
irrespective of safety implications.
Operating Costs. They are based on two main components: - a variable cost for
operating the road tanker (RT), and the cost of the driver's time (Dn.
Total Vehicle Operating Cost = (RT X D) + (DT xT)
where
RT : Unit cost component for the road-tanker ($Ikm; SFrlkm)
D : Distance traveled (km)
DT : Unit cost component for the driver's time ($Ihr; SFrlhr)
T : Time taken to travel the distance (hr)
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FACI'ORS 35
These factors could be reflected by the distance traveled and the travel time along
the route.
For the above, it is indicated that the main cost criteria when assessing or comparing
alternative routes for the transportation of hazardous materials is the expected increase
or decreases in travel time (the main component that influence operating costs).
An increase or decrease in operating costs of over 10% is considered to have a
significant effect on the cost of transporting hazardous materials. It is also considered
that the distance cost could increase further as long as the travel time was within the
10%. For example, a longer distance route could have, or be developed with less con-
gestion thereby resulting in a travel time about the same as the shorter route.
Summary
This chapter focused on the analysis and assessment of transportation routes for the
carriage of hazardous materials. It introduces issues related to the integrated risk
assessment approach to the safety transportation of hazardous substances, namely:
transportation risk and environmental land use factors, capability of the existing road
network and cumulative traffic implications and economic distribution considerations
and operator's requirements for practical transportation economics. The nature of the
major risk to be assessed is included in this chapter. Systematic methodological issues
are outlined for the assessment of transport risk, and procedures for the development of
estimated truck accident rates, transport operational costs, etc., are proposed.
36 CHAPfER 1
References (Chapter 1)
I BUWAL, Bundesamt fiir Umwelt, Wald und Landschaft, Bern.

2 Luckritz, R. T.; Schneider, A. L., Decision making in hazardous materials transportation, J. Hazard. Mater.
4 (1980), pp. 129-143.
3 RID, Regulations concerning the Internatioanl Carriage of Dangerous Goods by Rail, an annexe to
Convention concerning International Carriage by Rail (COTlF).
4 RSD, "Reglement Suisse Concernant Ie Transport des Marchandises Dangereuses par Chemin de Fer".
(RSD is the Swiss regulation corresponding to the international RID regulation). (RIDIRSD): - Ordnung
fiir die internationale und schweizerische EisenbaImbefllrderung gefllhrlicher Giiter (RIDIRSD) vom
1.1.1990.
'ADR, European Agreement Concerning the International Carriage of Dangerous goods by Road
(Europiiisches Uebereinkommen tiber die internationale Befllrderung gefllhrJicher Gilter auf der Strasse
(ADR), vom 30. September 1957 (Stand: I. Mai 1985).
6 SDR, Verordnung tiber die Befllrderung gefahrlicher GUter auf der Strasse, vom 17. April 1985. (This is the
Swiss regulation corresponding to the international ADR regulation).
7 Substance Hazard Index (SID), American Petroleum Institute, Management of Process Hazards,
• Recommended Practice, 750. First ed., Washington, DC: APL (1990).
8 RSPA: Transportation Safety Information Report, 1987, Annual Summary, DOT-TSC-RSPA-88-3. U.S.
Department of Transportation, (November 1988).
• Nicolet-Monnier, M.; Gheorghe, A.V.• Integrated Regional Risk Assessment, Vol. II, Consequence Assess-
ment of Accidental Releases, KLUWER Academic Publishers B.V., Dordrecht, The Netherlands,
(1995).
10 Computer code ETA: Available from Science Applications Int. Corporation, 5150 EI Camino Real, Los
Altos CA (USA).
11 Computer code R1SKMAN: Available from Pickard, Lowe & Cranick 7760 University Drive, Newport
Beach, CA (USA).
12 Computer code CAFTA-PC: Available from Science Applications Int. Corporation, 5150 EI Camino Real,
Los Altos CA (USA). '
13 Computer code TREDRA: Availahle from JBF Associates Inc. Technology Drive, 1000 Technology Park
Center Knoxville TN (USA).
14 Computer code BRAVO: Available from JBF Associates Inc. Technology Drive, 1000 Technology Park
Center Knoxville TN (USA).
I' Russel, E.R.; Harwood, D.G., Procedures for the development of estimated truck accident rates and release
probabilities for use in HAZMAT routing analysis, in F. Saccomanno and K. Cassidy (eds.),
Transportation of Dangerous Goods: Assessing the Risk, Institute for Risk Research (lRR), (1993). pp.
247-262.
16 Barber, E.J., Hildebrand, L.K., Guidelines for Applying Criteria to Designate Routes for Transporting Haz-
ardous Materials, Report No. FHWA-IP-80-15, Federal Highway Administration (U.S.A.).
17 BUWAL, Handbuch ill zur SWrfallverordnung, Richtlinien fiir Verkehrswege, Dezember 1992. Bundesamt
fUr Umwelt, Wald und Landschaft, Bern.
18 Hammond, S.P., Severity ratings and incident trend analysis, Proceedings, ER91, AB; See also: Hammond,
S.P.; Smith, W.W., Historical Analysis of Dangerous Goods Spills in Alberta, in F. Saccomanno and K.
Cassidy (eds.), Transportation of Dangerous Goods: Assessing the Risk, Institute for Risk Research
(lRR). (1993).
19 Rose, N, Profitability and Product qUality: Economic determinants of airline safety performance. Journal of
Political Economy, 98 (1992), pp. 944-964.
20 McCullagh, P.; Neider, J., Generalized Linear Models, (1993). Chapman and HaIl, London.
21 Hausman, 1.; HaIl, B. H.; Griliches, Z., Econometric models for count data with an application to the
patents and R&D relationship. Econometrica, 52 (1984), pp. 909-938.
22 Cameron, A.C.; Trivedi, P., Econometric models of count data: Comparisons and applications of some
estimators and tests. J. of Applied Econometrics, 1 (1986), pp. 29-54.
23 Van Aerde, M.; Shortreed, 1. H., The Pathway Approach to Risk Assessment, in Risk Management for
Dangerous Goods, Ed. J. H. Shortreed, University of Waterloo Press, (1989), pp. 13-21.
CHAPTER. 2
ASSESSMENT OF TRANSPORTATION RISKS IN A COMPLEX AREA
2.1. Overview and Scope of Transportation Risk Analysis
There are new trends in the chemical industry to manufacture dangerous products at the
site of utilization or, when possible, to modify the synthesis method in such a way as to
avoid the use of hazardous substances. Integrated concepts are developed, that provide
for the minimization of hazardous wastes through reduction, reuse, recycling and
recovery at source, as well as addressing treatment and disposal of material not
managed at the source. Historically, the siting of hazardous waste treatment and
disposal facilities has been and is still a difficult undertaking, raising public fear. The
transport and distribution of hazardous substances, such as petroleum products,
liquefied petroleum gases, ammonia, chlorine gas, pesticides, chemicalslpetrochemicals
and radioactive materials, inevitably involve the potential for incidents and accidents
which 11Uly result in death or injury to people, property damage or da11Ulge to the bio-
physical environment through the effects of fire, explosion or toxicity. An increasing
number of transportation accidents involving hazardous substances have occurred
worldwide. Such accidents with their resultant effects on people and the environment
have increased awareness in government, industry and the community and resulted to
think again of the risk assessment process applied to transportation of hazardous
substances. In that context, it is now recognized that the safety planning of transport-
ation routes, accounting for the type and nature of surrounding land uses, is an integral
component of the safety 11Ulnagement of hazardous substances transportation.
Delineating hazardous substances transportation routes is, as such, a significant and
essential complementary measure to technical and operational safety and environmental
controls on the hazardous substances containers and associated regulatory processes. It
is relevant to note that fixed installations are more amenable to local, organizational and
operational hazard controls. Transportation systems are dynamic systems with
additional external variables (e.g., drivers, traffic conditions, etc.) difficult to bring into
one overall control system.
The scope of regional risk assessment concerning transportation of dangerous goods
has to include all transport systems, such as rail, road, ship, and pipeline networks. The
main goal of such studies is to provide the (Federal, Departmental, or Cantonal)
Authorities with facts on actual risks incurred in a given region, on which to base
political decisions regarding the course and costs of mitigation actions and improve-
ments to be brought to existing transportation networks and distribution systems. One
has to learn from the mistakes done in the past and improved concepts and design
should be aimed at when planing new traffic ways and traffic control systems. Risk
cannot be eliminated totally, but reduction to an acceptable level is, most of the time,
38 CHAPfER2
possible. The cost of increased risk reduction have to be balanced with the social
benefits which can be expected from such reduction measures, since fmally the extra
expenses must be financed by both industry and Community. Quite often, in recession
crisis, abatement measures for environmental protection (noise, pollution) or for
improving the security of transportation systems and distribution networks are
postponed, because public money is simply not available at the moment.
Although the Authorities are aware of the danger and drawbacks of actual or planed
transport systems, it is quite often not possible to chose a different concept or another
layout due to natural constraints such as the presence of rivers, ground water, valleys
and mountains. In many cities built during the middle age, and which escaped bombing
during the second World War, the traffic conditions are difficult due to tortuous or
narrow roads. In many cases the whole traffic has to pass through the center of the city
or village and by-pass road are non-existent. What concerns the implementation of new
industries or new residential areas, the case is different since alternative choices can be
found and it is usually possible to minimize risk by sound selection of sites, avoiding a
cumulating of activities presenting a risk to the population.
The orderly development of industry and the protection of community safety
necessitate the assessment of hazards and risks. Recently the Swiss Federal Office for
Environment, Forestry and Landscape (BUW AL)1 has issued different handbooks No. I
through III, in the form of guidelines, of which "Handbook No. m, was prepared for
"assessing the risks of traffic ways". It is a complement to the Ordinance of April 1,
1991 concerning the "Protection Against Incidents" (Storfallverordnung, StFV)2.
The assessment of risk necessitates the establishment of criteria against which
judgments can be made as to the compatibility of various land uses. To date, risk
assessment criteria in Switzerland, have been applied on a case-by-case basis by the
different cantonal departments (e.g., Dept. of Planning, of Road Construction, etc.). In
Switzerland the departments in charge of "regional planning" and "risk assessment"
should now consider it appropriate to agree upon and formally issue a set of criteria,
based on accepted international experience, which shall assist in a consistent approach
to regional risk assessment and in the decision making process in the Community.
At this place it is worth noting that the denomination for the same kind of department,
or cantonal office/authorities may differ quite a lot from one Swiss Canton to another,
depending also of the respective official language in use (i.e., French, German, Italian,
and Romansch).
2.2. Risk Estimation and Risk Assessment
2.2.1. QUANTITATIVERISK.CRITERIA
The results of a hazard assessment is a quantification of risk. The level of risk which is
acceptable is a difficult question to answer. Quantitative risk criteria must be developed
by the competent authorities, which can be used by industry to provide a basis for
ASSESSMENT OF TRANSPOTATION RISKS IN A COMPLEX AREA 39
responsible decision-making on risks. Some of the problems which arise in this area are
those of:
• individual and average risk
• death versus injury
• maximum risk to employees
• maximum risk to public
• value of human life
• methods and criteria for comparing human loss and property damage
• multiple fatalities
• engineering feasibility
• allocation of resources, etc.
Several indices of industrial injury risk are used in the literature:
• Lost Time Accident Rate (number of accident per 108 hours).
• Fatal Accident Frequency Rate (FAFR), defmed as number of deaths per lO8
hours. (The Fatal Accident Rate (FAR) of the chemical industry may be
expressed as deaths/l08exposed hours).
• Death Rate per Annum (DRA).
One method of assessing the value of human life is to calculate the value of the
future production or services which an individual may be expected to give to the
community. A variant of this is to estimate future earnings, although these are not
synonymous with the value of the work done.
Risk C, fteria
The primary criterion is that the maximum risk to an employee is limited to the Fatal
Accident Rate (FAR) of the chemical industry (Kletz 3, 1971; Kletz4 , 1972). Another
primary criterion is that the maximum risk to a member of the public is limited to a
level comparable with other involuntary risks (target: risk of death < 10-7 per year
averaged over the whole population). The secondary criterion is that the value of a life
method is used to assess additional expenditure aimed at eliminating other hazards.
The risk criteria just described need to be supplemented by an additional criterion
for multiple-casualty accidents. In such a case it is assumed that there should be no
special weighting designed to reduce the probability of death due to a multiple-casualty
accident as opposed to death due to an individual accident. This can be represented on
a log-log graph of accident frequency versus number of deaths, using a slope of -1.
A further criteria which needs to be satisfied includes limits on the total frequency f
of accidents and on the risk r to individuals_ Thus:
n
f = "2Ji
i=!
(2.1)
where, fi is the frequency of a particular type i of accident, and n the number of

accident types. The risk r to an individual may be computed as:
40 CHAPTER 2
1
r=-Lxj; ,
II
(2.2)
N i=1
where, Xi is the number of deaths in a particular type of potential accident; /; the

frequency of such an accident; n the number of types of potential accident; and N the
total number of people at risk.
In 1985, the Health and Safety Commission (HSC), in Great Britain, has examined
the potential risks of the transport of certain dangerous substances and the results were
presented in a HSC reportS in 1991. The terms of reference covered not only the
consideration of major hazard aspects of the transport of dangerous substances but also
the identification of appropriate control measures and advice on additional action that
might be necessary. They did not include radioactive substances, transport by air or by
pipelines or risk to the environment.
It was recognized that existing accident statistics alone provide an inadequate basis
on which effectively to judge the risks involved in rare but potentially high conse-
quence events. It was concluded that a more appropriate approach is to use the
quantified risk assessment (QRA) methodology. The report acknowledges that the risks
estimated by this method are subject to some uncertainty, but judges that QRA has
provided the best estimates of the risks involved, provided valuable insights and
assisted the Sub-Committee in reaching its conclusions.
QRA can only provide numerical estimates of the risks; it cannot, by itself, provide
conclusions as to whether or not those risks are low enough or too high. Such judgment
are reserved to the decision-maker.
2.2.2. DEFINmON OF INDIVIDUAL AND SOCIETAL RISK CRITERIA
Risk is defined as the likelihood of any adverse outcome. The suggested risk criteria are
probabilistic in nature. That is, they account for both the involuntarily consequences
(effects) and likelihood probability of hazardous events.
All activities associated with the storage, transportation and handling of dangerous
goods have an associated level of risk. Risks can be assessed and managed, but it is not
possible to eliminate totally a risk unless the activity itself is eliminated. In many cases
this simply leads to risk transfer (for example from train to road, or vice versa) which is
an important concept in risk assessment and management. The criteria are therefore
based on the concept of a residual risk, the acceptability of which should be established
in relation to various land uses.
A rational decision making process requires the establishment of a consistent
framework with standards to express the desired level of safety. Probabilistic Safety
Criteria (pSC), which are quantitative expressions for the probability of occurrence of
an undesirable event within a given period of time, can play the role of such standards.
The purpose of this section is to provide a general guidance concerning the setting
and applications of such criteria.
Public health risk should be assessed separately from environmental risk. In terms of
health impacts, occupational and public risks should be treated separately. Two
categories of risk apply as a result of direct or indirect impacts:
• Fatal effects, either immediate (resulting from direct exposure or accidental

situation) or delayed (resulting from chronic exposure to hazardous substances).
• Non-fatal effects, (injuries, diseases) of either an immediate or delayed nature.
In relation to environmental risk, categorization of risks can be made on the basis of
extent: local, regional and global; and on the duration of the effect: short or medium
term, and long term. Some environmental effects are of such a long term nature that
they are virtually (or actually) irreversible.
Risks from transport, transfer, or routine handling operations (like loading/unloa-
ding, or pumping operation) should also be differentiated from those resulting from
major accidents. The criteria proposed in this section refer to this latter type.
To date, emphasis has been placed on the development of risk criteria in terms of
acute (i.e., immediate) health effects, mostly fatalities, and in some cases, immediate
injuries to people. For the long-term effects of chemicals, the assessments have until
now relied mostly on translating animal tests results to people. Recommendations
established by National Health Councils are relied upon in that regard There are also
very few cases of probabilistic safety criteria that apply to accidental releases of
chemicals into the natural environment. The diversity of response mechanisms (in type
and nature) to the multitude of species within the different ecosystems, including the
issue of irreversibility and/or recoverability of damage make it difficult to establish a
uniform criteria in this area. Such criteria will largely depend on local circumstances
and may need to be developed on a case by case basis.
There are two dimensions of risk which should be considered separately, individual
and societal (Figure 2.1). On the one hand, the individual's concern about their own life
or safety is mostly independent of whether the risk is from an isolated incident or a
large-scale disaster. Society's risk perception, however, is mostly influenced by multi!,le
fatality or injury disasters.
There is a general agreement that societal or group risks should be considered
when assessing the acceptability of any hazardous activity or industrial facility. There
are two components to the societal risk concept. First, the number of people exposed to
levels of risk is important. Second, society is more averse to incidents that involve
multiple fatalities or injuries than to the same number of deaths or injuries occurring
through a large number of smaller incidents. For example, society reacts differently
(and generally with greater concern) to an aircraft crash involving a number of injuries
or fatalities than to repeated motor vehicle accidents involving smaller numbers of
injuries or fatalities at a time. To deal with this aspect of risk, the intensity of use and
the density of people need to be considered. The nature and scale of the incidents
contributing to the particular risk levels at particular points and the outcomes of those
incidents in terms of fatality and injury, also need to be considered.
Societal risk analysis combines the consequences and likelihood information with
population information. It is usually presented in the form of an F-N curve, which is a
graph indicating the cumulative frequency (F) of killing (N), or more people.
Societal risk is calculated from the same basic consequence calculations used to
estimate the individual risk. However, instead of presenting the risk summed for all
incidents, independent of the population density (which is the method used for
42 CHAPTER 2
individual risk), each possible incident outcome is considered in tum and its frequency
(F) and the numbers of people (N) that could be affected by it, are recorded as an F-N
pair. The calculation of the number of people affected includes an allowance for the
average geographical distribution of the population.
Some Definitions Conceming Risk

Annual fatality ra.tes are measures of societal risk, which is equal to the expected long-
term average number offatalities per year.
Mean fatalities per event express the average number offatalities in given accidents. This
is equal to the fatality rate divided byte event frequency.
Individual and societal risk criteria are used for ranking risk sources in order of priority
and to indicate whether the estimated risks are tolerable or whether remedial measures are
to be taken.
Individual risk is expressed as the likelihood of deaths per year for an individual at risk.
Societal risk is usually defined as the relationship between the number ofpeople killed in a
single accident and the chance or likelihood that this number will be exceeded. It is usually
presented in the form of an "F-N curve".
FN curves are presented as a graphic indicating the cumulative frequency (F) of killing N
or more people. They show the relationship between the number of fatalities (N) and
frequency (F) of 'N' or more fatalities per year.
Figure 2.1. Risk definitions.
o Definition 1: Societal risk is usually defined as the relationship between the

number of people killed in a single accident and the chance or likelihood that this
number will be exceeded. It is usually presented in the form of an "F-N curve", which
is a grapr.ic indicating the cumulative frequency (F) of killing N or more people.
Groli!' risk does not involve the calculation of the "individual risk of death" but
rather the "risk of a number of deaths".
There are many ways of expressing the societal impact of serious accidents, such as
the number of predicted, prompt, or latent fatalities; agricultural restrictions; large scale
evacuation and economic loss. There is no international consensus on which of these or
other measures should be chosen to develop societal risk criteria. Individual countries
will need to choose the impacts of greatest concern to them.
o Definition 2: Individual risk is usually defined as the probability per year that
anyone person will suffer a detrimental effect as the result of exposure to an activity.
Probabilistic safety criteria for individual risk are proposed under the consideration
that risks arising from accidents in hazardous installations/activities should present only
a small increment to the risk to which individuals are already exposed.
The criteria is intended for application to an individual risk calculated using the
following assumptions:
the individual should be considered to be resident at the location off-site, or
close to the transportation road, yielding the largest risk for a representative
period of time
the individual should be considered to be an average individual with respect to

dose susceptibility
atmospheric dispersion calculations should be realistic, i.e., making allowance
for the variability in weather and wind direction.
Whilst individual fatality risk levels include all components of risk (i.e., fire,
explosion, and toxicity), there may be uncertainties in correlating toxic concentrations
to fatality risk levels. The interpretation of "fatal" should not rely on anyone dose-
effect relationship, but involve a review of available data.
The following notes are provided to assist in the formulation and implementation of
appropriate risk assessment criteria:
• The individual fatality and societal risk criteria should include all components of risk,
i.e., fire, explosion and toxicity.
• The implementation of the criteria must acknowledge the limitations and in some
cases the theoretical uncertainties associated with risk quantification. Two approaches
are usually adopted to account for such uncertainties: a ''pessimistic'' approach, i.e.,
assumptions are on the conservative side with overestimation of the actual risk; or "best
estimates" using realistic assumptions with an estimated risk that could either be an
overestimate or an underestimate of the actual risk. The criteria suggested in this section
are set at a realistic level.
• the risk to an individual and/or to the public in the vicinity of an industrial site, arise
from all industrial activities in the area, including storage, handling, and transportation.
The basic risk criteria (to various land uses) need to be related to the site. It may also be
appropriate to plan for sub-criteria for each individual site to account for cumulative
impact of developments.
• In a large industrial complex, risk criteria should also provide for the potential for
accident propagation. The risk of an accident at one plant triggering another accident at
another neighboring plant should be kept low. Adequate safety separation distances
should be maintained. This also applies to railways in urban areas, where the distance
from the railroad to the next row of houses is quite often too small.
2.2.3. ENVIRONMENTAL RISK CRITERIA
In addition to the risk to people and property, the impact assessment process for
potentially hazardous storage, or loading/unloading installations, or transport systems,
must also consider the risk from accidental releases to the biophysical environment. Fire
and explosion hazards are of less relevance to the environment, in comparison to the
effect these hazards may have on people. Acute and chronic toxicity impacts are those
which must be chiefly addressed. Generally, there is less concern over the effects on
individual plants or animals. The main concern of fire and explosions hazards is instead
with whole systems or populations.
44 CHAPTER 2
The assessment of the ultimate effects from toxic releases into the natural ecosystem
is difficult, particularly in the case of atypical accidental releases. Data are limited and
factors influencing the outcome variable and complex. There may be no immediate loss
of plants or animals or other observable effects from single releases, but there may be
cumulative and synergistic effects. It is therefore appropriate to ensure that a thorough
review of available data are undertaken and best available information used in the
assessment process.
In many cases, it may not be possible or practicable to establish the final impact of
any particular release. It may be appropriate in such circumstances to assess the
likelihood of identified concentrations of concern occurring in the air, water or soil.
Where such criteria are used the assessment should remain on the conservative side.
Because of the complexities of such assessment and case-to-case differences, it is
inappropriate to specify general criteria. The acceptability of the risk will ultimately
depend on the value of the potentially affected area or system to the local community
and wider society. For example, where a rare or endangered ecosystem or species is
involved, a much lower risk level is necessary than where the potentially exposed area
or system is degmded and/or common.
2.2.4. HOW TO COMBINE CONSEQUENCES AND FREQUENCY ESTIMATES

TO PRODUCE RISK MATRICES
The basis for a quantification of the risk from a hazardous or industrial activity is a list
of hazardous events, or groupings of like events which can be considered to produce
similar consequences. The frequencies of these events may be estimated. There may be
a range of possible outcomes from each event, depending on the different circumstances
which may prevail: for example wind direction, weather category, location of people
etc. Each of these cj.~cumstances must be defined and a probability assigned to them.
The aggregation of frequency and consequence analysis can therefore be complex,
although it is conceptually simple, because all analyses follow essentially the same
procedure.
Damage-causing events must be related to the undesired initiating events: for
example, the various possible outcomes arising from a release of flammable material
may be modeled using an event tree. The conditional probability for factors such as
wind direction towards ignition sources and chance of ignition at each source can then
be used to produce a frequency for the damage-causing event from the frequency of the
initiating event.
The consequences of each damage-causing event must be assessed: The usual
approach is to define ranges to selected casualty probabilities from a combination of
effect and vulnerability models (Figure 2.2). These causality probabilities may be
selected apd limiting ranges to each value estimated - for example, the probability of
casualties occurring at various over pressures could be used in conjunction with an
explosion overpressure model to produce radii to selected casualty probabilities.
Determine Hazard Scenarios for

Storage Vessel, RailJRoad-tanker
Pipe break, valve leaking, release valve damaged
Vessel failure (type: cooled, pressurized)
Pressure vessel failure mode (partial or total)
Fire, explosion, BLEEVE
Establish Release Mode of Hazardous Product

Release of toxic liquidS/solids
Release of toxic vapors
Release of thermal/nuclear radiations
Determine Mode of Action I Transmission to Target

Airborne dispersion (lighterlheavier than air)
Waterborne dispersion (water-soluble, ecotoxic)
High thermal radiation effect
Overpressurelblast effect
Projectiles/missiles
Calculate Effects on Target

Exposure to toxic vaporslliquids
Thermal radiation effects
Nuclear radiation effects
Blast overpressure effects
-
Take Allow~~ for Mitigating pifects
Shelter aVailability
Evacuation I Rescue
First-aid I medical treatment
I
Assess the Consequences
Damage to people (probability of deaths, injury)
Damage to property (frrelstructural damage)
Damage to environment
Assess fmanciallosses and associated cost
(operating loss, remedial actions)
Figure 2.2. Method for assessing the consequences of hazardous

product release.
The selection of probabilities will usually depend on the available data underlying
the vulnerability model used - the analyst should be wary of using probit type relation-
ships to produce a large number of casualty probability bands as this not only compli-
cates the analysis, but the degree of detail would not be supported by the basic data.
46 CHAPI'ER2
Having obtained the frequency and casualty probabilities for each damage-causing
event under consideration, the risk relationships are derived in the following manner.
Taking each event in turn, the number of people present in the area covered by each
casualty probability band are multiplied by the appropriate casualty probability,
producing, by summation, the- total number of people predicted to be affected by each
event. The overall frequency-consequence relationship can then be drawn up from the
number (e.g., people, property, land, or water resources) affected and the frequency for
each event. Expressing the risk in terms of the frequency distribution of multiple
casualty events (N-F-curves) is known as "calculating the societal risk".
QUIlntijicalion of Risles
Step 1: Estimate the frequencies of these events.
~: Define the circumstances concerning a given scenario and assign a conditional
probability to each of them to produce a frequency for the damage-causing event from the
frequency of the initiating event.
S1HL1: Define, from a combination of effect and vulnerability models, several conse-
quence ranges which are related to the selected casualty probabilities.
Step 4: Derive the risk relationships for each event, and multiply the number of people
present in the area (corresponding to each casualty probability band) by the appropriate
casualty probability. Finally the total number of people affected by each potential event is
obtained by summation.
~: Calculate the societal risk by estimate the overall frequency-consequence relation-
ship from the number (e.g." people, property, land, or water resources) affected, and the
frequency for each event. Express the resulting risk in terms of the frequency distribution of
multiple casualty events.
Step 6: Calculate the individual risk at a given location by taking the casualty probability
at that location for each damage-causing event, and multiply it by the frequency of that
event. Obtain the individual risk from all such events, (and therefore from the activity as a
whale) by summation over all the events.
Figu~ 2.3. Methodology for risk quantification.
The individUill risk at a location is obtained by taking the casualty probability at that
location for each damage-causing event and multiplying it by the frequency of that
event. The individual risk from aU such events, and therefore from the activity as a
whole, is obtained by summation over an
the events.
The final expression of individual and societal risk then incorporates the likelihood
and severity of all the outcomes of the scenarios that have been considered for particular
location. The following sections illustrate the way in which the calculations can be
carried out (see also Figure 2.3).
Procedure/or Estimating Societal Fatality Risk Levels

For estimating the societal risk R. from a single event, one must evaluate the probability
s of a given number of people n being subjected to a specified level of injury in a
specified time interval.
= probability of a hazardous event, such as a pool fire, torch, flash fire,

fireball or explosion.
= likelihood that an individual at a defmed location will be subject to a
casualty with a specified level of injury from such a hazardous event.
= likelihood that the individual will be at such a location when the
hazardous event occur.
Procedure for Estimating Societal Fatality Risk Levels

Step 1: Define the type of injury for which risk is to be evaluated.
Step 2: Define the casualty numbers, N, which are to be used in expressing societal risks.
~: Select a release, j, with associated probability Prei.
Step 4: Define the hazardous event from that release with a probability Phaz = s.
Step 5: Use damage models to evaluate the number of casualties, n. This can be done by
allocating ranges of exposure to discrete values of Pc and evaluating the number
ofpeople n who fall within each exposure range.
~: Repeat steps 3-5 for all releases.
Step 7: Select a value for N.
Step 8: Sum all values of p. for which n > N.
Step 9: Repeat all steps 7 and 8 for all values of N.
Figure 2.4. Methodolo~y !or estimating societal fatality risk levels.
Note that in above formula the sum is computed over all people within the area
affected by the hazardous event.
When evaluating the number of people affected by a given hazardous event, the
following expressions are often used:
s = P",I P.IC and n = L,n(k)~ (k) (2.4)

R. = P",I P..c L,n(k) ~(k) (2.5)
where
P,.I = probability of a gaslliquid release
p ..c = probability of the hazardous event occurring as the result of the release
n(k) k-time-averaged number of people subject to an average casualty c, with
probability Pc (k).
Societal fatality risk from a number of events is normally expressed in a cumulative
form, i.e., as a probability of arriving at N or more casualties. This is usually expressed
for a number of values of N and is obtained by summing all values of R. for events
where n~N.
48 CHAPfER2
The procedural steps summarized in Figure 2.4 may be of relevance in the

estimation of societal risk levels.
Procedure for Estimating Individual Fatality Risks

Following the definition of 'individual fatality risk given in the previous section, the
individual risk for a single event and consequence type, can be detennined from the
following basic expression:
(2.6)
where
PIwz = probability of a hazardous event, such as a pool fire, torch, flash fire,
fire ball or explosion
Pc = likelihood that an individual at a defined location will be subject to a
casualty with a specified level of injury from such a hazardous event
P acc = likelihood that the individual will be at such a location when the
hazardous event occur.
Procedure for Eslimllting lrulividrull Falality Risks

Step I: Define the location of the individual relative to the point of release
Step 2: Define the type of injury for which the risk is to be evaluated
Step 3: Select a release scenario, with a probability P,.I

Step 4: Establish the probability of the hazardous event occurring, given the
release: This is equal to PIwz , (with PIwz = P,.I . Probability of escape for an
event becoming a hazardous event).
Step 5: Use an appropriate consequence rrwdel to determine Pc for the location selected -
Pc being the casualty probability due to the event consequence at this location.
For hazardous events with immediate ignition at or near the source, Pc is directly
related to PIwz . For dispersing toxic clouds, the probability of wind direction and
other meteorological conditions (e.g., stability class) would have to be taken into
account.
Step 6: Postulate the probability that an individual will be present at that location - P DeC
(assume P occ = 1 for the person rrwst at risk approach).
Step 7: Calculate the individual risk Re = PIwz p.: Poce .

Figure 2.5. Methodology for estimating individual fatality risks.
When evaluating individual risks, assuming the "person at most risk approach" is
=
adopted for a given location, then it is usually assumed that P DeC 1. Expanding PIwz in
terms of the likelihood of a release Prol and the probability that it will give rise to the
hazardous events p.se leads for the individual most at risk to the expression:
(2.7)
wbere
PreI = probability of gaslliquid release
Puc = probability of the hazardous event occurring as the result of the
release
For a number of releases and hazardous events, individual risks are summed, over
all possible accident scenarios and types of accidents, i.e.,
(2.8)
The steps summarized in Figure 2.5 may be give guidance in the computation of
individual risk of fatality.
2.2.5. DEFINITION AND ACCEPTANCE OF RISK LEVELS
Acceptability of a given level of risk involves many considerations of whicb safety is

only one, althougb safety is playing an increasingly important role in planning
considerations. Attitudes towards risk acceptability can vary widely depending on local
situations. In some cases, certain risks may only be acceptable when they are out-
weighed by certain advantages which people associate with the considered activity.
However, unacCeptable risks can be shown to exist in a region, whatever the advantages
maybe.
The basis for risk criteria is that, generally, various levels of risks are tolerated on a
daily basis, both to individuals and to society as a whole. Where risk is taken with the
choice and full knowledge, that risk can be described as voluntary risk. Examples of
voluntary risk include smoking, driving, and rock climbing, provided that the individual
knows and understands the risks. Where the individual does not bave knowledge of the
risks, or is not entirely free to choose to avoid the risk exposure, then the risk can be
termed non-voluntary. Examples of non-voluntary risks include meteorite strike, some
illnesses and some natural disasters. In reality, most types of risk exposure have degrees
of both the voluntary and non-voluntary. People in general are willing to expose
themselves to quite higb levels of individual risk by undertaking certain activities. On
the other hand, society offers growing resistance to risks perceived as being imposed on
one group of people for the benefits of others, or where the risk exposure of one group
does not fit with their share of benefits is usually perceived as a non-voluntary risk.
Societal Risk Levels

When a risk is to be imposed on an individual or a group of people (e.g., by locating
hazardous activities - e.g., transportation, storage, industry - in an area), the concept of
acceptability of that risk for the decision-making process is that it should be low relative
to otber known and tolerated risks. As societal risk acceptability is specific to eacb
society, it is very important that allowance be made to reflect differences between
societies and cultures. The F-N approach is a complex one whicb does not lend itselt to
easy implementation. It may not be possible in practice, for example, to account for
50 CHAPTER 2
small but incremental increases in population over a period of time which may
eventually have significant implications on a cumulative basis.
In view of these considerations we suggest that:
• Judgments on societal risk shall be made on the basis of a qualitative approach
on the merit of each case rather than on specifically set numerical values.
• Individual fatality and injury risk contours at the individual risk criteria levels
(appliCable to the various land use categories) should be established. People
density within each risk band should then be established and the suitability of a
particular regional development judged on the basis of the location, density and
vulnerability of the exposed population.
In practice, the frequency of each potential accident and the number of people that
may be affected by each accident should be estimated wherever practicable. F-N curves
should be generated and used as an input to the risk analysis. An example of societal
risk curves for some human-caused events, that was adopted by the USA, is shown in
Figure 2.6.
In general, the individual risk criteria for particular land uses are set with regard to
the lower end of the range of typical population densities for that use. Where individual
developments (industrial or residential), or a number of developments taken together
result in substantially higher densities, their location relative to the suggested criteria
contour should be carefully considered. Generally the closer to the limit for that type of
land use, the lower the acceptable population density should be.
The analysis should also consider the specific vulnerability of such development and
popUlations, taking inte account the specific vulnerability of the people (age.ct, young,
disabled, etc.); the topography and access to and egress from the development and
locality for emergency response and evacuation; other emergency infrastructure and
design of developments (e.g., glass facades towards an explosion hazard, people in the
open air where the hazard is toxic gas).
Individual Fatality Risk Levels

Individual fatality risk is the risk of death to a person at a particular location. The
International Community has recommanded a fatality risk level of one in a million per
year (1 x 1O-<i per year) as the limit for risk acceptability for residential area exposure.
Experience with implementation indicates that the criteria is practical and appropriate,
and as such should be maintained. It is necessary, however, to account for variations in
the duration of exposure to that risk at any particular location by anyone individual. It
is also necessary to account for variations in people's vulnerability to the hazard and
their ability to take evasive action when exposed to the hazard.
The one in a million criterion assumes that r~sidents will be at their place of
residence and exposed to the risk 24 hours a day and continuously day after day for the
whole year. In practice this is not the case and this criterion is therefore conservative.
People in hospitals, children at school or old-aged people are more vulnerable to

hazards and less able to take evasive action, as compared to the average residential
population.
A lower risk than the one in a million criterion (applicable for residential land use
areas) may be more appropriate for such cases. On the other hand, land uses such as
commercial and open space do not involve continuous occupancy by the same people.
The individual's occupancy of these areas is on an intermittent basis and the people
present are generally mobile. As such, a higher level of risk (relative to the permanent
housing occupancy exposure) may be tolerated A higher level of risk still is generally
considered acceptable in industrial areas.
Accordingly, the following risk assessment criteria are suggested for the assessment
of the safety of location of a proposed development of a potentially hazardous nature of
the land use planing in the vicinity of existing hazardous installations:
• Hospitals, schools, child-care facilities and old age housing development should
not be exposed to individual fatality risk levels in excess of half in one million per
year (0.5 x 10-6 per year).
• Residential developments and places of continuous occupancy, such as hotels and
tourist resorts, should not be exposed to individual fatality risk levels in excess of
one in a million year (1 x 10-6 per year).
• Commercial developments, including offices, retail centers, warehouses with
showrooms. restaurants and entertainment centers, should not be exposed to
individual fatality risk levels in excess of five in a million per year (5 x 10-6 per
year).
• Sporting complexes and active open space areas should not be exposed to
individual fatality risk levels in excess of ten ill a million per year (10 x 10-6 per
year).
• Individual fatality risk levels for industrial sites at levels of 50 in a million per
year (50 x 10-6 per year) should be contained within the boundaries of the site where
applicable. Whilst individual fatality risk levels should include all components of
risk (i.e., fires, explosions and toxicity), there may be uncertainties in correlating
toxic concentrations to fatality risk levels. The interpretation of 'fatal' should not rely
on anyone dose-effect relationship, but involve a review of available data.
52 CHAPTER 2
10
-><
1\
10-2
~
c:
~
....... 1Q-3
is'
c:
CD
::l
~ 10-4
LL
1fr5~----~~~--~----~-----4----~
1fr6~----~----~~----+-----~----~
10~L-----~------~~~~------~----~
10 102 103 1Q4 10S 1()6
Fatalities (x)
Figure 2.6. Societal risk curves.
Injury Risk Levels

Relying entirely upon fatality risk criteria may not account for the following factors:
• Society is concerned about risk of injury as well as risk of death.
• Fatality risk levels may not entirely reflect variations in people's vulnerability to
risk. Some people may be affected at a lower level of hazard exposure than
others. It is therefore appropriate that risk criteria also be set in terms of injury,
i.e., in terms of levels of effects that may cause injury to people, but will not
necessarily cause fatality.
Effect of Heat Radiation

The ultimate effect of various heat flux (radiation) as the result of a fife incident would
depend on the duration of people's exposure to the resultant heat flux (TABLE 2.1).
TABLE 2.1. Effects of heat radiation
Heat Radiation Effect

(kW/ml)
1.2 • Radiation received from the sun at noon in summer
2.1 • Minimum to cause pain after 1 minute

4.7 • Will cause pain in 15-20 seconds and injury after 30
seconds exposure (at least second degree burns will occur)
12.6 • Significant chance of fatality for extended exposure. High

chance of injury
• Causes the temperature of wood to rise to a point where it
can be ignited by a naked flame after long exposure.
• Thin steel with insulation on the side away from the fire
may reach a thermal stress level high enough to cause
structural failure.
23.0 • Likely fatality of extended exposure and chance of fatality

for instantaneous exposure.
• Spontaneous ignition of wood after long exposure.
• Unprotected steel will reach thermal stress temperatures
which can cause failure.
• Pressure vessels needs to be relieved or failure would
occur.
35.0 • Cellulosic material will ignite within I minute exposure

• Significant chance of fatality for people exposed instanta-
neously to the fire
For the purpose of injury, a lower heat radiation level (relative to that level which
may cause fatality) is appropriate. A heat radiation level of 4.7 kW/m2 is considered
high enough to trigger the possibility of injury for people who are unable to be
evacuated or seek shelter. That level of heat radiation would cause injury after 30
54 CHAPTER 2
seconds exposure. Accordingly, a risk injury criteria of 50 in a million per year (50 x
10.6 per year) at the 4.7 kW/m2 heat flux is suggested for residential areas. International
experience with the implementation of that criteria indicates that it is achievable and
appropriate.
Effect of Explosion Overpressure

TABLE 2.2 indicates the effect of various levels of explosion overpressures resulting
from explosion scenarios.
TABLE 2.2. Effects of explosion overpressure
Explosion Effects
Overpressure
(kPa)
3.5 • 90% glass breakage

• No fatality and very low probability of injury
7.0 • Damage to internal partitions and joinery (repairs are
possible)
• Probability of injury is 10%. No fatality.
14.0 • House inhabitable and badly cracked
21.0 • Reinforced structures distort
• Storage tanks fail
• 20% chance of fatality to a person in a building
35.0 • House inhabitable
• Wagons and plant items overturned
• Threshold for eardrum damage
• 50% chance of fatality for a person in a building and 15%
chance of fatality for a person in the open air.
70.0 • Threshold for lung damage

• 100% of chance of fatality for a person in a building or in the
open air
• Complete demolition of the house.
Using a similar analysis to that adopted in establishing a heat flux injury level, it can
be suggested that an explosion overpressure level of 7 kPa be the appropriate cut-off
level above which significant effects to people and property damage may occur.
Accordingly, an injury/damage risk criteria of 50 in a million (50 x 10-6 per year) at the
7 kPa explosion overpressure level is suggested for residential areas. International
experience with such implementation confirms this level as appropriate.
Toxic Exposure Criteria

Depending on the concentration, the nature of the material, the duration and mode of
exposure (i.e., via the respiratory tract, lungs, skin or ingestion), the effects of toxicants
range from fatality, injury (e.g., damage to lungs and respiratory system, damage to
nervous system, emphysema, etc.) to irritation of eyes throat or skin through to a
nuisance effect. Effects can be classified as acute, chronic or delayed.
There are a number of assessment criteria and dose-effect relationships that vary
from one chemical to another. Toxic criteria applicable to one chemical rrlay not
necessarily be appropriate for others. International experience conclusively shows that
the formulation of a uniform specific criteria to cover all toxic effects is not appropriate
or valid. Instead, each case should be justified on its merits using a thorough search of
available and known dose-effect relationships as the basis for assessment. Incidents
with injurious impact on people should be kept to low frequencies.
The suggested injury risk criteria for toxic gas/smoke/dust exposure are as follows:
• Toxic concentrations in residential areas should not exceed a level which would
be seriously injurious to sensitive members of the community following a
relatively short period of exposure at a maximum frequency of 10 in a million
per year.
• Toxic concentrations in residential areas should not cause irritation to eyes or
throat, coughing or other acute physiological responses in sensitive members of
the community over a maximum frequency of 50 in a million per year.
Risk Levels for Property Damage and Accident Propagation

The siting of a new hazardous storag~insta1lation or transport system (raiUroad- tanker)
must account for the potential of an accident at the installation/system causing damage
to the building, and also propagating to a neighboring building (storage), or a
neighboring raiUroad-tanker and hence initiating further hazardous incidents (so-called
domino effect).
The principle of setting risk criteria to reflect the potential for accident propagation
is that the risk of an accident at one plant triggering another accident in the neighboring
should be low and that adequate safety separation distances should be provided as part
of siting and layout of plant/storage areas. A similar approach should be followed for
railway shunting yards and loading/unloading stations in the vicinity of residential or
industrial areas.
Heat radiation levels of 23 kW/m2 as the result of fire incidents may affect a
neighboring installation to the extent that unprotected steel can suffer thermal stress that
may cause structural failure. This may trigger a hazardous event unless protection
measures are adopted.
Explosion overpressure levels of 14 kPa may damage piping and (low-pressure)
equipment at a neighboring plant/rail-tanker. It is therefore appropriate that the
probability of these levels of impacts at neighboring installations be relatively low. A
probability of 50 in a million per year is suggested for assessment purposes. The explo-
sion overpressure level of 14 kPa is also sufficient to cause significant damage to
56 CHAPfER2
buildings and as such is appropriate as the basis for a risk of damage criteria A
probability of 50 in a million per year is also suggested for assessment purposes.
These criteria do not remove the need to consider higher consequence levels at
lower frequencies. The hazard analysis should consider the whole picture not just the
nominated quantitative criteria
2.2.6. FINAL ASSESSMENT OF RISK LEVELS
The qualitative and quantitative results of the risk analysis can be applied in the
assessment process as follows:
a) Risk impacts at various distances from a planed or existing storage, respectively
loading/unloading installation, pipeline routes, or selected raiVroad routes may
be compared against safety targets or criteria A judgment can be made about the
hazard impact. A general principle of assessment is that the risk impacts from the
new development should be well below the levels of risk which people and the
environment are regularly exposed to from similar development and other
sources.
b) The analysis should particularly highlight the major contributors to the risk and
their nature and extent and, secondly, areas where risk could be eliminated, or
cost-effectively reduced. These results can be used to develop prevention and
protection measures including priority allocation of resources for hazard control.
2.3. Information and Data for Assessing Transportation Risks
The requirements for transportation of goods are changing due to new forms of
industrial production methods like JIT ("just in time"), minimal storage requirements
and trading. Transports of dangerous goods, especially by road, are recognized as a
problem by some members of communities and there are groups which consider it to be
a very risky activity, especially what concerns transportation of radioactive spent fuel
elements (that are, as every body knows, transported in very heavy fuel casks designed
to withstand a drop from several meters).
CMmical 7'rtmsporllltio" Emerge"cy Ce"ter (USA)

CHEMrREC, the Chemical Transportation Emergency Center (USA), operated by the
Chemical Manufacturers Association, can be reached, in the US, under toll free number 1-
800-424-9300. CHEMTREC provides immediate advice at the scene of the emergency. In
Swilzerlatul, it is the Toxicological Information Center'/} in Ziirich, who is responsible of
giving advice in case of emergency concerning poisoning with toxic substances or gases.
0) Tolrlkologisches Informationszentrum, Notfalldiensl, Klosbachstrasse 107,8032 ZIIrich, Tel. (01)- 251-51-

51.
A correct way of looking to the transportation problem is to find out if the

transportation of dangerous goods are a problem to the extent that this could be
considered a societal problem, and in such a case to search for a convenient solution or
course of action.
2.3.1. AVAILABILITY AND COMPARABILITY OF STATISTICAL DATA
Statistical Data on Reported Accidents

Case histories of selected major incidents (for storage, pipelines, chemical plant, and
other hazardous installations) have been discussed in literature. It should be emphasized
that there are frequently discrepancies in the accounts of such accidents. The number
of deaths and injuries should be regarded as approximate. Source of information on
major incidents are the accounts in journals such as the NFPA Quarterly and Chemical
Engineering ("Chemmentator"), the reports of investigations by the Bureau of Mines
and the NTSB (National Transportation Safety Board), The Fire Protection Manual by
Vervalin6 , the report by Eisenberg et al. 7 , and reviews by Strehlow8, by Davenport9 , by
Kletz IO , by McCrindle ll , by Marshall 12 and by Slater13• See also Davies and Lees I4,1S
(1991).
In order to estimate transport risks, one has to study the interaction of the transport
process with the surrounding systems (traffic ways, population density, environment
and ecological systems, etc.). An accident can occur as a result of a collision, or due to
brake failure or a defect in the fuel system which can cause a fire. Release of material to
the environment can take place as a result of a puncture of the tank wall, or due to the
break of a valve, or the rupture of the whole tank. The probability of accidents with
dangerous goods transports on a specific route cannot be assessed directly with an
acceptable degree of accuracy because rail/road-tanker accidents art; not too frequent
and may, even, have never happened on the rail/road segment under consideration.
Interpretation and Comparability of Data. Published data on accidents and failure

rates for both road or rail transport modes are available for several countries. But there
are international discrepancies in the estimation of the expected accidents and tank
failures rates in road and rail accidents. Part of the discrepancies observed between
figures arise from different methodological approaches, such as differences in accident
definition or in data collection procedures and reporting. Usually the international
traffic systems are ~ot comparable as such, due to fundamental differences related to
roadlhighway type and construction characteristics, speed limitations, the way of
coupling rail-wagons together, or to assemble dangerous rail cargoes (passenger cars
intermixed with rail-tankers for instance). In most risk assessment studies, probabilities
are assessed after reviewing heterogeneous data and are rarely drawn from a single
specific data source or with the same consistent definitions of events likely to occur.
None of the published figures are directly comparable, because definitions for accidents
and releases are numerous. A standardization in this field is absolutely necessary.
Usually traffic conditions and composition are poorly known and it is preferable to
deduce the accident frequency from generic lorry statistics where these are available.
58 CHAYfER2
Accidents involving "thin walled" tanker releases are, in some countries, high enough
for direct estimation of the conditional release probability. Looking at "thick walled"
tankers, releases are scarce, and conditional release probabilities can be determined by
applying a deterministic failure analysis to a statistical description of accident
environment stresses (Clarke, et al}6 1976).
True Accident Rates. Recently methods have emerged for assessing the true accident
rate, i.e., the anticipated number of accidents after a certain number of accidents have
occurred over a given period. One of these methods is the ''Empirical Bayes Method"
and the other make use of Poisson distributions, which means that the expected values
for accidents in a population of intersections or road segments of a similar type are
gamma distributed. To use these methods, a number of accidents that is not too small
must have occurred for every type of accident under consideration. This again may not
be the case with road-tankers, for which very few accidents have been registered.
Methods for estimating accident frequencies and comparing alternative routes shall be
presented in the next chapters.
Swiss Accident Statistics. There is generally a problem with the availability and
appropriate reporting of statistical data on dangerous goods' accidents, this especially
in Switzerland, where for instance the use of computer in public administrations is quite
recent. The problem is even enhanced due to the Swiss federalist system according a
large independence to the Swiss Cantons, which all have a different way of reporting
accidents and preparing statistics. Mostly in the past, there was no preferential diffe-
rentiation between the different kinds of traffic accidents, and it is actually still difficult
to know the exact number of accidents involving road-tankers or truck transports of
other dangerous goods. Usually, accidents are registered and reported by the Municipal
or Cantonal Police (urban/highway traffic) using different protocols and computer
software/systems. There is no Federal Accident Database available for Switzerland.
Reporting of Accidents. Most of the time, in case of material damages involving cars
and trucks only, the participants do announce the damage to their respective insurance
and the police is not involved or informed at all about that accident. When people are
injured, the police must be called and will register the accident. As long no material has
been released to the environment, the police will usually not record whether the truck
was empty, or transporting earth, construction material or some dangerous goods, nor
record what kind of product was involved in the accident. There is also another problem
with historical accident data that is due to traffic intensity and road communications
changing over the years, and thus the statistical data may not always reflect actual
traffic conditions. Therefore accident rates cannot always be satisfactorily extrapolated
on the basis of historical data (Abkowitz,17 1988).
Road Classification. A unified classification of roads and highways as a function of

environmental characteristics and according to construction type, such as lane sepa-
ration, road inclination, number of turns, speed limitation, and other traffic signs, the
availability of fences on central reservation or on road side, the presence of gutters
preventing the spreading of water pollutants is absolutely necessary. When actual

accident rate data are missing, such a classification scheme associated with appropriate
correction factors would allow to use accident frequencies of comparable road segments
for risk assessment studies and to modify them from case to case.
Improvement to this situation shall be brought to, when a new Swiss Federal data
base system, STRADA-DB ("Datenbank fUr SRAssen-DAten "), will be fully implemen-
ted. IS (see Chapter 10 for a description of STRADA-DB).
2.3.2. COMPARING ALTERNATIVE ROUTES VS. COMPARING ALTERNATIVE

TRANSPORTATION SYSTEMS
The performance of transport systems in a region and their capability to perform

satisfactorily are of great importance to the user respectively community. Although the
number and design of roads cannot be influenced by the transport system itself, the
latter can exert an influence on the traffic system and indirectly on the planning and
development of road/railway systems. What distinguishes dangerous goods from
ordinary goods is that in the event of an accident during the transfer of dangerous
goods, acute injury and damage to people and surroundings can result.
Comparing Alternative Routes

The problem of choosing alternative routes, in Switzerland, is a difficult problem due to
the geographical situation of the country which is strongly partitioned and does not
offer many alternatives for both rail and road traffic: The "Jura" mountain range from
Geneva to Basle, the land-plateau parallel to the Jura, and the "Alps" mountains with
their associated valleys, like the "Rhone valley", the "Graubunden region", or the
"Gotthard region". There are two main transit axis, North-South (from Basle to Chia3s0
through the Gothard tunnel) a!ld East-West (St. Margrethen - Zurich - Bern - Lausanne-
Geneva). For rail transit we have two North-South axis, i.e., the "Gothard tunnel"
connecting Goschenen to Airolo - Chiasso, and the "Lotschberg tunnel" connecting the
Rhone valley to Bern and the Swiss Plateau, and in the same N-S-axis the "Simplon
tunnel" connecting Brig to Domodossola - Milan.
Other constraints play also a role. For instance the goods imported to Switzerland
through Basle arrive either by rail, road or ship, depending on the nature of the
products. From this fixed point, the goods which were shipped must be transshipped to
road or rail transport systems. Petroleum products are also stored in large storage tank
areas (National strategic reserve). Truck movements from and to the Rhine harbor have
to take place through dedicated roads since no alternative :outing is available. It must be
kept in mind that alternative roads, which cannot be differentiated as regards risks on a
yearly basis, can very· well show differences in possible consequences at different
period of the year and time of the day.
Among the general criteria used for routing, we find:
• risk criterion (environment and population should be at minimum risk)
• worst-case consequence criterion (the most severe consequences of a single
accident should be as limited as possible)
60 CHAPfER2
• selective object criterion (selected objects like ecological systems, etc., should
not be damaged or harmed by water soluble or other substances).
Dynamic rerouting would be an alternative for managing the risk of dangerous
goods transports. Efforts in this direction are in progress, especially in the United
States. The method presupposes the availability of dynamic traffic control systems
based on the use of Global Positioning System (GPS) located on each truck, which
supply the exact position of vehicles and a radio-communication system which transmit
the data to the traffic manager. Such solutions are costly and no legislation, at present,
can oblige any trucker company to equip its vehicles with such systems. The congestion
of traffic is so severe today that it would be useful to take corrective measures. One task
for a traffic control system will be to distribute the demand for road resources in time
and to optimize the traffic flow (private cars, trucks and public transports) and to select
appropriate roads dynamically, passing the information to the road-user equipped with
an automatic display, to redistribute the traffic by proposing alternative roads. Another
solution which has already been adopted in several countries consists in controlling the
traffic flow by modifying the allowable speed on highways to avoid traffic queues.
Irrespective of which risk criterion is selected, methods of implementation are
always needed so that the routing will have the desired effect of reducing transportation
risks. In addition to a method for determining which roads present a lower risk, a way
of informing the drivers in time (i.e., in advance) of the presence of obstacles, road
repairs, deviations, and traffic queues has to be implemented. At present the technical
installations and the necessary legislation are missing. Actually information for road-
users is broadcast on special radio frequencies. Some car-radios allow the current radio
emission to be automatically interrupted for receiving special traffic messages. The
main dr3w-back of this system is that information is usually not quite up-to-date or
reaches the driver much too late, when this one has already been stopped by a queue on
a highway.
Road vs. Rail Transportation

The goal is to compare alternative (or combined) transportation systems in order to see
if one routing alternative and/or the choice of another transport medium could possibly
reduce the transportation risks. General guidelines are given in Road Transport of
Hazardous Chemicals by the CIA (1973, Item No.4). At the international level, the UN
transport code untitled The Transport of Dangerous Goods, published by the United
Committee of Experts (1977), constitutes an international code of practice which is
applicable to road transport. The international transport of hazardous chemicals in EEC
countries is carried out under the terms of the European agreement concerning the
international carriage of dangerous goods by road, ADR.
The comparative hazard assessment can be conducted in considering accident
scenarios. The effects of the accident may then be determined. The calculation may be
relatively simple, such as the determination of the concentration of a chemical spilled
into a lake, or more complex, such as the estimation of the concentration profile of a gas
cloud caused by a spillage. The assessment may include the probabilities for particular
accidents to take place and the usefulness of countermeasures. Transport by road are
often compared to rail transport on the basis of the number of fatalities per 1000 t, as
shown in the example of Figure 2.7.
Environmental data (weather categories, ecosystems, forests, urbanlland population)
and other roadway elements, such as the number of at grade road/rail crossings, number
of turns, type of road/highway, etc., have to be taken under considemtion as well and
risk factors and accident scenarios developed. The corresponding traffic, accident and
release data are compiled and corrected to reflect the conditions of selected road
segments. The population density along the transport corridors must also be estimated.
The dangerous goods are usually classified according to danger classes, the quantities
transported must be known or estimated and the transport risks must be calculated for
each danger class and each road or rail segment.
EXIlmJIle of a Risk Comparison Between Rail and Road Transport

(The exmnple applies to US traffic conditions)
Using published U.S. datafor accident rates per vehicle-mile, it is possible to calculate the
risk involved to the truck or rail drivers in transporting 1()(}() ton ofproduct over 200 miles
by road or rail:
• lf road transport is used, then assuming a load of 15 ton and an average speed of 30
mileslh, 67 loads have to be transported for 6.7 h each, making a total of 450 vehicle-
hours. The fatality rate for driving a vehicle is 57 x 10.[1Ivhc.hr, so that:
frequency offatality = 450 x 57 x 10.[1
= 2.5 x 10-4 fatalities/l ()(}() ton transported.
• lf rail transport is used, then assuming a load of23 ton and an average speed of 30
mileslh, 42 railcars have to be on the move for 6.7 h each, and assuming 42 railcars/train
anda train crew of 2 men, the exposure time is 6.7 x 2 = 13 h. Taking the same fatality rate
as above we find for the train with two drivers:
frequency offatality = 13 x 57 x 10·B
= 7.5 x 10~ fatalities/1OOO ton transported
On this basis the fatality risk for road transport is, in this exmnple, 33 times that for rail
transport.
Figure 2.7. Risk comparison between rail and road transportation.
One as to differentiate between the risk of an accident and the risk of a spill, fire or
explosion. Risk mitigation measures and the overall system sensitivity to these
measures have to be evaluated. To establish optimum routing conditions, two types of
modeling conventions may be used:
• Minimum time travel, where trucks are routed on the basis of least time from one
point to another. This inherently minimizes risks as it reduces the time exposure
of a vehicle on a roadway.
• Minimum accident probability, where trucks are routed on the basis of least
accident probability from one point to another.
Depending on the region being assessed, alternative routing could be found possible
or the advantage of one transport medium against another become more apparent.
Redistribution of some traffic may be considered in order to reduce the risk to popu-
62 CHAYI'ER2
lation or environment. The recommended routing strategy should adopt a minimum

accident probability convention. In the future it would be a good practice that results of
transportation risk assessment studies shall be used to inform the Community of pro-
posed implementations of risk mitigation programs and the cost thereof, and to have the
Community decide and agree on the risk level that she is willing to accept.
Comparative studies for chlorine and other dangerous goods transportation have
been reported and summarized further in this section.
2.4. ClassifICation of Regional Risk and Order of Priority
Large industrial areas accommodate an extensive number of risk sources and activities
of varying nature and extent. Such sources may include operating process plants,
storage terminal, transportation activities, etc. Ideally, a cumulative assessment of such
risk should include a detailed hazard analysis and quantified risk assessment for all
industrial facilities and associated activities. In many cases, however, because of limited
resources and time constraints, a preliminary evaluation of the various risks is needed,
in order to establish which activities should be the focus of detailed risk assessment in
priority to others and where ~sessment resources should be allocated in the framework
of regional risk assessment and land use planing. Method for classifying risks in a
region (which also can be applied to the risk of transportation, handling and storage of
dangerous products) have been proposed in an Inter-Agency program on the
"Assessment and Management of Health and Environmental Risks from Energy and
Other Complex Industrial Systems",19 The program in question was sponsored by
UNEP (United Nations Environment Programme); WHO (World Health Organization);
IAEA (lntematiomd Atomic Energy Agency), and UNIDO (United Nations Industrial
Development Organization).
Classification of Risk at a Regional Level and Setting of Priorilks

The method is based on the classification of hazardous activities in a region by way of
categorizing consequences and probabilities of occurrence of major accidents. It provides a
preliminary generalized quantitative overview of the different rislcs in a large industrial
area, based on the concept of societal risk, and enables the setting of priorities for further
analyzjng the different sources of risk.
The categorization of the consequences leads the user to approximately calculate

the number of fatalities caused by an accident in a fixed installation (storage area) or
during the transportation of hazardous materials by road, ship, or pipeline. The
estimation of probabilities yield information on the frequency of accidents. The results
of the analysis may be presented in a graphical form on an x-y system of coordinates,
where the x-axis shows the classes of consequences and the y-axis the probability
classes. Therefore, all hazardous activities in the area (including storage and transport-
ation) can be classified and presented in a matrix format. Once criteria for the
acceptability of societal risk have been established, the user can identify on the matrix
all the activities which do not satisfy the requirements.
Geographicallnfol7llQtion System (GIS)

A GIS system allows for the analysis of complex data associated with mapped information.
Map features of different types are stored within the computer system along with des-
criptive or numerical data. Both the graphic and the textual data can then be analyzed and
displayed using a variety of techniques.
Using a graphic information system it is possible to represent different kinds of risks

for a whole region. A method pertaining to risk representation has been developed by
Simoni,20 (1995), in the framework of the PPR&S.
An overview of the main tasks for risk classification and prioritization scheme is given
below:
• Classification of types of activities and inventories, selected rail-, road-,
waterways-, pipelines-segments presenting a risk.
• Estimation of consequences
• Estimation of probabilities for fixed installations and/or for transportation.
• Estimation of societal risks, and finally
• Prioritization of risk areas requiring detailed risk assessment studies
The estimation of probabilities of major accidents for fixed installation (storage,
loading/unloading area) is based on estimating the frequency of major accidents for
each of the activity under analysis, by deriving an average (standard) probability
number (defined as the absolute value of the logarithm of the yearly frequency of
occurrence of the accident in that "standard activity") and adding to it few probability
number correction parameters. These parameters account for: the frequency of loading/-
unloading operations; the safety systems associated with flammable substances;
.,rganizational and safety management; and the prob&bility of wind direction towards
populated areas in the affected zone.
The estimation of probabilities, resulting from transportation of hazardous ma-
terials, is based on selecting an average (standard) probability number for each portion
of road, railway, waterway, or pipeline under analysis, and the application to this
standard probability number of a few correction parameters. These parameters reflect:
the safety conditions of the transport system; the traffic density; and the probability of
wind direction towards populated areas in the affected zone.
The selection of the activity or road/rail segment to be considered for analysis is
based, in this method, on a distance criterion (distance from populated areas), and on a
traffic density criterion, both tabulated in the original UN-Interagency report First, the
activities are classified using special checklists mentioning the type of activity and the
substance handled, for which a reference is made to categorization tables. In these
tables the classification of substances is based on effect categories. The effect categories
themselves are related to the quantity of substance involved in a potential accident of a
given type. The maximum distance and area of effect can be deduced for each category
class. The external consequences (Ca." number of fatalities / accidents) of major
accidents can be calculated using the formula:
64 CHAPTER 2
(2.9)
A affected area in hectare (one ha corresponds to a surface of 104 m2 )

p population density (personlha)
fa area correction factor for the distribution of population in the affected
zone
fm correction factor for mitigation effects
• Step 1: For each dangerous good transported: select one route (road, railway, waterway,
pipeline); select segments of 1 kIn in length; choose the most hazardous segment out of
the list (high population density, high traffic load and bad traffic safety).
• Step 2: Select from the table the average probability number for each hazardous
substance (or group of substance) and for each road segment under consideration.
• Step 3: Estimate the probability number correction parameter 1\: taking into account the
safety conditions of the transport system.
• Step 4: Estimate the probability number correction parameter lit taking into account the
traffic density (number of transport units per year).
• Step 5: Estimate the probability number correction parameter IIp taking into account the
wind direction
• Step 6: Calculate the probability number Nt,s using the above formula.
• Step 7: Convert the probability number into a probability, P.
• Step 8: If a portion of a roadlraillwaterway/pipeline is exposed to a risk: of accident
from different hazardous substar.::es, the frequencies (accident frequency per kIn and per
year) calculated for each of them have to be grouped under the defmed classes of injuries.
• Step 9: Repeat all the steps for each road segment and each substance transported on
that road segment.
Figure 2.B. Procedure for estimating the probabilities of major accidents resulting from transportation
of hazardous materials.
The probabilities of major accidents for storage or fixed installations can be
estimated according to the formula:
(2.10)
Nj,s Average probability number for the installation i and the substance s
nf Probability number correction parameter for the frequency of loading!
unloading opemtions
nl Probability number correction parameter for the safety systems associated
with flammable substances
no Probability number correction parameter for the organization and
management safety
np Probability number correction parameter for wind direction towards the
populated area.
In order to calculate the frequency (Pi,., number of accidents per year) of accidents
during transportation, the so-called probability number should first be estimated:
(2.11)
Nt. Probability number

ttt.. Average probability number for the transportation of the substance s
nc Probability number correction parameter for the traffic density
nt Probability number correction parameter for the traffic density
np probability number correction parameter for wind direction towards the
populated area.
The probability number can then be converted into a probability from the
relationship between N and P:
(2.12)
The necessary data, tables and correction factors for this method can be found in the
afore-mentioned UN-Interagency report (1993).
The procedural steps for estimating the probabilities of major accidents resulting
from transportation of hazardous materials is given in Figure 2.B.
Fixed Installations. A similar approach is taken for fixed installations. The estimation
of the societal risk is done according to the procedure of Figure 2.9.
• Step A: Classify each activity using a scale of consequence classes (0-25, 26-50, S1-l(lO,
101-250, 2S1-5OO, >SOO fatalities/accidents) and a scale of probability classes (frequency =
10-3 to 10-7 accident per year)
• Step B: If a certain industrial activity presents risks to the population from different
substances which can cause accidents independently from each other, sum up the risk from
substances which have the same class of consequences.
• Step C: For risk classification, prepare a matrix of the activities showing frequency vs.
consequences.
Figu.re 2.9. Procedure for estimating the probabilities of major accidents resulting from fixed installations.
2.S. Hazards Resulting from the Transport of Dangerous Goods
Hazardous materials are moved by the foUowing modes of transport: road, rail,
waterway, pipelines, ship, aircraft. The type of material transport used may greatly
affect the storage requirements. A mode of transport which involves large but
infrequent deliveries is in general likely to require more storage than one which gives
continuous flow. Actually the trend for delivering electrical or mechanical components
or chemical products to the industry is to use the approach called ''just in time delivery".
This way of distributing goods is, unfortunately, requiring a much larger number of
66 CHAPTER 2
transports, usually taking place by road-trucks. The associated storage requirements

have corresponding economic and safety implications.
Transport nwy present a hazard to: (a) the transporter. (b) the crew. (c) the public.
The relative importance of these varies between the different nwdes of transport.
This section reports some analyses of transport accidents, and describes the main
factors contributing to accidents.
2.5.1. TRANSPORT OF CHEMICALS
Transportation H(J1Jlrtis
The hazard presented by the transport of chemicals are:
• Fire
• Explosion
• Toxic release of conventional toxic substances
• Toxic release of ultratoxic substances
Road accidents leading to the release of hazardous nwterials are due in most cases to a
tanker crashing or overturning.
Big road disasters are for instance: San Carlos de la Capita, Spain (1978), in which
a propylene tanker crashed into a holiday camp site, cost the lives of some 211 people.
The explosion at St. Amand-Ies Eaux, Frct.'lce (1973), and the toxic release on the
Southwest Freeway, Houston, Texas (1976), etc.
Rail accidents in which there is an escape of hazardous material are mostly due to
train crash, or derailment. Mechanical features of rolling stock and faults in the rail
track are frequent causes of these. Another hazard of railway is fires along the track
caused, in the past, by steam engine or break shoe sparks.
HatIlTfl of CMmical Transport

• Fire is nwre likely than explosion where there is a loss of containment of a flo.nunDble
from a rail. road, barge. or shiptank or from a pipeline.
• But both unconfined vapor cloud explosions and boiling liquid expanding vapor
explosions occur as a result of transport accidents.
• In addition an explosion hazard exists with substances which are to some degree unstable
(some chemicals and explosives).
• Loss of containment of conventional toxic substances from a tank or a pipeline nwy give
rise to a large toxic gas cloud or may pollute water supplies.
Figure 2.10. Transportation hazards.

Since the Flixborough disaster in 1974, Unconfined Vapor Cloud Explosions

(UVCE) have received much attention in the industrial, scientific, governmental, and
insurance fields. The state-of the-art up to 1978 has been treated by Gugan21 , who also
included a listing of known incidents up to that time. Further listing of UVCE can be
found in the papers of Strehlow/2 (1972), and Davenport. 9 Of 71 incidents up to the
year 1978,72% occurred within hydrocarbon processing plants, such as chemical plants
and refineries; 23% were rail, truck or pipeline accidents; and the remaining 5%
occurred in other places.
There is little published information on the extent to which ultratoxic substances are
transported. In view of the hazard of toxic contamination which may arise from even a
small spillage of such a substance, such transport is likely to be the subject of special
arrangements.
Events which can give rise to hazards include particularly:
• Containment failure
• Accident impact
• Loading and unloading operations
Failure Data and Failure Data Analysis

The application of reliability techniques creates a demand for data on equipment failure
and repair times, on other failure related events and on human errors. These data may be
obtained from the literature, from the data banks or within the works. The determination
of accurate data tends to involve much effort. The accuracy required in the data varies
considerably between different types of reliability calculation and even between
different parts of the same calculation. In a fault tree analysis, for example, some
branches of the tree may be sensitive to the failure rates used, while others may not be.
Often it is sufficient to lruow that the failure rate lies between certain broad lil. .1its for
the solution to the reliability problem to be clear (use also expert judgment).
Failure Distributions. We distinguish between the following types of failure distribu-

tions
T ABlE 2.3. Types of failure distribution used in engineering and economic applications
Binomial distribution Weibull distribution

Multinomial distribution Rectangular distribution
Poisson distribution Gamma distribution
Exponential distribution Pareto distribution
lognormal distribution Extreme value distribution
It is emphasized that there are many factors which determine the failure rate of an
equipment and that the range of failure rates observed for an equipment can be quite
wide. Two prime sources of failure rate data for fixed installations are the pUblication of
the UKAEA (U.K. Atomic Energy Authority), including the SRS (System reliability
Service) and the Reactor Safety Study (1975) by the AEC (Atomic Energy Commis-
sion, USA).
68 CHAPTER 2
Data Related to Transportation

Data on transport have been compiled in 1974 by Westbrook. 23 Some data on the
relative frequency of ignition of leaks are given by Browning. 24 Information on failure
and event data can also be found in the Canvey survey.
Miscellaneous Data and Information from the Canvey Report

The most comprehensive hazard assessment of non-nuclear installations in the U.K. is
the Convey study carried out for the HSE (Health and Safety Executive) in 1981 by the
UKAEA Systems Reliability Directorate (SRD), during the preparation of the Second
Canvey Report. The overall approach taken was among others the identification of the
principal hazards of the installations and activities in the Canvey Island trhurrock area
and th~ assessment of the associated risks to society and to individuals.
This kind of investigation requires the estimation of the probabilities of various
occurrences and of their consequences. Some data taken from the Canvey study are
listed in Figure 2.1 I.
Spontaneous Failure of Pressure Vessels. Among the scenarios considered, this is a

possible initiating event for release of LPG, ammonia and HF. Failure of pressure
piping was also considered for LPG and LNG. If a large pipe under pressure (15-25 cm
diameter) were to fail, there would be a vapor plume, the plume would find a source of
ignition and would bum back to the pipe and there form a burning jet (up to 50 m
length) or a burning sphere (10 m diameter). In burning back there would be the
possibility of semi-confmed vapor cloud explosions which could damage other
equipment such as LPG storage spheres.
Failure of Methane Pipelines. A serious failure of a pipt"line (35 cm diameter) would

lead to a large release of methane, but the gas would be highly buoyant and would
undergo turbulent mixing with air, so that there is no chance of the formation of a large
vapor cloud heavier than air. The frequency for such an event is less than 1O·4/year.
Spillage and Evaporation of LNG and Ammonia on Water. For a rapid spill of LNG on
water the evaporation time is short and a puff dispersion model is applicable. But for a
continuous spill the evaporation rate rapidly reaches the spill rate and this evaporation
rate then determines the pool diameter reached. The evaporation rate has been estimated
to be about 0.19 kg/m2 s, which corresponds to a liquid regression rate of 4.7xlO-4m/s.
Similarly, it is stated that for a rapid spill of ammonia on water the evaporation time is
short and that all the ammonia will evaporate, except for about 20% which dissolves in
water.
A cloud of cold LNG tends to behave like as a heavy gas cloud. The model of the
source which is assumed can have significant effects on the results. The release may be
continuous or instantaneous. The source may be of fixed size or increasing size, as with
LNG spilled on to water. The process of release mayor may not involve the
entrainment of large amounts of air. The description of the source elements must be
done precisely in order to model correctly the scenario.
Canny struly
Potentially haz/lrdous material in the Canvey Island IThurrock areas, their location and the
quantities stored and in process were identified. The investigation identified several
principal haz/lrds in the area. These are:
I) oil spillage over bund
2) UfG vapor cloud release (l,OOO t)
3) [PG vapor cloud release (1,000 t)
4) ammonium nitrate explosion (4,500 t of 92% solution)
5) ammonia vapor cloud release (1,000 t)
6) hydrogen fluoride cloud release (100 t)
The following cases were considered in the Canvey study:
• a severe fire might occur if there is an escape offlammable liquids from storage.
• a severe vapor cloud fire and/or explosion might occur if there is a spillage of LNG or
LPG so that a vapor cloud forms and ignites.
• a severe explosion might occur if there is a rupture of an ammonium nitra~1! storage tank.
• a severe toxic release might occur if there is a spillage of ammonia or a rupture of
storage or process plant containing hydrogen fluoride.
Canvey Study: Probabilities of Various Occurrences and oftlwir Consequences
The frequencies indicated below are for:
• Spontaneous failure ofpressure vessel (LPG, ammonia, HF): 1(J5_10-4/year
• Failure of LPG pipework (whole refinery installation): 5 x 1(J3/year
• Large vapor release (LPG filling point): 5 x 10·3/year
• Serious fatigue failure for a UfG tank (above ground): 2 x 1rr/year
• Overpressurization by overfilling for a LNG tank (above ground): 10·s _10-4/year
• Rollover involving structural damage for a LNG tank (above ground): 1(Js -1 rr /year
• Unconfined vapor cloud explosion in a refinery: 10.3/year
• Major fire in refinery: O.I/year
• Failure of butane pipeline (15-20 cm dia.): 3 x lO-4/km.year
• Derailment of rail tank car: 1 x 1O.fi!train km traveled. Probability of overturning given
a derailment: 0.2
• Frequency of accident of road tanker involving spillage: 1.6 x 10-8/Ian traveled
Figure 2.11. Principal hazards considered by the Canvey study.

For safety analysis, the question arises whether it is safer to handle a given quantity
of hazardous material in a few large units or in many small ones. Many of the hazards
associated with transport arise during loading or unloading at terminal. The use of the
same container to carry different chemicals is quite normal and it is fairly common
practice with road tankers, barges, ships and pipelines.
There is a certain amount of data available on incident rates for the various modes
of transport, but this should be interpreted with care, due to technical and administrative
70 CHAPTER 2
improvements during the last decade. Assessment of the comparative safety of the
different modes of transport may consider only the risk to the public or they may take
into account also the risk to terminal operators and to transport crews such as drivers.
The latter may be injured or killed in road accidents and the inclusion of such accidents
make a significant difference to safety calculations for road transport. There is however
no transport crew for a pipeline.
2.5.2. FREQUENCY ANALYSIS
Three main approaches are used for transportation frequency analysis:

a) use of simple historical failure rates for all accidents lumped together for a class
of commodities or containers.
b) use of historical failure data for the specific containers typically used for the
commodity class (or container class) for individual accident scenarios or classes
of accident scenarios.
c) use of cumulative probability distributions of force magnitude, along with
engineering models to determine failure thresholds for individual accident
scenarios.
A frequency analysis consists of determining not only the frequency of the accident
initiator (i.e., train or truck accident) but also other factors directly affecting the release
scenario (e.g., probability of leakage, fire, explosion), given an accident.
Leimkiihlec5 et al. (1961) performed a statistical analysis of the frequency and
severity of truck accidents. Regression analysis were used to determine accident
frequency as a function of cargo type and season of the year. Cumui~.tive probability
functions were cl:!termined for the net impact velocity of collisions with other trucks and
with automobiles, each at different angles of impact, and for overturns and other
collisions combined. An updated analysis (Leimkiihle~, 1963) included a correlation
of cargo damage with transport vehicle damage for trucks transporting new automobiles
and bulk petroleum products. Vehicle damage was correlated to highway type and
accident type. Risk reduction investigation included controls on truck speed and
rerouting to avoid congested two-lane roads.
Appropriate statistical methods (Chira-Chavala27 , 1991; Hu 28 , 1989)) must be
applied to historical data (national statistics) extracted from governmental databases in
order to obtain both accident and travel data for the specific conditions/routes desired.
When examining, the analyst must distinguish between: (a) truck and car data; (b)
different types of truck; (c) the accident rate and the vehicle involvement rate; and (d)
the estimated reporting efficiencies for different type of accidents (e.g., fatal vs. pro-
perty damage only).
A second option is to access state databases for the specific routes. A statistical test
must then be performed with specific values to determine if the calculated value should
be replaced because the sample size is too small (Harwood and Russell29 , 1990). The
expected value could be used if data were missing for a highway segment. The expected
value can be chosen according to the following method, which consists in using one of
the existing analyses of the state databases and to apply the result to the specific routes
of interest.
Train accident rates are more straightforward than those for trucks because there is
basically one source of accident data for a train and one source of exposure data. In the
USA, there exists a Railroad Accidentllncident Reporting System (RAIRS), which is a
database of accident reports (Abkowitz and List30 , 1986) submitted by the railroads to
the Federal Railroad Administration (FRA). The Association of American Railroads
(AAR) also maintains a database (TRAIN II) on railcar movements. The FRA Office of
Safety (OS) publishes an annual accident/incident report, entitled Accidentllncident and
Inventory Bulletin.
The AAR, the Railway Progress Institute (RPI), and the Chemical Manufacturers
Association (CMA), in the United States, are jointly developing a hazardous materials
transportation risk analysis mode1. 31 The accident cause portion of the model will
include the effects of accident type (collision, or non-collision), train speed, track class,
railroad type, train length, car placement within the train, tank car type, and car
protective features.
2.5.3. FAILURE FREQUENCIES FOR RAILROAD TRANSPORT
In the case of railroad transport, many countries have negligible history of serious
accidents to liquefied gas rail tank cars, and reliable statistics directly assessed from
these are usually not available. However one may find a large amount of data on
collisions and derailments and the type of location. It is from this information that a
series of rail tank car failures rates has to be assessed for a variety of locations.
2.5.4. BULK TRANSPORT MODES
The four principal modes for bulk transport are highway, railroad, pipeline, and
waterborne. In the case of highway transport, single loads up to 10-25 tons move under
various constraints of time or route. With railroad transport, the routes are more
restricted, and many other operating constraints can exist. Unit loads can reach up 100
tons and, in Europe, full train loads of up to 1500 tons can be carried. In North
America, mixed train loads with several thousand tons of hazardous chemicals are not
uncommon. Water transport varies from small unit loads, often on barges moving in
canal systems, up to bulk carriers moving loads in one or more large refrigerated tanks
over the seas.
In the first three modes of transportation, the presence at any location of the
hazardous material in tanks is usually of a transient nature, as the vessel or vehicle
passes. For most of the time it will not be present. There are some locations, however,
where it may be present for extended periods of time, such as during loading or
unloading operations, or when it is stored in large storage yards, locks, or vehicle parks.
In contrast, the material will always be present in a transport pipeline at location
through which it passes. In assessing the failure frequencies at each location, due
account of these factors must be taken.
72 CHAPTER 2
The U.K. Data Base and Risk Methodology Approach

Details of the methodology developed for the analysis of the risks arising from the
carriage, in bulk, of toxic and flammable substances by road and rail in Great Britain
has been reported by Purdy,32 (1993). The novel aspects of consequence and human
impact modeling are considered. The relative risks of transporting hazardous materials
by road or rail are explored and it is shown that the inclusion of motorist and rail
passenger populations significantly affects the calculated risk levels.
In Great Britain there has been few transportation incidents with casualties and the
British data base on accidents cannot be used for frequency analysis. For this reason a
different approach was followed (Figure 2.12).
Metlwds for Deriving Spin Frequencies

• One approach (A) would be to use an event tree similar to that developed by Hubert et
al. 33 from French data. In such a method the values given to the critical probability
associated with "escalation" to puncture is critical, yet very uncertain.
• The other approach (B) to deriving appropriate spill frequencies is to generate rates by
statistical techniques from an "accident free" history, providing a useful upper bound
check. Last method was used for deriving spill frequencies in Great Britain.
Figure 2.12. Method for estimating spill frequences.
Although data on US rail incidents are easily available, it was felt the differences
between the design standards and operating practice made this data inapplicable to the
British situation. However US road dnta could be used, and by appropriate modification
to exclude those events which could not or were unlikely to occur in Great Britain, spill
frequencies were derived. Fault tree analysis was used to develop the possible cause of
events which could lead to ~uipment leaks. These were then used to derive approp;·iate
equipment spill frequencies for both rail and road transport of LPG, ammonia and
chlorine. For flammable spill events it is also necessary to consider the probability that
spill will then be ignited and whether this will take place initially, or at some later time
once a flammable cloud has developed. In the paper of Purdy,32 tables of frequencies of
spill against cause, substance, and transport mode, as well as ignition probabilities for
flammable substances are given for conditions prevailing in Great Britain.
Regarding consequence analysis, there are several constraints which act to limit the
range of possible events:
• Simultaneous loss of content for multi-compartment tankers is extremely unlikely.
• Small release of flammable material are unlikely to ignite.
• Above a certain hole size, either the release of pressurize, liquefied gas will be so
rapid that it can be considered equivalent to an instantaneous release on vessel rupture,
or the hole will lead to a propagating failure of the pressure vessel.
• In the rail environment ignited jets of LPG are unlikely to create significant hazards,
unless they impinge on other LPG tankers which then BLEVE. The BLEVE frequency
used should normally include such a cause.
Analysis of Rail and Road Transport Risks (U.K. Case)

For the "UK transport study" three failure cases were used: these were vessel rupture, a
large hole and a nominal equipment leak (Figure 2.13).
• It was assumed that 10% of the release from pressure vessels were instantaneous and
could be modeled as the entire loss of contents.
• In the case of toxic materials the cloud was assumed to contain 100% of the tanker
contents
• For LPG, twice the adiabatic flash fraction was assumed to enter the vapor cloud.
• Containment effects of roads and drains can significantly affect the shape and
dimensions of the hazard zone (especially for flammable liquids).
For motor spirit two cases were considered:
a) tanker remaining on rail or road (confined by kerbs, drains, etc.), and
b) tanker left the road or rail track (unconfined case to be modeled as a circular pool).
Figure 2.13. Failure cases used for the analysis of rail and road transport risks in Great Britain.
Spreading pool expressions can be coupled with a "drain model" and a ''fire model"
to estimate the maximum area affected:
• The immediate ignition case were calculated assuming that the release occurred
over a finite time and that the pool size was the maximum possible after
regression (regression rate is equal to the spill rate).
• In the delayed case, the pool was allowed to spread to its maximum before
iL~tion took place.
• The possibility of soft BLEVE fire ball of a gasoline tanker does not seem likely, since
the it will at least take an hour of prolonged heating before a rupture occurs with only
two relief valves (out of four) working. For LPG the type and extent of hazard depends
on the mode of release, and whether and when it is ignited. BLEVE are much less likely
on the road.
• Vapor Cloud Explosions (VCE) are very unlikely on both road and rail. The risk and
consequences can be estimated using a simple consequence model such as TNT.
Outside the flammable cloud, the probability of death due to the effects of overpressure
from a VCE is low.
• The risk from released toxic gases, such as chlorine and ammonia, is very dependent
on the accuracy of the dispersion modeling. When calculating the societal risk, the
crosswind extent of the cloud is as important as the downwind hazard range.
In 1985, the Health and Safety Commission (HSC), in Great Britain, has examined
the potential risks of the transport of certain dangerous substances and the results were
presented in a HSC report34 in 1991. The terms of reference covered not only the
consideration of major hazard aspects of the transport of dangerous substances but also
the identification of appropriate control measures and advice on additional action that
74 CHAPfER2
might be necessary. They did not include radioactive substances, transport by air or by
pipelines or risk to the environment.
Transport in Bulk, of Toxic and Flammable Substances by Road and RaU in Great
Britain
It is concluded, that for Great Britain, the safe routing of materials with large hazard
ranges may be more easily achieved by road. While the natural separation afforded by the
rail system may make this mode more suitable for lower hazard materials.
It was recognized that existing accident statistics alone provide an inadequate basis
on which effectively to judge the risks involved in rare but potentially high
consequence events. It was concluded that a more appropriate approach is to use the
quantified risk assessment (QRA) methodology. The report acknowledges that the risks
estimated by this method are subject to some uncertainty, but judges that QRA has
provided the best estimates of the risks involved, provided valuable insights and
assisted the Sub-Committee in reaching its conclusions.
QRA can only provide numerical estimates of the risks; it cannot, by itself, provide
conclusions as to whether or not those risks are low enough or too high. Such judgment
are reserved to the decision-maker.
2.6. Legislation Concerning Transportation of Dangerous Goods by Rail, Road,

and Air
2.6.1. INTERNATIONAL REGULATIONS
The international transport of hazardous materials by road or rail, is the concern of

several bodies of European and world level:
• The UN Transport Code, 1977 (code of the United Nations for the transport of
dangerous goods), constitutes an international code of practices. This code gives a
classification of hazardous substances.
• The ADR ordinance, "Accord Europeen Relatif au Transport des Marchandises
Dangereuses par Route" (ADR) is regulating the international transport of hazardous
materials by road. 35
• The RID regulation "Reglement International Concernant le Transport des
Marchandises Dangereuses par Chemin de Fer (RID)" is regulating the international
transport of hazardous materials by rail.
There. exist two practically identical international regulations concerning air
transportation:
• ICAO-TJ (Technical Instructions), which are the recommendations of the International

Organization for Civil Air Traffic, a subdivision of the UNO, located at Montreal,
Canada, and
• IATA-DGR (Dangeous Goods Regulations),36 the recommendations of the
International Air Traffic Association.
The transport of radioactive substances is the concern 'of the International Atomic
Energy Agency (/AEA), which has issued "Regulations on the Safe Transport of
Radioactive Materials, 1973".
Red Book on Transportation of Hazardous Materials, and other Guidebooks

The Red Book on Transportation of Hazardous Materials 37 is written as a guide for a
wide range of people concerned with the transportation of hazardous materials,
including hazardous substances and hazardous wastes designated by the US
Environmental Protection Agency (USEPA). It is designed to enable people to better
understand the nature and scope of the US Federal Transportation Regulations issued
and enforced by the Department of Transportation (DOT),). Environmental Protection
Agency (EPA), and related State and international agencies. Appendix A presents
information on vapor pressure curves for selected flammable liquids. Appendix B
summarizes Chapter 9 of the United Nations "Orange Book" - Recommendations on the
Transport of Dangerous Goods. This Chapter 9 of the UN recommendations contains
perfopnaDce-oriented packaging standards being adopted by various national and
international bodies, including RlDIADR in Europe, the International Maritime
Organization, the International Civil Aviation Organization, and Transport Canada. It
has been proposed, in the USA, for adoption as a replacement for 49 CFR Part 178 in
DOT's rule making Docket No. HM-181. Appendix C describes selectefl DOT rule-
making notices and amendments dockets Nos. HM-l to HM-198, whereas o\ppendic D
mentions pertinent hazardous materials statutes.
Another guidebook similar to the first one, entitled Transportation of Hazardous
Materials -A Guide to Compliance,38 has been written by Cheremisinoff as a guidebook
to understanding and applying the US Department of Transportation Hazard Materials
Regulations (HMR) to the shipment and handling of regulated hazardous materials. The
intend of the guidebook is to provide a general reference that can be used by American
industry managers to provide training of employees involved in regu.Iated hazardous
materials and articles transportation. The book gives valuable explanations on
classification terms, packing groups for hazardous materials, labeling, and the like.
2.6.2. PACKAGING AND LABELING
It is essential that chemicals which are to be transported are adequately packaged and
labeled. The packaging and labeling of chemicals is the subject of BEC Directive
67/5841EEC, as amended by subsequent directives. The provision of these directives are
implemented in the Packaging and Labeling of Dangerous Substances Regulations
1978. Further recommendations on the marking and labeling of dangerous chemicals
are given in "Marking Containers of Hazardous Chemicals and Labeling Chemicals", A
76 CHAPTER 2
Guide to the Regulations for the Labeling of Hazardous Chemicals by the CIA (1970
Item 13), 1977 Item 14).
Some aspects of packaging have been covered by Kieselbach in his report. 39 The
various aspects of packaging and labeling of chemicals for road transport include:
(1) packaging and labeling of the chemical, (2) loading and securing of the container
on the vehicle, and (3) labeling and documentation of the vehicle. Loads should be well
secured to the vehicle. The Code of Practice - Safety of Loads on Vehicles by the DOE
(Department of Environment) covers this aspect.
2.6.3. SWISS REGULATIONS FOR DANGEROUS GOODS TRANSPORTATION

ON LAND
It should be noted that in Switzerland there does not exist a Ministry of Traffic. The
responsibility for ensuring the correct application of the regulations is imparted to the
Cantons and to the Swiss Confederation.
Road Transportation of Hazardous Materials

In Switzerland road transportation of hazardous materials (including radioactive
materials) is regulated by the following ordinances (Figure 2.14).
Road Transport
• ADR ordinance, "Accord Europeen Relatif au Transport des Marchandises Dangereuses
par Route (ADR)",35 or "Europiiisches Uebereinkommen uber die internationale Beforde-
rung gefiihrlicher Guter auf der Strasse (ADR)", vom 30. September 1957 (status by: 1. Mai
1985)., and by its Swiss equivalent below.
• SDRIRSD ordinance: "SDR, Ordonnance Suisse Relatiw' au Transport des Marchandises
Dangereuses par Route, 1972)". This ordinance has beenfdly revised in 1985 antifollows
now the description of the international ADR agreement. The German terminology is now:
"RSD, Beforderung gefiihrlicher Guter auf der Strasse". Verordnung jUr die Beforderung
gefiihrlicher Guter auf der Strasse (RSD) vom 17. April 1985.
Figure 2.14. Swiss ordinances concerning road transportation of hazardous materials.
• Bundesgesetz uber den Strassenvekehr (Strassenverkehrsgesetz) vom 19.12.1958. status

by 1988 (SR 741.01).
• Verordnung uber die beforderung gefiihrlicher Guter auf der Strasse (SDR) vom
17.5.1985. Status by 1.1.1990 (SR 741.621).
• Anlagen der SDR und ihre Anhiinge vom 1.5.1985. status by 1.1.1993.
Figure 2.15. Other Swiss regulations and Federal laws concerning the transportation of dangerous goods
by road..
Other regulations concerning the transportation of dangerous goods by road in

Switzerland are mentioned in Figure 2.15.
Rail Transponation of Hazardous Materials

In Switzerland railroad transportation of hazardous materials (including radioactive ma-
terials) is regulated by the following ordinances (Figure 2.16).
• RID regulation: "Reglement International concernant le transport des marchandises

Dangereuses par chemin de fer (RID)".
• RSD regulation: "Reglement Suisse Concernant le Transport des Marchandises Dange-
reuses par Chemin de Fer, 1967 (RSD)", which is actually based on RID and follows its
content. The Swiss German equivalent is: "Ordnung flir die internationale und schweizeri-
sche Eisenbahnbeforderung gefdhrlicher Gater (RIDIRSD) vom 1.1.1990".
Figure 2.16. Regulation concerning railroad transportation in Switzerland.
Other regulations concerning the transportation of dangerous goods by rail, in

Switzerland, are:
• Bundesgesetz aber den Transport im offentlichen Verkehr (Transportgesetz) yom 4.10.

1985 (SR 742.40).
• Verordnung aber den Transport im offentlichen Verkehr (Transportverordnung) yom
5.11. 1986. Status by 1.1.1990 (SR. 742.401).
• Anlage I zur Transportverordnung (RSD). Status by 1.1.1993.
Figure 2.17. Other Swiss regulations concerning the transportation of dangerous goods by rail.
The Swiss ordinances and other Swiss regulations can be obtained from the Swiss
Federal Printing and Materials' Office, EDMZ ("Eidgenossische Drucksachen und Ma··
terialzentrale") in Bern, Switzerl~d, .
Since 1.1.1992, the Swiss Federal Dangerous Goods Inspectorate (EGI, "Eidgenos-
sische Gefahrgut-Inspektorat") of the Swiss Association for the Control of Pressure
Vessels, SVDB ("Schweizerisches Verein fUr Druckbehlilteriiberwachung") is now res-
ponsible for the examination and licensing qf transport containers.

BY AIR
The carriage of hazardous material by air is the concern of the Restricted Anicles Board
of the International Air Transpon Association (lATA). The lATA Restricted Anicles
Regulations place restrictions on the type and quantity of chemicals which are carried
and specify requirements for packaging and labeling. These regulations are, also in
Switzerland,40binding on the lATA member airlines (Figure 2.18).
The regulations list some 2300 substances requiring special packing and handling.
These include flammable liquids and solids, combustible liquids, explosives, oxidizing
materials, organic peroxides, corrosive materials, poisons, noxious or irritating sub-
stances and radioactive materials. A distinction is drawn between passenger and cargo
aircraft. The carriage of radioactive substances is governed by the lATA Regulations on
78 CHAPTER 2
Radioactive Materials which embody the IAEA Regulations on the Safe Transport of
Radioactive Materials.
Dangerous Goods Regulations (lATA)

• The lATA Dangerous Goods Regulations are published in order to provide procedures
for the shipper and the operator by which articles and substances with hazardous
properties can be safely transported by air on all commercial air transport.
• Dangerous goods are articles or substances which are capable of posing a significant
risk to health, safety, or to property, when transported by air, and which are classified
according to section 3 of the Dangerous Goods Regulations (lATA).
Figure 2./8. IAT A dangerous goods regulations.
The ICAO-T/ (Technical Instructions) and the IATA-DGR (Dangerous Goods

Regulations) regUlations are liable for Switzerland through the following additional
Swiss regulation (Figure 2.19).
• Verordnung uber die Beforderung gefiihrlicher Guter mit LuftJahrzeugen (SR 748.411)
vom20.9.1984
Figure 2./9. Swiss regulation complementing the lATA regulations.
The Swiss Federal Office for Civil Aeronautics (BAZL)i is responsible for the
application of the regulations concerning air transportation of dangerous goods. This
instance, however, does not have its own experts for the domain concerning the
transportation of dangerous goods.

ON WATERWAYS
Transports with Swiss Sea Vessels

There are in Switzerland maritime companies owning sea vessels registered under
Swiss flag. Their port of attachment is Basle. As a consequence, Switzerland is a
member of the International Maritime Organization, IMO, located in London, a
subdivision of the UNO, and has atso signed the SOLAS agreement concerning the
protection of human life on see (Figure 2.20).
• Bundesbeschluss betreffend die Ermachtigung des Bundesrates zur Annahme von

Aenderungen des Uebereinkommens von 1974 zu Schutze des menschlichen Lebens auf See
vom 19.6.1981, BBl1980 [[709.
Figure 2.20. Swiss Federal decree concerning the international SOLAS agreement.
i BAZL, Bundesamt fUr Zivilluftfahrt, Bern.

This means that the IMoo code of the IMO is applicable to Switzerland which has to
fully comply to it. The Swiss Federal Ministry responsible for the application of this
regulation is the "Schweizerisches Seeschiffahrtsamt Basle".
Transports on the Rhine

Such transports are regulated by the following ordinance (Figure 2.21) .
• Verordnung iiber die Beforderung gefdhrlicher Guter auf dem Rhein (ADNR) yom
15.7.1987 (SR 747.201).
Figure 2.21. Swiss ordinance concerning the transportation of hazardous materials on the Rhine.
2.6.6. REMARKS CONCERNING THE REVISION OF THE REGULATIONS AND

ORDINANCES AND THEIR APPLICATION
Since 1.1.1993, there are several changes, modifications and improvements which were
brought to the different ordinances. Detailed information concerning such changes can
be found in Kieselbach 39 (1993).
TABLE 2.4. Swiss Authorities responsible for the application of the ordinances and regulations
RSD (Rail Transport) Swiss Federal Office of Transportationi

The Swiss Railroadii and Workshop Servicesiii
in Bern are exerting a controlling task.
SDR (Road Transports) "Cantonal Authorities"
IMOO Code (See Transport) Swiss office for shipping on lakes, Basleiv
IATA-OOR (Air Traffic) Swiss Federal Office for Civil Aeronauticsv,
with the help of SWISSAIR company
The Swiss Authorities mentioned in TABLE 2.4 are responsible of the application
of the different regulations and ordinances.
2.6.7 . TRANSPORTATION OF RADIOACTIVE SUBSTANCES
The transport of radioactive substances is the concern of the International Atomic

Energy Agency (IAEA), which has issued "Regulations on the Safe Transport of Radio-
active Materials, 1973".
In Switzerland such transport are usually related to the activities of the four nuclear
power plants Miihleberg (Canton of Bern), Beznau 1+11, and Leibstadt (Canton of Argo-
I BAY, Bundesamt fiir Yerkehr, Bem.

n SBB/cfF. Abteilung Zugftirderung. Bern .
.. ZfW, Werkstiittedienst, Bern.
iv Schweizerisches Seeschiffahrtsamt, Basel.
v Bundesamt fiir Zivilluftfahrt. Bern.
80 CHAPI'ER2
via), and of other institutions such as the nuclear research institute PSI (paul Scherrer
Institute) at Villigen, the AC laboratory of the Swiss army (Spiez), universities, hos-
pitals, and the two Swiss Federal Polytechnical Schools ETH (ZUrich) and EPFL
(Lausanne). The radioactive substances may involve radio-nuclides for medical, indus-
trial, or research application. The radioactive sources may be sealed or unsealed, and the
radio-nuclides are classified according to their toxicity. Handling and storage of
radioactive sources is submitted to the Swiss regulation concerning the peaceful use of
atomic energy and the oIdinance on radiation protection (Figure 2.22).
• Atomic Law: "l.oi fedirale du 23 dicembre 1959 sur l'utilisation pacifique de l'energie
atomique et la protection contre les radiations"
• Radioprotection ordinance: "Ordonnance Concernant la Protection Contre les Radia-
tions, 30th of June, 1976"; ("Verordnung aber den Strahlenschutz vom 30. Juni 1976")
• SDRIRSD ordinance: "SDR, Ordonnance Suisse Relative au Transport des Marchandises
Dangereuses par Route, 1972)". The German terminology is: "RSD, Verordnung flir die
Beforderung geftihrlicher GUter auf der Strasse, vom 17April 1985".
Figure 2.22. Swiss regulation concerning the peaceful use of atomic energy and the ordinance on
radiation protection.
Transport authorizations are not required for radioactive sources, as far as their
activity is below 20 pica-curie per gram for alpha emitters, and 20 nano-curie per gram
for beta emitters. Special transport authorizations are required for transporting larger
quantities of radioactive sources. Any transport must comply with special packaging
requirements prescribed in the ordinance.
The authorities in charge of controlling the Swiss nuclear activitiel', the HSK,
"Hauptabteilung fUr die Sicherheit der Kernkraftwerke" is responsible of the application
of the ordinance and regulation and is c;lelivering transport permits. The international
transport of nuclear fuel pins from a nuclear power station to an intermediate storage or
to the nuclear reprocessing facilities of La Hague (France), as well as from a fuel
manufacture to a Swiss nuclear power station is taking place in special heavy transport
casks under police escort. Such transports are, of cause, subjected to special authoriz-
ations and must be announced to the Swiss Government (Department of Inland Affairs)
as well as to International Atomic Energy Agency (lAEA) in Vienna.
Transport Inside a Nuclear Facility. The ordinance prescribes also that for radioactive
sources transported inside an authorized facility but outside a controlled radioactive
zone, the dose should not exceed O.lmSvlh (10 millirem per hour) at a distance of one
meter from the surface and 2 mSvlh (200 millirem per hour) on the surface.
Transport Outside a Nuclear Facility. Transports of radioactive substances outside a
facility, i.e., on private or public roads, must comply with the RSD/SDR ordinance. The
maximum allowable activity for free parcels is limited accoIding to tabulated values,
whereas a distinction is made regaIding solids, liquids and radioactive gases. The
contamination on parcels, trucks, containers, tanks and other packaging material (which
is of non-adherent in nature) has to be as low as possible and is not allowed to exceed
the following values (Figure 2.23 and Figure 2.24).
A) Parcels Subjected to Free TrafJic with UN·Nr. 2910

Maximum allowable surface contamination for transport on private or public roads (non-
adherent surface radioactivity):
• Beta and Gamma emitters with low toxicity: 0.4 Bq/cm2 (lO-SIlCi/cm2)
• for all other alpha emitters: 0.04 Bq/Cm2 (1O-6IlCi/cm2)
• Maximum allowable dose per hour measured at the sUrface of the parcel:
• S~vlh (O.S mremlh)
Figure 2.23. Transports of radioactive substances outside a facility: - Parcels subjected to free traffic.
B) Parcels not subjected to free traffic with UN-Nr. 2912, 2913, 2918, 2982
Maximum allowable Dose per hour for parcel belonging to this class but not subjected to
supplementary restrictions:
• 2 mSvlh (200 mremlh) measured at the surface of the parcel
• 0.1 mSvlh (1Omremlh) at a distance of one meter from the parcel surface
Maximum allowable dose per hour for parcel belonging to this class but subjected to
supplementary restrictions:
• higher than 2 mSvlh (200 mremlh) but in any case less than max. 10 mSvlh (1,000
mremlh).
Figure 2.24. Transports of radioactive substances outside a facility: - Parcels not subjected to free traffic.
Maximum allowable sUrface contamination for Beta, Gamma and Alpha Emitters of low
toxicity (non-adherent radioactive confllmination):
• 0.4 Bq/cm2 (J O-S IlCi/cm2) for parcel subjected to free traffic and/or non radioactive
goods
• 4 Bq/cm2 (10-4 IlCi/cm2) for all shipment
• all other alpha emitters
• 0.04 Bq/cm2 (10-6 IlCi/cm2) for parcels subjected to free traffic and/or non radioactive
goods
• 0.4 Bq/cm2 (1 O-SIlCi/cm2) for all other shipment
Figure 2.25. Transports of radioactive substances outside a facility: - maximum allowable surface conta-
mination.
Note: KfK, "Kernforschungszentrum Karlsruhe" (Germany) has prepared a report concerning

the Transport of Radioactive Substances on Public Roads in Germany (Berger, 1992).41 The
report is a guide to help the scientific and technical personal of this nuclear research center in the
choice of the modalities for the transport of radioactive materials on public roads in accordance to
international regulations and authorizations prevailing in Germany. Not only the Atomic Law, the
Radiation Protection Ordinance and the Ordinance on Dangerous Goods on Roads, which are
binding in any case, are considered, but also the internal rules and regulations of the nuclear
research center. This guide contains useful information regarding transport conditions prevailing
in Germany.
The maximum allowable surface contamination (non-adherent radioactive conta-
mination) are given in Figure 2.25.
82 CHAPTER 2
2.7. BUWAL's Guidelines for Assessing the Risks ofTraftic Ways in Switzerland
The Swiss Federal Office for Environment, Forestry and Landscape, BUW AL(i) has
issued a "Handbook No. Ill, to the Incident Regulations: - Directives for Traffic Ways
(December 1992)". This handbook complements the regulations and ordinance on
incidents and is to be used for assessing the risks of traffic ways2 (see below). The
methodological approach proposed by BUWAL for risk assessment, and the factors to
be used for risk calculation differ slightly from those methods described in our book.
We are, therefore, presenting in this Chapter, in a shorted form, the general and legal
aspects of Handbook No. III, whereas the BUWAL methodology for estimating the
specific risks of transportation of dangerous goods by road, rail, and waterways (i.e., the
Rhine river) will be presented in the respective chapters concerning such subject The
main topics from the original BUW AL handbook were translated from German into
English and reproduced in a summarized form to complement the chapters two to four
of this book.
The official BUWAL document in German, French, or Italian language may be
purchased from the Swiss Federal Office for Printed Matters and Office Material
(EDMZ(ii».
2.7.1. LEGAL BASIS TO THE SWISS "REGULATION CONCERNING

INCIDENTS"
Based on the act from February 27, 1991 decreed by the Swiss Federal Government,
which came in effect on April I, 1991, the Swiss regulations concerning the protection
against incidents (so-called "Regulation Concerning Incidents", abbreviated in German
as StFV, 1991)42 came into effect on April 1st, 1991, and is based on Article 10,
paragraph 4 of the "Environmental Protection Law" (USGt 3 from October 7, 1983;
and Article 26 paragraph 1 of the "Waters Protection Law" (GSchG. 1991).44
Both regulations allow the Swiss Federal Government to issue prescriptions con-
cerning the safety measures of processing plants, by applying their right in the form of
a prescription.
Based on Article 39, Paragraph 1 of the "Environmental Protection Law (USG,
1983)", and Article 47, Paragraph 1 of the "Waters Protection Law (GSchG. 1991)",
the "Regulation Concerning Incidents (StFV, 1991)" describes more precisely different
tasks and duties mentioned under the Articles 10, Paragraph 1-3 and 44-47 of the USG
law, namely official surveys; control and information right; non-disclosure and informa-
tion duties.
(.) BUW AL. BWldesamt fiir Umwelt, Wald WId Landschaft. Bern. Schweiz.
(n) EDMZ. Eidg. Drucksachen- und Materialzentrale. EDMZ, 3000 Bern, Schweiz.
2.7.2. MAIN TOPICS OF THE SWISS "REGULATION CONCERNING

INCIDENTS"
The objective of the Swiss "Regulation Concerning Incidents (StFV. 1991)" is the
protection of the population and the environment against severe damages, which might
occur as a result of (abnormal) plant operation or transportation incidents.
The main topics of this regulation are summarized below:
• Identification of risk concerning the population and the environment
• Risk reduction
• How the Owner has to manage the risk
• Control of the self-responsibility of the Owner by the Competent Authority
• Improving the information of the population concerned, in order for it to become
more aware of the risks involved.
The competent authorities are consequently allowed to issue prescriptions in the
form of handbooks to cover the following three domains:
1) Processing plants dealing with substances. products or special wastes
2) Processing plants dealing with micro-organisms
3) Traffic ways on which dangerous goods are transported
In the following chapters/sections we shall only be concerned with item (3), dealing
with the risk assessment of traffic ways on which dangerous goods are transported.
2.7.3. HANDBOOK III, TO THE "REGULATIONS CONCERNING INCIDENT,

STFV - GUIDING PRINCIPLES FOR TRAFFIC WAYS"
The Handbook III (Dec. 1992) to the "Regulations Concerning Incidents, StFV -
Guiding Principles for Traffic Ways" concerns more precisely the Owners of traffic
ways (railroads, transit roads and highways, and the Rhine river) where dangerous
goods are transported or loadedlunloaded.
Task and Duties of the Owner of Traffic Ways

The tasks and duties that are of the responsibility of the Owner of traffic ways can be
categorized into three groups according to:
• Validity range: railways; traffic ways; and the Rhine river.
• Prevention of incidents: implementation of preventive safety measures; prepa-
ration of a "short safety report"; assessment of the risks of traffic ways.
• Managing the incidents: mastering incidents; announcing and reporting of inci-
dents.
Transportation of Dangerous Goods by Rail

The dangerous goods are defined as those classified in the RSDIRID list in Annex 1
("Rules for the Transportation of Dangerous Goods by the Swiss Railways, RSD from
November 5th, 1986)45 to the regulation from November 5th, 1986 concerning
84 CHAPTER 2
''Transports Through Public Services" or the corresponding goods and objects defined
in international agreements. Each dangerous goods transport for which use is made of a
railway installation is to be considered as a transport and is consequently submitted to
the prescription.
Transportation of Dangerous Goods by Road

The transit roads (for truckers) are defined in the Swiss legislation.46 They include the
National Highways and main roads, and the road segments interconnecting a highway
to a main traffic way.
The following installations are also considered as being part of a traffic way:
custom-house installations, highways restaurants and parking lots, gasoline stations as
well as other installations related to traffic ways (in its function as transit road).
According to SDRIADR, the goods/products listed in "Appendix A" to the pres-
cription concerning the "Transportation of Dangerous Goods on the Road (SDRlADR
from April 17, 1985) or the goods and objects mentioned in corresponding international
agreements. Every forwarding of dangerous goods by road is considered as a transport
and is submitted to the prescription.
Transportation of Dangerous Goods on the Rhine River

The prescription concerning incidents is also applicable to the Swiss part of the
waterways of the Rhine river between Basle (-Hiiningen) and Rheinfelden, with
cantonal borders involving the cantons of Argovia. Basle-Landscape and Basle-City;
and national borders involving Germany, France and Switzerland.
Regarding the waterways jurisdiction, the Owner of the waterways is in this case the
State where the waterways segments are located. For the Swiss part of the river the
waterways segments are place under the responsibility of the respective cantons
represented by their competent administrative Authorities.
This situation creates thus some coordination problems regarding the assessment of
the risk presented by the Rhine river.
The dangerous goods forwarded on waterways are dermed according to the
"Appendix A" to the prescription ADNR from April 29. 1970. 47 The transportation of
dangerous goods by ship includes transport with inland ship and ocean ship.
Preventive Safety Measures

The Owner of a transport way is obliged by law to take any remedial action in order to
reduce the risks:
• Risk reduction measures
• Measures to avoid any incidents
• Measures to limit the consequences of a possible incident/accident
The owner has also to produce a short report concerning and describing the risks of
the transit way located under his responsibility. This report has to include, in short,
information on:
• Data and information concerning the traffic, traffic composition, and accidents
on the traffic segment.
• Information concerning the safety measures.

• Estimation of the likelihood of an incident/accident with severe consequences
for the population and the environment.
The main factors which influence the estimation of the situation are the technical
details concerning the category and type of construction of traffic ways; special infor-
mation regarding environment; the traffic load and composition, and the accident
statistics for the traffic segment under consideration.
Examination and Assessment of the "Short Report"

In a first step the Swiss Executive Authorities assess the correctness and completeness
of the so-called "Short Report" they have received from the Owner of the traffic ways
and verify the plausibility of the probabilities assumed for the incidents leading to
severe consequences. After an eventual visit in the field they make sure that the
assumptions are correct and the likelihood of incident occurrence is really small.
Depending on the results of this evaluation the competent Authorities are entitled to
request from the Owner a more detailed risk analysis, with particular emphasis on the
systematic investigation of the risk to the population and environment presented by the
transportation of dangerous goods on the traffic ways.
In a second step the competent Authorities examines the detailed risk analysis
assessment and have to make their own judgment on whether the resulting risk is
acceptable or not.
2.7.4. DUTIES OF THE EXECUTIVE AUTHORITIES
The competencies and duties between the Swiss Confederation, the Cantons, and the
Executive Authorities were defined as Indicateri below:
Duties of the Executive Authorities at Local Level

The responsibility of the executive authorities, concerning the application of the inci-
dents regulations include:
- Information and control function regarding their jurisdiction domain
- Examination and judgment concerning the short report
- Decision regarding additional risk assessment studies
Examination and judgment of the risk assessment study
- Decision concerning the implementation of additional safety measures
Information duty concerning the results of the official control
Dispensation of preparing a short report
Duties of the Swiss Cantons

The duties of the Swiss cantons can be summarized as follows:
Creation of an Incident Reporting Center
Information and alarm in case of incidents
86 CHAPTER 2
Coordination of the emergency services

Informing the Swiss Federal Office for Environment, Forestry and Landscape
(BUWAL) of any incident involving dangerous goods transports.
Incidents Reporting Center. The cantons have to designate an incidents reporting

center. The duty of such an office has to collect any incidents announcements and to
inform immediately the special emergency services responsible for managing such
incidents. Moreover the cantons have to designate a centralized office, which is
responsible of transmitting immediately the information to the Alarm Bureau of the
Swiss Confederation located in the Swiss Meteorological Institute (ARMA).
Duties of the Swiss Confederation

The Swiss Confederation has to take action, as control instance concerning the
achievement of the requested compliance with the incidents regulation.
Data gathering
- Information of foreign countries
- Expert commissions
- Directives
In that respect the "Office for Environment, Forestry and Landscape (BUWAL)" is
acting as a central place for collecting and assessing the data received from the owners
of traffic ways. In the case of incidents, which may possibly have important
consequences and extend beyond the Swiss border, the competent Authorities of the
Swiss Confederation have to take contact with the Swiss representatives in the foreign
country and inform the foreign national competent authorities of the possible extent of
the incident
Information and Alarm in Case of Incidents. It is the duty of the cantons to inform in
good time the population concerned and eventually to release the alarm and indicate the
appropriate measures to be taken. They have also to inform and eventually alarm the
neighboring cantons and states when the incidents might have severe consequences
across the cantons or states borders.
2.7.5. LIMITATION OF QUANTITATIVE RISK ANALYSIS IN DANGEROUS

GOODS TRANSPORTATION SYS1EMS
An important limitation of quantitative risk analysis is that it can provide no guarantee

that all accident initiators or accident scenarios have been identified. Especially
collision scenarios present a problem due to the non-repeatability of event sequence,
due to varying vehicle type, vehicle speed, road conditions, angle of collision, etc.
These effects will playa role regarding a possible leak and will influence the size of the
hole through which dangerous products may escape. Models can be validated with one
or more experiments, but it cannot be assumed that a model is appropriate for all
situations.
The accuracy of absolute risk results depends on whether all the significant risk
contributors have been included, the realism of the models, and the uncertainty
associated with the input data. If the principle risk contributors can be estimated from
adequate historical data, the resulting uncertainties will be small, but they can be one to
two orders of magnitude larger in case of rare events for which the statistical data are
scarce. This statement was confirmed by the result of a benchmark exercise prepared
during an a recent international conference on the risk of transporting dangerous goods
(Saccomanno, 1993).48
2.7.6. REMARK CONCERNING THE SWISS METHOD
The substances chosen by BUW AL to be included in the Swiss study were judged to be
representative of those categories of substances which could cause large numbers of
fatalities or cause a severe environmental pollution, and which represented a significant
proportion of the overall traffic in dangerous substances.
Summary
The scope of regional risk assessment concerning transportation of hazardous materials

has to include all transport systems such as rail, road, ship and pipeline networks. The
present chapter introduces the probabilistic risk criteria against which judgments could
be made as to the comparability of various land uses. Individual and societal risk
criteria, as well as environmental risk criteria are highlighted. The concepts of risk
matrices, acceptance of risk levels and the use, availability and comparability of
sta':i.;tical data are discussed at length. Methods and techniques for risk prioritization at
the regional level (including the risk from the transportation of dangerous goods in a
given region) were included in the economy of the chapter. The international and Swiss
legislation concerning transportation of dangerous goods by rail, road and air is pre-
sented. According to the Ordinance of April 1, 1991 concerning the "Protection Against
Incidents" (StOrfallverordnung, StFV), the Swiss Federal Office for Environment,
Forestry and Landscape (BUWAL) has proposed and recommended a methodology for
assessing the risk of traffic ways in Switzerland. The method is of general interest and is
described with some details here and in the next chapters. It can be used by others after
modifying the coefficients to reflect the traffic and accident conditions prevailing in a
different country.
88 CHAPTER 2
1 BUWAL, Bundesamt fiIr Umwelt, Wald und Landschaft, Bern, Schweiz.

2 BUWAL, Handbuch m zur SWrfallverordnung, Richtlinien fiIr Verkehrswege, Dezember 1992, Bundesamt
fiIr Umwelt, Wald und Landschaft (BUWAL), Bern.
3 Kletz, T.A., Hazard analysis - A quantitative approach to safety, in Major Loss Prevention, (1971), pp. 75.
• Kletz, T.A., Specifying and designing protective systems, in Loss Prevention, 6 (1972), pp. 15.
S USC: Major Hazard Aspects of the Tranport of Dangerous Substances: - Report and appendices, Health
and Safety Commission (HSq, Advisory Committee on Dangerous Substances, HMSO, LOI1(!on, (1991).
6 Vervalin C.H., Fire Protection Manual for Hydrocarbon Processing Plants, Gulf Publisher, 1st. Ed., (1964);
2nd. Ed., (1974).
7 Eisenberg, N.A., et aI., Vulnerability Model, A Simulation System for Assessing Damage Resulting from
Marine Spills, Nat. Tech. Inf. Service, Report AD-A0l5-245, Springfield, Va, (1975).
8 Strehlow, R.A., Unconfined vapour cloud explosions - An overview, Fourteenth Symp. on Combustion,
Combustion Institute, Pittsburgh, Pa, (1973).
9 Davenport, J.A., A study of vapour cloud incidents, Eleventh Loss Prevention Symp., New Engng Prog.,
73(9), (1977), pp. 54.
10 Kletz, T.A., Unconfmed vapour cloud explosions. An attempt to quantify some of the factors involved.
Eleventh Loss Prevention Symp., Am. Inst. Chem. Engrs, New York, (1977).
11 McCrindle, W.I., A First Guide to Loss Preventiton, Instn. Chem. Engnrs, London, (1977).
12 Marshall, V.C., How lethal are explosions and toxic escapes, Chem. Engr, London, 323 (1977), pp. 573.
13 Slater, D.H., Vapour clouds, Cherny Ind., May 6, (1978), pp. 295.
14 Davies, P.A.; Lees, F.P., Impact speed of heavy goods vehicles, J. Hazard. Mater., 26 (1991), pp. 213-217.
IS Davies, P.A.; Lees, F.P., Accident speed of freight trains, J. Hazard. Mater., 28 (1991), pp. 367-370.
16 Clarke, R.K.; Foley, J.T.; Hartman, W.F., Larson, D.F., Severities of transportation accidents, Vol. IV,
Report SLA 74-0001, Sandia Laboratories, Albuquerque, USA, (1976).
17 Abkowitz, M.; Cheng, P.D.M., Developing a risk/cost framework for routing truck movements of hazardous
materials, Accident Analysis and Prevention, 20 (I), (1988), pp. 39-51.
18 Petersen, G.; Rosenthaler, Ch.; Scheifele, D., STRADA-DB, Kurz-Information 1993, Bundesamt fiIr Stras-
senbau, EYED, 3003 Bern. (Zu beziehen rei EDMZ, 3000 Bern).
19 Manual for the Classification and Prioriti.."ation of Risk from Major Accidents in Process and Related
Industries. (Draft November 1991), Inter Agency Programme on the Assessment and Management of
Health and Environmental Risks from Energy and Other Complex Industrial Systems, International
Atomic Energy Agency (IAEA), Vienna
20 Simoni, R., Einbezug von SWrfal1risiken in die Raumplanung. Ein methodisches Konzept zur Erarbeitung
der raumbezogenen StOrfailrisikosituation aufgrund ortsfester Anlagen und Verkehrswege als raum-
planerische Entscheidungsgrundlage. VDF-Verlag der Fachvereine Zllrich, ETH, Swiss Federal Institute
ofTechnology, Zllrich. (1995).
21 Gugan, K., Institution of the Chemical Engineers, Gulf Publishing Co., Houston, (1979).
22 Streblow, R.A., Unconfmed vapor cloud explosions - An overview, Presented at the 14th Symposium
(International) on Combustion at Pensylvania State University, (August 1972).
23 Westbrook, G.W., The bulk distribution of toxic substances: a safety assessment of the carriage of liquid
chlorine, in Loss Prevention and Safety Promotion, Vol. I, (1974), pp.197.
24 Browning, R.L., Estimating loss probabilities, Chern. Engng., Albany, Dec. IS, (1969), pp. 135.
2S Leimkiihler, F.F; Karsen, M.J.; Thompson, J.T., Statistical analysis of the frequency and severity of
accidents to potential highway carriers of highly radioactive materials, Report No. NYO-9771, U.S.
Atomic Energy Commission, (1961).
26 Leimkiihler, F.F, Trucking of radioactive materials: Safety vs. economy in highway transport, (ed.) The
John Hopkins Press, Baltimore, Maryland, (1963).
27 Chira-Chavaia, T., Data from TRB-proposed national monitoring systems and procedures for analysis of
truck accident rates, in Transportation Research Record, , Transportation Research Board, Washington,
DC, 1322 (1991), pp. 44-49.
28 Hu, P.S., Estmating commercial truck VMT of interstate motor carlers: - Data evaluation, Report ORNU
TM-11278, Oak Ridge national Laboratory, Tennessee, (1989).
29 Harwood, D.W.; Russel, E.R., Present Practices of Highway Transportation of Hazardous Materials, Report
FAWH-RD-89-013, U.S. Department of Transportation, (May 1990).
30 Abkowitz, M.; List, G.F., Hazardous Materialss Transportation: Commodity Flow and Incident/Accident
Information Systems, U.S. Congress Office of Technology Assessment, (January 1986).
31 Bendixen, L.M.; Barkan, C.P.L., Development of an interindustry risk model for hazardous material
transportation by rail. Paper read at the International Consensus Conference: Risks of Transporting
Dangerous Goods, April 6-8 (1992), University of Waterloo, Toronto, Canada (1992).
32 Purdy, G., Risk l:Ilaiysis of the transportation of dangerous goods by road and rail, J. Hazard. Mater. 33
(1993), pp. 229-259.
33 Hubert P.; et aI., Les risques d'accident majeurs dans les transports de mati~res dangereuse, Centre d'Etude
sur l'Evaluation de laProtection Nuc16aire, D 114, France, (1985).
:w HSC: Major hazard aspects of the tranport of dangerous substances: - Report and appendices, Healt and
Safety Commission (HSC), Advisory Committee on Dangerous Substances, HMSO, London, (1991).
3~ European Agreement concerning the International Carriage of Dangerous Goods by Road (ADR), and
protocol of signature, Economic Commission for Europe (ECElfRANS/I00) Inland Transport Commit-
tee, United Nations, New York, (1993).
36 Dangerous Goods Regulations, lATA Resolution 618, Attachment "A", International Air Transport Asso-
ciation, Montreal-Geneva, 1993.
37 Red Book on Transportation of Hazardous Materials, Second Edition, (eds.) Lawrence W. Bierlein, Van
Nostrand Reinhold Company, New York, (1988).
38 Cheremisinoff, N.P., Transportation of Hazardous Materials - A Guide to Compliance, (ed.) Noyes Publi-
cations, New Jersey, USA, (1984).
39 Kieselbach, R., Transport geflihrlicher GUter in der Schweiz, - Basisinformationen I Standortbestimmung
1992, Swiss Federal Laboratories for Materials Testing and Research, Diibendorf I ZUrich, EMPA
Bericht Nr.227, (1993).
40 Dangerous Goods Regulations, lATA, International Air Transport Association, 34th Edition, (lst January
1993).
<I Berger, H.U., Transport radioaktiverstoffe auf offentlichen Strassen, unter Beri1cksichtigung der dem KfK
erteiltcn Genehmigingen, Bericht KfK 5085, (Oktober 1992).
<2 Verordnung yom 27. Februar 1991 fiber den Schutz vor StOrflillen (SWrfailverordnung, StFV), SR 814.ot2.
43 Bundesgesetz Yom 7. Oktober 1983 fiber den Umweltschutz (Umweltschutzgesetz, USt;), SR 814.01.
.. Bundesgesetz VOD' 24. Ianuar 1991 fiber den Schutz der Gewlisser gegen Verunreinigung (Gewlisserschutz-
gesetz, GSchG), SR 814.20.
<~ Verordnung yom 5. November 1986 fiber den Transport im offentlichen Verkehr (Tranportverordnung,
TV), Anlage 1 (Ordnung fur die Schweizerische EisenbahnbefOrderung geflihrlicher Gfiter, RSD), SR
742.401.
46 Durchgangsstrassenverordnung yom 18. Dezember 1991, Anhiingen 1 und 2 .
., Verordnung Yom 29. April 1970 fiber die BefOrderung geflihrlicher Gfiter auf dem Rhein (ADNR), SR
747.224.141.
<II Saccomanno, F.F.; Yu M.; Shortreed, J.H., Risk uncertainty in the transport of hazardous materials, in
Transportation Research Record 1383, Transportation Research Board, National Research Council.
Washington, DC, (1993), pp. 58-66.
CHAPTER 3
RISK ANALYSIS OF HAZARDOUS MATERIALS TRANSPORTATION

BYROAD
General information and know-how related to truck accident rates, accidents in tunnels
and their consequences are presented in this chapter. The Swiss methodology recom-
mended by BUWALi for assessing the risks of hazaIdous materials transportation by
road is explained in details. Finally, new developments in traffic and vehicle control are
presented.
3.1. Information and Knowledge about Truck Accidents and Consequences
3.1.1. EXPERIENCE CONCERNING TRUCK ACCIDENTS
Frequency and Severity of u.s. Truck Accidents

In 1961, Leimkuhler et aI. I performed a statistical analysis of the frequency and severity
of truck accidents as a function of cargo type and season of the year (TABLE 3.1).
TABLE 3.1. Truck accident rates (corresponding to U.S.lraffic con-
ditions)
U.S. Truck Acddent Rate

Highway Accidents per million vhc.-miles
Class
Area Type Roadway Type U.S. Weighted
Average"
Rural Two-lane 2.19

Rural Multilane undivided 4.49
Rural Multilane divided 2.1S
Rural Freeway 0.64
U:rban Two-lane 8.66
Urban Multilane undivided 13.92
Urban Multilane divided 12.47
Urban One-way street 9.70
Urban Freeway 2.18
Source: Harwood and Russel, 1990
• Weighted by vhc-miles of truck travel.
• BUWAL: Bundesamt fiIr Umwelt, Wald und Landschaft, Bern. Switzerland.

92 CHAPTER 3
In the late 1960s and early 1970s, Brobse (USAEC 3 , 1972) developed a scheme for
categorizing highway and rail accident severity in terms of impact speed and fire
duration. TABLE 3.2 shows Brobst's values for accident frequencies associated with the
u.s. accident severity subcategories (USAEC 1972), which were based on Leimkuhler's
speed distribution.
TABLE 3.2. Accident frequency based on severity category
Frequency per Vehicle Mile

Severity Category Vehicle speed Fire duration Truck Rail
mph hr
Minor accidents 0-30 <112 6 x 10.9 6 X 10.9
0-30 0 4 x 10.7 4.7 X 10.7
30-50 0 9 x 10.7 2.6 X 10.7
Total <112 1.3 x 10.6 7.3 X 10.7
Moderate accidents 0-30 112-1 5 x 10. 11 9.3 X 10. 10

50-70 <112 1 x 10.8 3.3 X 10.9
50-70 0 5 x 10.9 9.9 X 10. 10
50-70 0 3 x 10.7 7.5 X 10.8
Total 3 x 10.7 7.9 X 10.8
Severe accidents 0-30 >1 5 x 10- 12 7.0 X 10- 11

30-50 >1 1 x 10- 11 3.9 X 10- 11
30-50 112-1 1 x 10- 10 5.1 X 10- 10
50-70 112-1 6 x 10- 12 1.5 X 10-10
>70 <112 1 x 10- 10 1 X 10- 11
>70 0 8 x 10-9 8 X 10- 10
Total 8 x 10.9 1.5 X 10-9
Extra severe 50-70 >1 6 x 10- 13 1.1 X 10- 11

accidents
>70 112-1 2 x 10- 13 1.6 X 10- 12
Total 8 x 10- 13 1.3 X 10- 11
Extreme accidents >70 >1 2 x 10.14 1.2 X 10- 13

Source: USAEC 1972
Since the early 1980s, it has been known that the truck accident rate varies with road
type and with population density. The most widely cited (TABLE 3.1) values valid for
the USA are those of Harwood and RusseI4 (1990). Both used U.S. data bases that
included data on highway geometries, truck volumes, and truck accidents. The above
mentioned weighted average values (TABLE 3.1) are proposed by FHWA as national
default values, that is, for use in U.S. risk analysis when better data are unavailable.
There is however a substantial variation in accident rate between some states an the
national U.S. average value (Hobeika and Kims, 1991). These authors pointed out that
the largest variations occur where the sample sizes were the smallest, and the weighted
average minimizes the effect of such values. Therefore the average values have an
uncertainty of as much as ±2. There are numerous earlier studies that substantiate the
RISK ANALYSIS OF HAZARDOUS MATERIALS TRANSPORTS BY ROAD 93
values in TABLE 3.1 (Smith and Wilmot6 , 1982; Abkowitz et al.7, 1984; Graf and
Archuleta8, 1985; Jovanis et al.9, 1989).
A set of data that are consistent with preceding (TABLE 3.1) are presented in
TABLE 3.3.
TABLE 3.3. u.S. Truck accident rates by highway type
Highway Type Accident Rate

Controlled access 3.8 x lO-6/mile
Non-controlled access 28.4 x 10·6/mile
Local streets 15.6 x 1O-6/m ile
Compded from Jovarus et aL9. (1989).
TABLE 3.4. Partial listing offactors considered to affect truck accidents (USA)
Truck Type or Configuration Highway

Number of trailers Function
Number of axles on Access control
tractor/trailer Number of lanes
Cab type Lane width
Cargo area configuration Shoulder width
Truck Size and Weight Shoulder surface
Width of trailer Median width
Length. overall Horizontal alignment
Length. trailer Vertical alignment
Emptylloaded Surface condition (wet/dry, etc.)
Weight, gross Pavement condition
Weight, trailer Pavement type
Truck Operations Traffic
Cargo type Volume (AOT)
Operator type Volume (day/night)
Trip type Percent trucks
Truck Driver Environment
Age Visibility
Experience with rig Weather
Hours of service Light
Driver condition Temporal
Location Monthlseason of year
Country/State Day of week
Urban/rural Time of day
Source: US DOT 1990. Note: ADT means average daily traffic.
A word of caution must be given here: not all database analyses are equally valid
The user of reported data analyses must develop its own confidence level with the
methodology used to analyze the data. One way of doing this is to compare the results
reported by several risk analysts. ldea1ly. the values reported in the literature should
describe any differences from previously reported values and explain the variations.
Sometimes the total number of accidents reported by some source may differ from the
94 CHAPI'ER3
number of accidents mentioned by another investigator, due to counting only accidents

above the regulatory reporting threshold. It must be noted that the uncertainty in the
truck accident rate is small compared to the uncertainty in some of the other parameters
used in absolute risk calculations.
In Switzerland, for instance, the police is only called when people are injured or
killed, thus minor accidents will not appear in the statistics of the police of the different
cantons, but such accidents will probably be announced to some insurance company and
will then appear in the statistics of the respective Swiss insurance companies. Moreover,
in the past, reporting was not done everywhere on the same basis, and the reports
concerning truck accidents did not mention whether a truck was carrying dangerous
goods or not
Factors Influencing Truck Accident Rates

The U.S. DOT (Harwood and Russel to, 1990) has studied the factors thought to have an
influence on truck accident rates. A partial listing of these factors is contained in
TABLE 3.4.
Impact Speed Of Heavy Goods Vehicles

For appraisal of cars, trucks and tankers accidents information on impact speed are quite
useful. In Great Britain, such informationll on impact speed of heavy good vehicles at
the moment of collision were obtained from tachometer readings (Figure 3.1). Some
110 tachograph-based reports covering the period 1978-1982 were analyzed. Data are
available for built-up areas, outside built-up areas, and motor ways. The speed was
ranging from 0 to 60 mph:
Impact Speed of Heavy Goods Vehicles at the MOlMn' of a Collision (Great Britain)
• Built-up areas: A total of68 accidents, of which 36 were heavy goods vehicles collisions.
• Outside built-up areas: A total of 32 accidents, of which 21 were heavy goods vehicles
collisions.
• Motor ways: A total of 13 accidents, of which 7 were heavy goods vehicles collisions.
• The speed was ranging from 0 to 60 mph (96 km/h). It must be noted that in Great
Britain, the road conditions and habits of driving are different from those in Switzerland.
Figure 3.1. Impact speed of heavy goods vehicles at the moment of a collision (Great Britain).
3.1.2. GASOLINE TRANSPORTS BY ROAD
The size and configuration of gasoline-hauling vehicles are constrained by road-use

laws that limit gross weight, number of axles, and axle loading.
In the USA, statistical data on truck accidents from the Bureau of Motor Carrier
Safety can be used. A review of the literature indicates that only a few studies have
focused on the hazards of transporting gasoline by road. Krasner l2 investigated
transportation accidents involving motor carriers of hazardous materials. Jones et al. 13
(1973) have analyzed the risk of hazardous material transportation.
A comprehensive study of the risk of transporting gasoline by truck was conducted

by the Battelle Pacific Northwest Laboratory14 in 1978. Fatalities were used a measure
of risk involved in transportation of gasoline. The probability of the occurrence of
individual accidents that result in a large number of fatalities was found to be relatively
low. For example, accidents that result in 10 or more fatalities were expected to occur in
the United States at the rate of once every 45 years.
The feasibility of reducing such risk by changing vehicle design has in the past not
been investigated. Although rollover accidents have been recognized as an important
cause of hazardous spills, the extent to which rollover involvement of vehicles can be
reduced by increasing the rollover threshold has not been examined. Another problem is
that of the size of a puncture hole caused by an accidental collision. A crash test
program for road vehicles (road trucks, tankers) has been presently initiated in
Germany, but the repetition of the same crash test does not lead to the same kind and
extent of damage. In the real world every collision happens in a different way, at
various speeds and angle of collision, involving one or more vehicles.
Gasoline and LPG Truck Releases: • A Canodian Study

The rrwst common accident types that were observed to cause releases of these two
commodities were:
• Overturn, collision followed by overturn, and collision.
Similar accident types were recorded for both commodities; however, the frequency of non-
transportation releases for LPG was higher, likely due to different reporting requirements
or practices. It was found that overturn and collision/overturn accidents resulted in similar
release sizes, while collision accidents resulted in statistically lower average release sizes.
The other spill types were examined qualitatively:
• Gasoline fire: for example, fire wru found to result in a 98% release of lading,
3 out of 4 times. Fire only occurred with gasoline spills.
• Spills: LPG spill sample size was much smaller than for gasoline.
Of the three truck types represented in the database, tractor trailer, tractor trailer with pup,
and tanker truck, the latter was the rrwst common, and hence was used for the statistical
analysis. It was found that the former two truck types could be considered as one sample in
terms of release percentage, but they could not be combined with the tanker truck sample.
Figure 3.2. A Canadian study on gasoline and LPG trucks releases.

The analysis of gasoline tanker accidents, a/o, in Michigan (USA), has shown that
tanker rollovers are the principle cause of gasoline spills and fires.
3.1.3. FACTORS WHICH CONTRIBUTED TO GASOLINE AND LPG TRUCK

RELEASES (CANADA EXPERIENCE)
The Canadian Data Base and Transportation Risk Model

Transport Canada has been collecting data on dangerous goods spills since 1979. In
July of 1985, reporting of dangerous goods occurrences in Canada became mandatory
96 CHAPTER 3
under the Transportation of Dangerous Goods Regulations (SOR/85-77). Since that date
the Transport Dangerous Goods' Directorate of Transport Canada has collected data on
numerous characteristics of each reported spill in a comprehensive data base. IS
The 1986 and 1987 data for gasoline and liquefied petroleum gas (LPG) truck spills
that occurred en route were examined with the objective of calibrating the accident and
release submodels for use in a transportation risk model that was developed previously
for Transport Canada.
Analysis of Gasoline and LPG Truck Releases (Canada Case)

The results of an empirical analysis of Canadian gasoline and LPG truck releases are
summarized (Figure 3.2).
3.2. Risk Assessment of the Transportation of Hazardous Materials Through

Road Tunnels
There are only a few statistical data and studies concerning the risk of accidental
releases and fIres in road tunnels, that were reported. In this section and in Chapter 10
we will review the results of some risk studies dealing with the risk of hazardous
materials transports in road tunnels.
3.2.1. U.K. STUDY OF THE RISK OF ROAD TUNNELS
Survey of Freight Traffic of Hazardous Materials

In Great Britain, the Department of Transport has commissioned a risk as~essment to
provide quantitative information on risks to the public and possible damage to tunnel
structure from incidents involving the transport of hazardous substances (Considine,
1986).16
Information concerning freight traffic of dangerous substances may be found in
Transport Statistics Great Britain 1971-1981. 17 It can be seen that approximately 5%
of the freight moved annually by road is classed as petroleum products, 3% as chemi-
cals, and 1% as fertilizers. The most comprehensive UK survey of dangerous goods is
that carried out by Wilson from the DOT. This survey was reviewed by Hills (1982)18,
and others. 19 The results from another survey by the UK Department of the
Environment, prepared in 1973, show the following proportion of vehicles carrying
hazardous goods, of which petroleum represents approximately 70% of all dangerous
goods transported by road (TABLE 3.5).
TABLE 3.5. Breakdown of dangerous goods trans-

ported by road in the UK in 1973, according to classes
of materials
Percentage Type of Material

56 Gasoline
8 Other flammable liquids
3 Flammable gases
3 Toxic gases
3 Other gases
19 Corrosives
3 Toxic solids and liquids
5 Flammable solids
Spill Probability. In UK, an overall spill probability of approximately 5 x 10.8 per

vehicle-km was estimated for road tanker traffic. Tanker spill data also led to the
conclusion that leakage and spills from liquefied gas carriers occurred no less frequently
than from road tankers as a whole. There is a problem in estimating spill frequencies
from packages, due to the lack of pertinent information. MacLean's survero indicated
that the frequency of packaged spills was about twice that of spills from bulk carriers
making a total spill probability for packaged goods of 1.5 x 10.7 per tanker-km (for the
year 1981). It was also found that the frequency of important load fires in packaged
goods was about 50% of that attributed to bulk tankers, making an overall load fire
probability of 1.5 x 10's per tanker-km.
It must be recognized that, in general, statistics generated from regional surveys are
l~~ikely to apply with any degree of reliability to re!;ions other than those were the
~urveys were conducted. Usually the degree of details is insufficient to permit the
prediction of dangerous goods traffic flows for all roads. Moreover depending on the
country of origin, the traffic and road conditions, as well as the speed limitations are
very much different. This remark does also apply to railway installations and the type of
wagons and coupling device in use, which may vary quite a lot from on country to
another.
Accident Scenarios for Road Tunnels

The development of accident scenarios for tunnels can be realized by applying the
techniques of fault-tree and event-tree analysis. Such techniques involve the systematic
decomposition of an event into combination of sub-events (fault-tree analysis) or the
consideration of all possible outcomes of a single initiating event (event-tree analysis).
Using such methods, one finds out that the most important mechanisms by which
people or tunnel structures could be affected by the various classes of hazardous
substances are:
• Inhalation of toxic or asphyxiant vapors (fumes).
• Dermal contact with toxic material (due to liquid spills or finely dispersed
solids).
• Contact with corrosive, caustic, or cryogenic material.
98 CHAPTER 3
• Fire (effect of temperature affecting tunnel structures, or people, and effect of

asphyxiating fumes).
• Explosion (effect of blast and blisters). There are three type of explosions to
consider: detonation, deflagration, and sudden release of energy stored in
compressed gases (or, physical explosion from the bursting of a vessel subjected
to an internal explosion).
Fire and Explosion Risk. The fIre and explosion risk of a hazardous material road
tanker in a highway tunnel is a function of the frequency with which an incident may
occur and the magnitude of such an incident. Fires resulting from the ignition of
hazardous substances may take on a number of forms including pool fIres and vapor
fires w.here the volatile liquid spillages (such as gasoline) are involved and, addition-
ally, for pressurized liquefied gas releases, torches (and fireballs). The frequency of
tunnel fire is usually very low, but the magnitude of such a fire in a highway tunnel is
significant. The pool fire mode will be by far the most important in terms of event
frequency. A pool fire inside a tunnel will cause heating of the walls by radiation and
convection from flames and hot gases. Significant damage may occur by thermal
spalling of tile grout and concrete. Usually, all fittings engulfed in the fire will be
damaged. Both thermal effects from flames and hot gases, and toxic effects of smoke
and combustion products will represent a threat to people in a tunnel fire. A fire
involving a 113.56 liters (30-gal) spill or a 75.71 Vmin. (20 gpm) leak of a liquefied
flammable gas or Class I flammable liquid, or involving a 605.7 I (l60-gal) spill or 100
gpm leak of a Class II or Class II combustible liquid, will endanger all people within the
tunnel but will probably not cause structural damage. On the other hand, a fire
involving a 378.54 liters (lOO-gal) or 151.42 Vmin.(40 gpm) leak of a liquefied
flammable gas or Type I flammable liquid, or a fIre involving 1.892.71 liters (500-gal)
spill or 757.08 Vmin. (200-gpm) i,~ of Class II or Class II combustible liquid, will
present a severe fire exposure to the tunnel structure, with ceiling temperatures ap-
proaching 1,093 °C (2,000 OF) for longer than 1 hour.
When a signifIcant vapor cloud builds up before ignition, a vapor fire is possible.
The hazards from vapor fires, in which significant overpressures are not involved, will
be dominated by the damage produced by flame contact of usually short duration. In
such a case cars present in the tunnel may take fire, but the tunnel structures will not be
damaged to a large extent. A hazardous material cargo spill involving a liquefied
flammable gas, or Type I flammable liquid that does not involve an immediate fIre can
create a significant explosion potential in a tunnel. The result of such an explosion
(subsonic deflagration or supersonic detonation) can create a blast overpressure that will
cause structural damage to the tunnel. Similarly, an explosion potential exists if a fire
involving a liquefied flammable gas or Class I flammable liquid is extinguished before
all of the available fuel is consumed or contained. Class II or Class III combustible
liquids do not present a significant explosion potential unless they are heated above
their flash point by an exposing fIre.
Quasi-instantaneous and continuous release of pressurized liquefIed flammable
gases will, when ignited almost immediately, bum as fireballs and torches, respectively.
Tunnel Temperatures During a Fire. Assuming a reference tunnel with approximately

following dimensions: 10 m wide, 4.85 m high, and 1.7 kIn long, with a horizontal
tunnel bore, it is possible to estimate the tunnel temperatures (TABLE 3.6 - TABLE
3.8) as a linear function of distance from the fire, for a given fire intensity (MW).
Exposure to temperatures above 50°C will quickly cause second degree bums and
is considered life threatening. In addition, smoke will fill the tunnel, in the direction in
which the fire ventilates, with toxic combustion products, making human survival
improbable beyond the point of fire origin.
TABLE 3.6. Estimated nmnel temperature as a function of fire intensity (100 MW) and dis-
tmce from fire
m 0 125 250 375 500 625 750 875

·C 856.0 7SO.3 644.6 539.0 433.3 327.6 221.9 116.3
TABLE 3.7. Estimated nmnel temperatures as a function of fire intensity

(SO MW) and distmce from fire
m 0 125 250 375 500

·C 856.0 678.4 500.8 323.2 145.6
TABLE 3.8. Estimated tlmnel tem-

perature as a function of fire
intensity (20 MW) and distmce
from fire
m o
·c 405
A hazardous material tunnel fire burning at an intensity of 20 MW can endanger the

lives of all people who are in the tunnel, but it will probably not cause serious structural
damage to the tunnel, because the ceiling temperature is not expected to exceed ca. 482
0c. A fire bumi~g at an intensity of 100 MW will, moreover, damage the structures of
the tunnel, because the ceiling temperatures within several hundred feet of the ftre will
approach 1,093 °C. It is worth noting that a tunnel emergency ventilation system can
reduce the temperatures in a tunnel during a ftre. The intensity of a highway tunnel ftre
involving a spilled hazardous material depends on the area of the spilled liquid, the
availability of combustion air, and the availability of the smoke to escape from the
tunnel. There is a linear relationship between the quantity of spilled liquid and the area
of unconfined spill. The duration of a hazardous material highway tunnel fire will
depend on the volume of available fuel and the depth of the spilled fuel, or the fuel
spill rate. A small unconfmed spill will spread to an average depth of 0.635 mm (0.25
in.). A fire involving such a spill will usually last less than 5 min. On the other side a
100 gpm leak (i.e., 100 x 3.7854 liters) in a 30 t gasoline tank truck could result in a
100 MW fire that will. last for approximately 80 min.
100 CHAPTER 3
Estimation of Societal Risks for Road Tunnels in the U.K.

Societal risks, in the U.K. 16, have been estimated for road tunnels and published. The
findings are reproduced here (TABLE 3.9). From historical international data, it appears
that major hazardous material fires in tunnels, once started, can be controlled only by
heroic efforts under fortunate circumstances in which effective support systems are also
present.
TABLE 3.9. Societal risk from tunnels in the UK
Societal Risks
(Frequency of N, or more fatalities per 10~ years)
Event
Motorists Public
Roule N=1 N=lO N=IOO N=1 N=IO N=IOO
Section
Corrosive/toxic liquid T 5.8 - - - - -
release OR 170 - - - - -
Flammable liquids T 45 15 15 - - -
OR 1,200 380 - 7 - -
Liquefied gas release T 1.6 0.61 0.44 - - -
OR 47 5.3 - 0.29 0.11 -
Toxic gas release T 3.2 3.2 0.19 - - -
OR 57 21 0.66 6.5 1.8 -
Asphyxiant gas release T 0.19 - - - - -
OR 6.6 - - - - -
Condensed phase OR 0.04 0.04 0.02 - - -
explosion T 0.88 0.44 - 0.12 0.8 -
Total 1,600 420 16 14 1.9 0
~: T: tunnel; OR: open road
3.2.2. U.S. STUDY OF THE RISK OF ROAD TUNNELS
Information on Tank Truck Fires in Highway Tunnels

Few hazardous material tank truck fires have occurred in highway tunnels in the United
States, primarily because such trucks have been prohibited from using most highway
tunnels since the 1949 Holland tunnel fire. At present, no U.S. statistics is available for
fires in tunnel, therefore other kind of statistics must be used to predict the frequency of
accidents and fires of hazardous material tank trucks. The best is to obtain infonnation
on the frequency of open highway accidents and fires of hazardous material tank trucks.
A very few reports are available:
ATA National Truck and Industrial Contest, 1982. American Trucking
Association, Inc., Alexandria, Va.
- Analysis of Accident Reports Involving Fire - January through June 1968.

Bureau of Motor Carrier Safety, Washington, DC.
- Summary of Motor Vehicle Accidents in the Petroleum Industry for 1981.
American Petroleum Institute, Washington, DC (August 1982).
- Accidents of Motor Carriers of Property (1979). Bureau of Motor Carrier Safety,
Washington, DC.
- New Drive for Truck Safety, Journal of American Insurance Association (May!
June 1970).
- Motor Vehicle Standards for Hazardous Material Transportation. Factory Mutual
research Corporation, Norwood, Mass. (January 1970).
Since the data cannot be extrapolated readily to European traffic conditions, some
findings from these reports will be presented here, however the data should be used
cautiously.
Tank trucks, in general, had a lower accident frequency than other trucks. Only
1.7 % of all truck accidents resulted in fire. Hazardous material tank trucks had a fire-
to-accident ratio that was 70% higher than that of general trucking industry, in which
2.9% of all accidents resulted in fire during the period July 1966 to December 1968.
Approximately half of the reported fires were caused by collisions. The remaining
50% were the result of non-collision accidents involving overheated brakes or tires,
defective exhaust systems, and defective electrical systems.
The spillage frequency of hazardous tank truck cargoes, as a result of accidents,
amounted to 8.5% of the accidents.
It was noted that the cargo was involved in 87% of the fires involving hazardous
material tank truck transports.
Prevention and Control of Highway Tunnel Fires

In the United States, a study similar to the British survey was also prepared to assess the
risk presented by the transportation of dangerous goods through highway tunnels and to
prevent and control the risk of fire?t
3.2.3. SWISS STUDY OF THE RISK OF ROAD TUNNELS
A Swiss study, entitled MORIKA, concerning the risk presented by the "Baregg" road
tunnel (National highway Nl, Baden, Switzerland) has been prepared recently for the
Canton of Argovia. 22 The knowledge and results of this survey will be presented in
Chapter 10.
3.3. Swiss Methodology for Assessing the Risks of Hazardous Materials

Transportation by Road
Based on article 5 paragraph 2a-d of the Ordinance of April 1, 1991 concerning the
"Protection Against Incidents" (Storfallverordnung, StFV)23 the Swiss Cantons have to
102 CHAPTER 3
estimate the probability of incidents on through-roads occasioning severe injuries to the

population or severe damages to the environment.
Recently the Swiss Federal Office for Environmental Protection, Forestry and
Landscape, BUWALi has issued a "Handbook Ill, to the Regulations Concerning
Incidents, StFV - Guiding principles for Traffic Ways" (December, 1992). This hand-
book is a complement to the Ordinance of April I, 1991.
We shall describe in this section the BUWAL procedure for assessing the risk of
dangerous goods transportation in Switzerland (on through-roads), and discuss the main
topics to be covered by the "detailed risk assessment report" addressed to the competent
authorities (i.e., in this case BUWAL).
Infonnation on Swiss Roads/Highways

Source: CIA, USA (over Internet)
Highways (1992)
62,145 km total (all paved), of which 18,620 km are canton and 1,057 km are national
highways (740 km Autobahn); 42,468 km are communal roads.
Highways (1994)
Total 71,106 km; paved 71,106 km (including 1,502 km of expressways).
In Switzerland, there are still several highways and new roads in construction. As
soon as a longer segment is ready, it is open to the traffic. This explains the differences
found in above mentioned figures, anyway this give an idea of the situation.
3.3.1. SUBDIVISION OF THE ROAD TRACKS INTO ROAD SEGMENTS
In a first step, the cantonal road network is to be represented on a general map. For a
clear presentation of the statements asked for in the short report, the road track should
be subdivided into several road segments. The length of a segment is determined in
such a way, that the architectural and technical configuration, the environment, the
traffic and the safety measures should be homogeneous within the selected segment.
The length of a segment should not be less than a kilometer.
Structural and Technical Aspects

The report must mention briefly the structural and technical aspects of the road/street,
that is to say the design, the road type, the important structural safety measures, and the
traffic control systems. This description is to be performed globally for the chosen
segment, in the case of longer segments, for groups of road segments. The description is
to be supplemented with general maps at a scale of 1 : 25'000 or 1 : 10'000. The starting
and ending points of the road segments must be clearly defined.
Road track: In a general map, the horizontal and vertical layout of the road is to be
described in short, particularly in reference to the disposition of the route in the terrain:
i.e., at ground level, in a cutting, on dam, on a slope. This description is to be completed
• BUWAL: Bundesamt fiir Umwelt, Wald und Landschaft, Bern, Switzerland.

through typical transverse sections. Fu~er crossings, junctions, unusual slopes, or

transverse slopes, narrow curve radius, etc., are to be mentioned.
Road type: Separated lane; amount tracks per direction; bridges, viaducts, galleries and
longer tunnel.
Safety measures: Structural safety measures (buildings of the drainage, emergency
exits, safety galleries of tunnels).
Side-buildings along national roads: Roadside restaurant, gasoline stations, and
adjoining refreshment rooms; restaurants, lodging and rest areas.
Side-buildings along roads: Public parking lots, look-outs with sidings and/or parking
places. The amount of parking lots should be mentioned.
Custom houses and toll-roads: Here the driveways and other connecting roads are to be
indicated, as well as the amount of dispatch lots
3.3.2. INFORMATION AND DATA CONCERNING THE ENVIRONMENT
The data to be recorded are: population density, groundwater resources, surface waters,
as well as other objects worth considering. Furthermore one has to determine an
observation corridor, which shall be, in case of disturbances, subdivided into a close-
range compartment seriously affected, and a far-range compartment much less affected.
The close-range compartment extends 200 m on the left and on the right of the road
sides; the adjacent far-range compartment extends both sides up to 1500 m. The
indicated distances are standard values, which suffice as a rule for the recording of the
en'.'ironmental data as part of the short report. In case of a particular disposition of the
row in the terrain, or due to a particular environment (narrow valleys; groundwater
resources) it may be necessary to adjust these distances according to the circumstances.
For each road segment, the statements about the environment are to be tabulated, and
whenever possible, entered in a general map at a scale of 1: 25'000.
Population Density. The population density is to be indicated for the close- and far-
range along the road, and should mention the number of inhabitants per square
kilometer (Inh.1km2). If there is no data available, than the population density could
possibly be estimated as follows:
urban population density: > 5000 Inh.1km2
small-town population density: 2000 to 5000 Inh.1km2
village population density: 100 to 2000 Inh.1km2
slight or no settlement: < 100 Inh.1km2.
Groundwater Zone. On the basis of the available maps mentioning waters protected
against pollution, all protected groundwater zones are to be recorded in the close-range,
S. For each protected zone the unsaturated zone as well as the flow direction of the
groundwater should be estimated and, if possible, the importance of the groundwater
resource. Likewise, for each protected groundwater zone, one has to indicate the
104 CHAPTER 3
available or planned groundwater deep-wells, and the known or estimated delivery

values of the pump-rooms.
Maps mentioning waters protected against pollution: The cantonal authorities are
in the possession of maps indicating the zones which are to protect against water
pollution. Such maps show the protected groundwater zones S, and the protected waters
areas A, B, and C.
Protected groundwater zone S: The following zones SI, SIT, and SII belong to the
protected zone S, as well as the protected groundwater areas.
Unsaturated zone: Data concerning the distance from the soil surface to the
groundwater table (unsaturated zone) can usually be obtained from the cantonal Offices
of Waters Protection. The permeability of the protective earth layer is to be estimated
roughly in order to calculate the probability value, as stated in the next section.
In the far-range compartment, the protected groundwater zones S, that are located
close to the road must be indicated, in case they could be polluted by waters present in
the close-range along the road, or if they are part of the main drainage channel of the
road segment under consideration. The protected waters zones A can also be mentioned;
the protected waters areas B and C, however, need not be mentioned.
Surface Waters. In the close-range all surface waters (flowing waters and standing
waters) and their distance to the road are to be recorded, or if along the road, they are
part of the main drainage channel of the road segment under consideration.
In the far-range, the surface waters are to be recorded, if they can be polluted
through waters, which lie in the proximity of the road, or are acting as receiving waters
of a main draining segment under consideration. For the flowing waters the yearly
average discharge quantity (m3/s) is to be estimated; whereas for the standing waters it
is the water volume (m3). These data can be taken from the "Hydrological Year-book of
Switzerland" .
The waters are to be described shortly with regard to:
- the direct use for drinking water supply (i.e., direct drawing out the waters)
- the indirect use for drinking water supply (i.e., natural or artificial feeding into a
groundwater stream, from which drinking water will be taken)
the importance as fishing waters.
Objects to be Considered. The objects listed below will only be considered in the
close-range:
• Schools, homes for old-age people, market centers, sport yards, camping yards,
and exhibition halls.
• Cataloged objects that are of national interest, like protected environment/-
landscape, hill moor, pa~ture-Iand, tale moor, and moor landscapes.
• Cataloged large fruit cultures.
• Industrial processing plants with a potential danger.
• Water supply plants.
• Other important traffic ways.
3.3.3. INFORMATION CONCERNING TRAFFIC RISE, TRAFFIC STRUCTURE

AND ACCIDENT HISTORY
The traffic data described in the following section must be determined for each road
segment.
Traffic Rise
The traffic volume is defmed here, as the Average Daily Traffic per 24 hours (ADT-24).
The Heavy Traffic Share (HTS) corresponding to this ADT-24 must be indicated.
Average Daily Traffic per 24 Hours. This is defined as the yearly total of vehicles at a
certain road cross section divided through 365. This can be determined for certain
through-roads from the published statistics of the Swiss traffic counting, which takes
place every five years. Since such counts records catch hold of the ADT-24 data, a
conversion is necessary. If records concerning the average (hourly) traffic of motor
vehicles at day (Nd) and night (Nn ) are available from the cantonal noise-pollution
registers, then the ADT-24 data can be determined from such records.
Heavy Traffic Share of ADT-24 (HTS). The defmition of the heavy traffic includes all
trucks, tractor-trailer units, articulated lorry, and special-purpose vehicles (> 3.5 t total
weight, without touring bus). The Heavy Traffic Share, HTS, can be estimated like
ADT, because of the above- mentioned Swiss traffic counts; or alternatively, the Nd2 and
Nnl data about average traffic at day and night (or, hourly traffic) can be used instead.
These data are available from the cantonal noise pollution-registers concerning motor
vehicle traffic. When corresponding data are lacking, one can apply the Swiss mean
value of 6% to '!8timate the HTS value. This ratio is to be adjusted accordingly by
means of estimates to reflect the local/regional conditions prevailing on a given road
segment.
Traffic Composition
In order to assess the composition of the Swiss traffic, one has to indicate the Share of
Dangerous goods traffic based on the Heavy traffic (SDH) and the ratio of the different
SDR-classesi corresponding to the dangerous goods traffic (RSC).
Share of Dangerous Goods based on the Heavy Traffic (SDH): According to the latest
road traffic survey of the year 1984, the share of the dangerous goods traffic related to
the total heavy traffic amounts to 8% (Swiss average). Depending on road segment and
regional particularities this proportion can vary between 5% and 15%.
, SDR-class: All defined in the Swiss Ordinance on the transportation of dangerous goods on roads. (SDR,
Regelung tiber die schweizerische Beforderung gefabrlicher Gtiter auf der Strasse). See also ADR,
International Ordinance on the Transportation of Dangerous Goods on Roads. (ADR, Regelung tiber die
internationale Beforderung gefabrlicher GUter auf der Strasse).
106 CHAPfER3
Ratio o/the Different SDR-Classes based on the Dangerous Goods Traffic (RSC): For
transit roads, we can assume the following distribution, which is based on the Swiss
average (TABLE 3.10):
TABlE 3.10. Ratio of the different SDR-classes to the dangerous goods traffic (RSC)
SDR-class 1 2 3 4 5 6 7 8 9
RSC 0.001 0.07 0.70 0.07 0.Q1 0.07 - 0.08 -
Depending on the regional aspects (presence of a harbor, stores, loading/unloading
station, chemical plant, or processing plant) the RSC-share corresponding to a given
road segment may be corrected on the basis of estimates.
Accident Statistics
The Accident Rate of the Total traffic is to be considered (ARn. This is to be calculated
according to the Swiss VSS-directive24 (VSS-Norm SN 641'310a). When some accident
statistics are not available, the accident rates corresponding to different road categories
can be selected from the data listed in TABLE 3.11, showing the accident rates, and in
brackets, the confidence limits.
TABlE 3.11. Accident rates for the total traffic (Average values AR-total)
Road Type Accident Rates

Highways 0.45 (± 0.20) x 10-6 , Vehicle· Ian
Semi-Highways 0.50 (± 0.10) x 10-6 , Vehicle·1an
Main roads (outside localities) 1.20 (± 0.40) x 10-6 , Vehicle·1an
Main roads (inside localities) 2.10 (± 0.40) x 10-6 , Vehicle·1an
If statistics concerning Accident Rates for the Heavy traffic (ARB) are available for
the different road segments, they must be used for the calculation. Where this data are
not available, then the accident rate for the total traffic is to be multiplied by a factor
0.5. This is because the share of the heavy traffic accidents to the total traffic accidents
(Swiss average) is approximately hiM as large as the share of the heavy traffic to the
total traffic. In special cases, e.g., strong ramp, this value can be higher and an
appropriate correction must be made.
3.3.4. ESTIMATION OF THE LIKELIHOOD OF AN INCIDENT WITH SEVERE

CONSEQUENCES TO PEOPLE OR ENVIRONMENT
Descrip.tion o/the Method

The goal of this assessment is the identification of road segments with comparatively
high probabilities of severe damage to people and environment for further more
detailed investigation.
The methodology followed by the Swiss Authorities is based on the most actual
national and international knowledge and experience in this field. Due to some
simplifications the method cannot cope equally well with all kinds of situations, like for
instance very long tunnels. The method allows to coarsely assess, for each road
segment, the probability of an incident causing severe damage to people, groundwater
resources, and surface waters, on the basis of representative incident scenarios.
Severe Damages
For this study, the following figures defining severe damages have been adopted:
Population: 10 or more deaths.
Groundwater: the exceeding of the tolerance value of the food ordinance for mineral
oil products, or for greatly water-polluting fluids at a drinking water catchment
in a protected groundwater zone S.
Surface waters:
Running waters: the contamination(a) of a river over a length of 10 km, or
more.
Stagnant waters: the contamination(b) of a lake over a surface of 1 km2 , or
more.
The RSS-factors have to be adapted to local/regional conditions.
Representative Incident Scenarios

"Representative incident scenarios" have to be dermed, which show comparable effects
for similar incidents involving different dangerous goods. These are:
• Severe Damage to thf Population: scenarios involving burning, explosio!l
and release of toxic gases.
• Severe Damages to Groundwater Resources and Surface Waters: Scenarios
involving the release of mineral oil products and greatly water-polluting
fluids.
A reference dangerous product is attributed to each representative incident scenario.
The reference dangerous product stands for the sum of all products of a SDR-class with
comparable effects (for the considered scenario). The relevant products are usually
belonging to one dangerous goods category, eventually two. A quantitatively important
product must be chosen as relevant dangerous product under the listed dangerous
goods.
(a) Here the contamination is defined as the overstepping of a solids loading of a river (including bank zone)
of 15 glm2 through mineral oil products, and/or 5 g/m2 of greatly water-polluting fluids.
(b) Accordingly, it is assumed, that rivers of less than 10 km length, and lakes of less than I km2 area can not
be badly damaged. However, it must be considered that, by the intermediate of these smaller surface waters,
other rivers or lakes can possibly be polluted. In a such situation, one is to decide case wise, whether a
frequency should be determined for the larger river or lake, depending on their size. The RSS-values are to
be eventually adjusted depending on the local realities.
108 CHAPTER 3
Frequency of Incidents with Severe Da11Ulges

The frequency of representative incidents leading to different categories of severe
damages is defined as follows:
• Severe Da11Ulges to the Population: one has to consider the sum of the
frequencies of the incident scenarios burning, explosion and release of toxic
gases.
• For Groundwater Resources, respectively Surface Waters: the respective sum of
the frequencies for the representative incident scenarios "release of mineral oil
products", respectively "greatly water-polluting fluids", which lead to severe
damages.
The frequency (per km and year) of representative incident scenarios leading to severe
damages, is to be determined for each road segment as follows (Eq.3.1):
F, = ADT·365·HTS·AR·SDH·RSC·RRP·RRI·RSS (3.1)
where
ADT: ADT(24) (Le., Average Daily Traffic) is to be converted on a year basis,
i.e., average number of vehicles per year (vhc.lyear); assuming that a
year has 365 days
AR Accident Rate (vhc.lkmr 1
Fs Frequency of a representative incident scenario with severe damages
[(km.yearr 1]
HTS: Heavy Traffic Share based on average daily traffic (ADT-24), (dimen-
sionless)
RRI: Relevant Release Rate, and for burning and explosion, the Ignition rate
(dimensionless)
RRP: Ratio of the Relevant Product of the SDR-class applicable to the repre-
sentative incident scenario (dimensionless)
RSC: Ratio of the different SDR-Classes corresponding to the dangerous
goods traffic (dimensionless)
RSS: Ratio of the representative incident Scenarios leading to Severe
damages (dimensionless)
SDH: Share of Dangerous goods traffic based on the Heavy traffic.
The computation of the traffic and accident factors HTS, AR, SDH, and RSC has
already been given previously.
Ratio of the Relevant Product of the SDR-class Applicable to the Representative

Incident Scenario (RRP)
This RRP-value is expressed as the ratio of the relevant products of the particular SDR-
class representative of a given incident scenario.
Relevant Release Rate and Ignition Rate

It is assumed, that all materials relevant for a representative incident scenario are trans-
ported more or less in similar quantities and containers, so that we can admit a uniform
release rate, and in the case of burning and explosion, that ignition will follow. This
value is valid both for open rail tracks, as well as for tunnels. The particular disadvan-
tages presented by the tunnels, with regards to release and ignition, are largely compen-
sated through the additional safety-measures usually present in tunnels.
The relevant release rates pertaining to each individual representative scenarios are
explained in the respective sections. Foreign information sources2S.26.27.28 and the data
available in the Switzerland served as a basis for the determination of these release and
ignition rates.
Ratio of the Representative Incident Scenarios leading to Severe Damages

The RSS-factor stands for the probability of a severe damage under the condition, that a
relevant release, and for burning and explosion, the ignition has already occurred.
The RSS-value for the population is largely dependent on the traffic density (ADD,
and on the population density.
For the determination of the RSS-value for surface waters, the influence of the
railing, the kind of drainage and the possible ply of the vehicle during the incident (on I
beside the road) were considered. Determining are, however, the distance of the surface
water to the road and the discharge volume of the river.
For surface waters, the possibility of an infiltration in a protected groundwater zone
S is considered additionally for the determination of RSS.
The general safety precaution available today were considered for the determination
of RSS. This counts both for the generally usual structural, technical and organizational
measures. In rare cases, where specially effective safety precautions should be consid-
ered, the Owner has the possibility to adjust the RSS values, which were indicated here.
3.3.5. ASSUMPTIONS CONCERNING THE REPRESENTATIVE INCIDENT

SCENARIO FOR THE "POPULATION"
Representatives Incident Scenario "FIRE"

Reference Substance. GASOLINE was chosen as representative flammable product.
The share of gasoline (incl. similar products) in the SDR-class 3 amounts to 40% (RRP
= O.4).i The other SDR-classes are not playing an important part for the scenario
"FIRE".
RRl: It was assumed a rate of 0.04 for a relevant release (quantity> ca. 1.5 t), and a
probability value of 0.05 for an ignition following such a release. As a result RRl is
taking a value of 0.002.
• The RRP-value is based on data published in the annual report of the "Schweizerische ErdOlvereinigung",
(i.e., the Swiss Petroleum Products Association).
110 CHAPTER 3
RSS: The consequences of a fire extend rarely over the road breadth. As a result it is
the road users that will mostly be affected. The RSS-value (thermal radiation damages)
for road users, corresponding to high ADT-values on open road (0.3), were taken from
the HSC-study and adjuste4 to reflect Swiss conditions. The remaining values for RSS
were derived by this. RSS takes thus, depending on the average daily traffic (AD7) and
population density (PD), one of the values listed in TABLE 3.12.
TABLE 3.12 Delennination of the RSS:factorfor the scenario ''Fire"

(affecting the road users)
ADT Population Density [Peoplelkm2]

in the close range
~2000 <2000
>30000 RSS=0.3 RSS=0.3
IS 000 - 30 000 RSS=0.25 RSS=0.2
5000-15000 RSS=0.15 RSS=O.1
<5000 RSS=0.05 RSS=O.ot
For tunnel other RSS-factors are necessary, since incidents in tunnels lead to
somewhat larger consequences. Additionally, a distinction is to be made between
tunnels comprising one a tube, and tunnels with two tubes because the exposition to
heat radiation is different.
The values for RSS, valid in such a case, are shown in TABLE 3.13.
TABLE 3.13. Delennination of the RSS-faa.ors for the scenario "Fire"
(affecting the road users in a twmel)
ADT Tunnel Type

Single-Tunnel Double- Tunnel
>30000 RSS=O.4 RSS =0.2
15 000 - 30 000 RSS=0.25 RSS=0.15
5000- 15000 RSS=0.15 RSS=O.1
<5000 RSS =0.05 RSS =0.01
Representative Incident Scenario "Explosion"

Reference Substance. "PROPANE" was chosen as representative flammable/explosive
product. The share of "propane" in the SDR-class amounts to 25% (RRP = 0.25).
RRI: A value of 0.002 was adopted for relevant releases (quantities> 2.0 tons) followed
=
by ignition. In this case RRI 0.002.
RISK ANALYSIS OF HAZARDOUS MATERIALS 1RANSPORTS BY ROAD 111
RSS: The effects of an explosion extend further than for a fire. Beside the road users, a
larger part of the population will also be affected. The value for explosion damage
concerning road users, corresponding to high ADT on open road (0.8), were taken from
the HSC-studrs and adjusted to reflect Swiss conditions. The remaining values for RSS
were derived from this value (0.8). RSS takes thus, depending on the average daily
traffic (ADn and population density (PP) one of the values of TABLE 3.14.
TABLE 3.14. Determination of the RSS·fa£tDr for the sa:nario ''Explosion"

(affecting road users and the population)
ADT Population Density [Peopa.Jkmz)

in the close-range
~2000 <2000
>30000 RSS=0.8 RSS=0.8
15 000 - 30 000 RSS = 0.55 RSS = 0.5
5000·15000 RSS=0.3 RSS = 0.2
<5000 RSS=0.15 RSS = 0.05
For "tunnels", the same procedure as for the scenario "fire" was adopted. The
corresponding values can be taken from TABLE 3.15.
TABLE 3.15. Determination of the RSS·fac:tor for the scenario "Explosion"

(affecting road users in a tunnel)
ADT Tunnel Type

~i~g1e- Tunnel Double-Tunnel
>30000 RSS = 0.85 RSS=0.55

15 000 . 30 000 RSS = 0.6 RSS=0.3
5000·15000 RSS = 0.35 RSS=O.1
<5000 RSS=O.1 RSS=0.05
Representatives Incident Scenario "Release o/Toxic Gas"

Reference Substance. CHLORINE was chosen as representative substance for the
scenario involving a leak and release of toxic gases. The share of chlorine (incl. similar
products) in the SDR-class 2 amounts to 15% (RRP 0.15). =
RRI: A rate of 0.001 was assumed for a relevant release (quantity> ca. 2 t). As a result
RRI takes a value of 0.00 1.
112 CHAPTER 3
RSS: The effects of a toxic gas release are concerning the road users, as well as the
residential population. When unfavorable conditions are encountered, one is to include
the far-range, as well, into the risk aspect considerations.
TABLE 3.16. Determ.ination of the RSS-factor for the scenario "Release of Toxic
Gas" (affecting road users and the population)
ADT Population Density [PeopleJkmz]

in the close-range
~2000 <2000
>30000 RSS=0.65 RSS=0.65
15000- 30 000 RSS =0.5 RSS=0.45
5000- 15000 RSS =0.35 RSS=O.3
<5000 RSS=0.25 RSS=0.15
ADT Population Density [PeopleJkmz]
in the close-range
~2000 <2000
>30000 RSS =0.65 RSS=0.6
15 000 - 30 000 RSS= 0.5 RSS=O.4
5000 -15 000 RSS =0.3 RSS=0.2
<5000 RSS=0.15 RSS=0.05
TABLE 3.17. Determination of the RSS-factor for the scenario "Release of Toxic
Gas" (affecting road users in a tunnel)
ADT Tunnel Type

Single-tunnel Double-tunnel
>30000 RSS=0.65 RSS=0.3
15 000 - 30000 RSS=0.35 RSS=0.15
5000- 15000 RSS=0.2 RSS=O.1
<5000 RSS=O.1 RSS=0.05
The RSS-values for toxic gas release concerning road users, corresponding to high
ADT on open road (0.6), were taken from the HSC-studis. The remaining RSS-values
were derived from this study and modified to reflect Swiss conditions. As a
consequence, RSS takes one of the values listed in TABLE 3.16, depending on the
average daily traffic and population density.
Especially with long tunnels, and in particular in urban regions, there is a problem
with aeration exhaust chimneys. This must be considered when computing a modified
value for RSS. The procedure adopted for tunnels was similar to that for fire. The RSS-
values are shown in TABLE 3.17.

SCENARIO FOR "GROUNDWATER CONTAMINATION"
The probability of a incident resulting in severe damages to the groundwater resources

is, as a rule, only to be considered for those protected groundwater zones (and/or
protected groundwater areas), in which wells of importance for public drinking water
supply are present (respectively are foreseen).
Severe damages are then possible only, if the catchment are to be found at
immediate vicinity of the road, or if the groundwater flows from the road in direction of
the catchment.
Representative Incident Scenario "Release of Mineral Oil Products" (Fuel Oil)

Reference Substance: FUEL OIL (for domestic heating) was chosen as reference subs-
tance. The share of "Fuel Oil" of the SDR-class 3 amounts to 100% (RPP = 1.0).
RRI: A rate of 0.004 was assumed for relevant releases of fuel oil (quantity> 10 tons).
Consequently RRI is equal to 0'()04.
RSS: Protected groundwater zones S are usually delineated in a way such that incidents
involving the release of mineral oil products (e.g., fuel oil) outside the protected zone
cannot conduct to a severe contamination of the groundwater close to the catchment
point. Therefore, the probability of an incident leading to severe contamination of the
groundwater resources is only to be estimated for protected groundwater zones S, which
are crossed by the road, or are affected by it
The RSS-factor can be found in TABLE 3.18. Fm the case that particularly efficient
measures have been taken for the prevention of water pollution, then RSS is to modified
accordingly.
TABLE 3.18. Determination of the RSS-factor for the scenario "Release of MineraI Oil
Products" (affecting groundwater)
Thickness So i I Permeability
of Soil Cover small medium high
<2m RSS = 0.05 RSS=0.2 RSS=0.5
2m-10m RSS = 0.01 RSS=0.05 RSS = 0.2
>lOm RSS=O.Q1 RSS =0.Q1 RSS=O.05
Representative Incident Scenario "Release of Strongly Water-Polluting Liquids"

Reference substance: "TETRACHLORETHEN" was chosen as representative reference
substance. The proportion of tetrachlorethen in the SDR-class 6 amounts to 20% (RPP=
0.2).
114 CHAPTER 3
TABLE 3.19. Determination of the RSS-facror for the scenario "Release of Liquids Potentially
Very Noxious to Warer"
Thickness of Soil Permeability

Soil Cover Low I Medium I High
Distance from the protected zone to the road: <50m
<2m RSS =0.2 RSS=0.5 RSS= 1.0
2m-10m RSS=0.05 RSS=0.2 RSS=0.8
>lOm RSS=O.Q1 RSS =0.05 RSS=0.5
Distance from the protected zone 10 the road: 50 m to 200 m
<2m RSS=O.Ol RSS =0.05 RSS=O.l
2m-10m RSS=O.Ol RSS =0.Q1 RSS =0.05
>lOm RSS=O.Q1 RSS =0.01 RSS=O.Ol
RRl: A mte of 0.02 was assumed for the release of relevant quantities of tetrachlorethen
(quantity> ca. 2 t). In this case RRI is equal to 0.02.
RSS: The RSS-factors are listed in TABLE 3.19. For the case that the road is built as a
an impervious sump, a coefficient for RSS of 0.01 should be used.
3.3.7. ASSUMPTIONS CONCERNING THE REPRESENTATIVE SCENARIO FOR

"SURFACEWATERS"
Representative Incident Scenario "Release of Mineral Oil Products"

Reference Substance: FUEL OIL (for domestic heating) was chosen as reference
substance. The share of "Fuel Oil" of the SDR-c1ass 3 amounts to 100% (RPP 1.0). =
A rate of 0.004 was assumed for a relevant release of fuel oil
RRl (for flowing waters):
(quantity> ca 10 t). As a result RRl(FIowing Walen) has a value of 0.004.
RRl (for lakes): Severe damages to lakes (as defined previously) can not be achieved
with quantities smaller than 50 tons. Since the release of a fuel oil quantity of 50 tons or
more on a road is practically impossible, this scenario is irrelevant, and does not apply
to the pollution of a lake.
RSS: The value of the RSS-factor for flowing waters is shown in TABLE 3.20.
Petroleum products can reach some flowing waters independent from the distance
to the road, if no oil trap, and catchment basin are present Rivers acting as main
dminage channel, with a flowing quantity of 10-75 m3/s, are especially endangered. In
such cases a value of 0.2 shall be used for RSS without evident inflltration, or a value
of 0.3 for RSS with proven infiltration.
TABLE 3.20. Determination of the RSS-factor for the scenario "release of mineral oil products"
(affecting surface waters)
Discharge Rate Distance from the Road

(m3/s) <SOm I SOm-200m
Without evident infiltration
10-75 RSS=0.4 RSS=O.1
75 - 125 RSS = 0.2 RSS=0.05
> 125 RSS=O.1 RSS = 0.Ql
With proven infiltration
10-75 RSS=0.5 RSS=0.15
75 - 125 RSS=0.3 RSS=O.1
> 125 RSS=0.3 RSS=O.1
Representative Incident Scenario "Release of Strongly Water-Polluting Liquids "

Reference Substance. "TETRACHLORETHEN" was chosen as representative product
The share of tetrachlorethen in the SDR-class 6 amounts to 20% (ARS =0.2).
TABLE 3.21. Determination of the RSS-factor for the scenario "release of liquids potentially
very noxious to water" (affecting flowing waters)
Discharge Quantity Distance from the Road

(m3/s) <SOm I SOm-200m
10-75 RSS=O.4 RSS=O.1
75 - 125 =
RSS 0.2 RSS=0.05
> 125 RSS=O.1 RSS = 0.Ql
10-75 RSS=0.5 RSS=0.15
75 - 125 =
RSS 0.3 RSS=O.1
> 125 =
RSS 0.3 RSS=O.1
RRI (jor flowing waters): A rate of 0.02 was assumed for a relevant release of
CHLORETHEN (quantity> ca. 2 t). As a result RRI(Flowing Watenr) has a value of 0.02.
RRI (for lakes): Quantities smaller than 5 tons are not going to cause severe damage to
lakes. A rate of 0.005 was assumed for a relevant release (quantity> ca. 5 t). As a
result RRl(lAke.) has a value of 0.005.
RSS: The value of the RSS-factor for flowing waters is shown in TABLE 3.21, and for
lakes in TABLE 3.22.
116 CHAPTER 3
TABLE 3.22. Determination of the RSS-factor for the scenario

"release of liquids potentially very noxious to water (affecting lakes)
Distance from the Road

<SOm I 50m-200m
RSS = 0.05
I RSS=O.01
RSS=0.2 I RSS=0.05
Strongly water-polluting liquids can, by the intermediate of a canalization, reach

some flowing waters independent from the distance to the road, if no catchment basin
is present. Rivers acting as main drainage channel, with a flowing quantity of 10-75
m 3/s, are especially endangered. In such cases a value of 0.3 shall be used for RSS
without evident infiltration, or a value of 0.4 for RSS with proven infiltration.
3.4. New Developments in Traffic and Vehicle Control
Traffic control systems, such as the railway signaling system and the traffic light control
system are basically complex distributed control systems.. These large scale systems
and networks are made up from a hierarchically structured subsystems which will
interact with one another. In order to control such highly complex systems, conven-
tional mechanisms have become inefficient and impractical. The reliance on sophisti-
cated computer control systems to manage such systems and networks is becoming
increasingly essential in order to match with ever growing demands and needs. New
developments in traffic safety management, software and computer system for vehicle
and traffic control are emerging.
The Research Programs DRIVE I PROMETHEUS

The DRIVE project is being carried out by a consortium consisting of the Institute of
Traffic Safety, rov Rheinland in Cologne (Germany); Program Validation Limited, and
Safety Critical Computing Group of Leeds University, both in the UK; Department of
Vehicle Dynamics, TNO Road Vehicle Research Institute in Delft (The Netherlands).
DRIVE is a pre-competitive and pre-normative research and development program which
is funded by the Commission of the European Communities (CEC), and that aims to secure
convergence between CEC policies on road transport and information technology and
telecommunications. This combination has been termed Road Transport Informatics (RTf).
DRIVE complements the industry research program PROMETHEUS. 29
Figure 3.3. The Research programs DRIVE I PROMETHEUS.
The DRIVE program is investigating and coordinating the development of standards

for software and computer systems in vehicles and in traffic management systems. It is
hoped that a viable solution to current and foreseeable problems in European road
transport can be found. A major assumption that has influenced the direction of the
DRIVE project is the idea that whereas we would appear to be faced with an unlimited
demand to use the road networks by both private vehicles and increasingly by inter-
national freight, there is a limit, due to both political and environmental pressures, to
the number of new roads that can be built in Europe. A solution to this problem is seen
in the implementation of a new and effective information systems, navigation systems
and traffic management systems.
Within these three areas the following ideas are being seriously considered:
• Use of artificial intelligence.
• Automatic or semi-automatic vehicle control systems (so-called intelligent co-
pilot).
• Lane and positioning control and collision avoidance systems.
• Road usage and access control.
In view of the problematic of alternative routing in some parts of the country, the
implementing of vehicle and traffic control systems shall definitely help improving
traffic safety and reduce the number of accidents involving trucks or road-tankers.
Summary
This chapter introduces the methodological framework for performing a risk analysis of
hazardous material transportation by road in the context of regional risk assessment and
safety management. A large variety of .info~tion and knowledge about truck accidents
and associated consequences are needed for complex transportation risk analysis.
Special attention is given to the risk assessment of the transportation of hazanlous
materials through road tunnels: examples include data and cases from UK, USA, as well
as from Switzerland The Swiss methodology for assessing the risks of hazanlous
materials transportation by road includes the need for the subdivision of the road tracks
into road segments, various information and data concerning the environment such as
traffic rise, traffic structure and accident happening. It allows the estimation of the
likelihood of an accident with severe consequences to people and environment.
Scenario generation for risk assessment is of a fundamental importance; various
examples on how to design risk scenarios are included in this chapter (e.g., scenario for
surface waters, for release of mineral oil products, groundwater contamination, etc.).
118 CHAPTER 3
1 Leimkuhler, F.F.; Karsen, J.M.; Thompson, J.T., Statistical Analysis of the Frequency of the Severity of
Accidents to Potential Highway Carriers of Highly Radioactive Materials, Report No. NY0-9771, U.S.
Atomic Energy Commission, (1961).
2 Brobst, W.A., Transportation accidents: how probable? Nuclear News 16(5), (1973), pp. 48-54.
3 USAEC (U.S. Atomic Energy Commission). Environmental Smvey of Transportation of Radioactive
Materials to and from Nuclear Power Plants, WASH-1238, (December 1972).
4 Harwood, D.W.; Russel, E.R., Present Practices of Highway Transportation of Hazardous Materials. Report
FAWH-RD-89-013, U.S. Department ofTransportation, (May 1990).
S Hobeika, A.G.; Kim, S., Databases and needs for risk assessment of hazardous material shipments by
trucks, in Proceedings of Hazmat Transport '91 - A National Conference on the Transportation of
Hazardous Materials and wastes, Northwestern University, Evanston, Illinois, (1991).
6 Smith, R.N.; Wilmot, E.L., Truck Accident and Fatality Rates Calculated from California Highway
Accident Statistics for 1980 and 1981, Report SAND82-7066, Sandia National Laboratories, (1982).
7 Abkowitz, M.; Eiger, A.; Srinivasan, S., Assessing the Release and Costs Associated with Truck Transport
of Hazardous wastes, Report PB84-224468, U.S. Environmental Protection Agency, (1984).
8 Graf, VD.; Archuleta, K., Truck Accidents by Classification, Report FHWAlCAffE-85, U.S. Department
of Transportation, (1985).
9 Jovanis, P.P.; Chang, H.L.; Zabaneh, I., A comparison of accident rates for two truck configurations, in
Transportation Research Record 1249, Transportation Research Board, National Research Council,
Washington, OC, (1989), pp. 18-29.
10 Harwood, D.W.; Russel, E.R., Present practice of highway transportation of hazardous materials, FHWA-
RD-89-Ot 3, U.S. Department of Transportation, (1990).
11 Davies, P.A.; Lees, F.P., Impact speed of heavy goods vehicles, J. Hazard. Mater. 26 (1991), pp. 213-217.
12 Krasner, L.M., Motor Vehicle Standards for Hazardous Material Transportation, National Highway Safety
Bureau, Washington, OC, (Jan. 1970).
13 Jones, G.P., et aI., Risk Analysis of Hazardous Material Transportation, Vol. I, Report TES-20-73-4-I, U.S.
Department of Transportation, (1973).
14 Rhoads, R.E., et aI., An Assessment of tl>e Risks of Transporting Gasoline by Truck, Report PNL-2133,
Battelle Pacific Northwest Laboratory, Richland, Wash., (Nov. 1978).
IS Stewart, A.M.; Van Aerde, M., An empmcal analysis of Canadian gasoline and LPG truck releases, J.
Hazard. Mater. 25 (1990), pp. 205-217.
16 Considine, M., Risk assessment of the transportation of hazardous substances through road tunnels, in
Recent Advances in Hazardous Materials Transportation Research, An International Exchange, Confer-
ence of Lake Buena Vista, Florida, November 10-13, 1985. State of the Art Report 3, Transportation
Research Board, National Research Council, Washington, OC, (1986).
17 Transport Statistics Great Britain 1971-1981, Department of Transport, Her Majesty's Stationary Office,
London, England, (April 15, 1977).
18 Hills, P.J., The Relative Dangers of Carrying Conventional Fuels in Britain by Inland Service Modes, Oyez
International Business Communications Seminar on the Risk of Fuel Transport, London, England, (June
16-17, 1982).
19 Monitoring the Movement of Hazardous Freight by Road, Research Report 45, Transport Operations
Research Group, University of Newcastle upon Tyne, England, (1982).
20 Maclean, AD., 1980 Chemical Incidents Survey: Statistical Analysis, Home Office Scientific Advisory
Branch, Report 4181, United Kingdom, (March 1981).

21 Egilsrud, Ph., FHWA Report: Prevention and control of highway tunnel fires, in Recent Advances in
Hazardous Materials Transportation Research, An International Exchange, Conference of Lake Buena
Vista, Florida, November 10-13, 1985. State of the Art Report 3, Transportation Research Board,
National Research Council, Washington, OC, (1986).
22 MORlKA: Mobile Risiken Kt Aargau, Nl Baregg-Tunnel, Sicherheitsanalyse Transport gef8hrlicher
GUter. Gruner AG, Ingenieurunternehmung, Basel. Auftraggeber: Kantonales Laboratorium, Gesund-
heitsdepartement des Kt. Aargau, Schweiz, (30. September 1991).
23 BUWAL, Handbuch m zur Sttirfallverordnung, Richtlinien fUr Verkehrswege, Bundesamt fUr Umwelt,
Wald und Landschaft, Bern, Switzerland, (Dezember 1992).
1A VSS: Analyse der StrassenverkehrsunfaIle, Schweizer Norm SN 64I'31Oa (1990), Vereinigung Schwei-
zerischer Strassenfachleute (VSS).
2S Health & Safety Commission (HSC): Major hazard aspects of the transport of dangerous substances, Report
and appendices, HMSO, London, (1991).
26 Inter-Agency Program on the Assessment and Management of Health and Environmental Risks from
Energy and Other Complex Industrial Systems: Manual for the classification and prioritization of risks
from major accidents in process and related industries, International Atomic Energy Agency, Vienna,
(199\).
27 U.S. Federal Emergency Management Agency, U.S. Dep. of Transportation, U.S. Environmental Protection
Agency: Handbook of chemical hazard analysis procedures, Washington IX, (1989).
28 Fermaud, C.H.; Bohnenblust, H.; Biihlmann, B., Gewasserschutz und Transport gefahrlicher Gilter, Risiko-
analyse fiir die Nationalstrassen des Kantons Uri, SIA 46191, (1991).
29 Williams, M, "PROMETHEUS is Rolling", Colloquium on the Car and its Environment - What DRIVE
and PROMETHEUS have to offer, lEE, (January 1990).
CHAPI'ER4
RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPORTATION

BY RAIL
In this chapter we are first reporting the findings on train accidents experienced in the
USA and Great Britain. Then we present information concerning the situation in
Switzerland, and the Swiss Federal Rail Networ~ (SBB/CFFIFFS) is briefly described.
The methodology proposed by BUWALii for assessing the risks of hazardous materials
transportation by rail is explained in details, and finally, selected results of a risk
assessment study prepared by SBB, that concern incident frequencies and the indicators
used to represent the extent of expected potential damages are presented at the end of
this chapter.
4.1. U.S. Experience with Train Accidents
4.1.1. RAIL ACCIDENT RATES
Train accident rates are more straightforward to obtain than those for trucks, because
there is usually one source of accident data, i.e., the railway companies collecting their
own accident data pertaining to a given railroad-network. In the USA the accident
reports must be submitted to the Federal Railroad Administration (FRA) by the
different railway companies and are then recorded in the RAIRS database (Railroad
Accident/incident Reporting System). Accident reports must be filled in when (a) there
is any impact with rail equipment and road users at a rail-road, respectively rail-
highway crossing, (b) the opemtion of the railroad results in death or injury, and (c) any
accident results in damages exceeding the reportable threshold.
The Association of American Railroad (AAR) also maintains a database (TRAIN IT)
on railcar movements. Actually 90% of the railcars are included, and 100% of their
movements and the commodity being carried are recorded.
A good source of information is the Accidentllncident Bulletin published annually
by the FRA Office of Safety (USA). TABLE 4.1, showing train accidents by track class,
has been taken from this FRA Bulletin. This organization publishes also an annual Rail-
Highway Crossing Accidentllncident and Inventory Bulletin.
For the calendar year 1991, the number of train miles amounted to a total of 5,768 x
108 , of which 0.88 x 108 for yard switching. An overall accident mte of 4.6 x lO-6/train-
mile (excluding gmde crossing accidents) can be computed for the same period, and an
i SBB: Schweizerische Bundesbahnen; CW: Chemins de fer fedCraux suisses; fFS: Ferrovie federali
svizzere (i.e., the Swiss Federal Railways).
h BUWAL: Bundesamt fUr Umwelt, Wald und Landschaft, Bern, Switzerland, i.e., Swiss Federal Office for
Environment, Forestry and Landscape.
122 CHAP1ER4
accident rate of 14.4 x 1O-6/train-mile is obtained for yard switching. The accident rate
without the yard or crossing values resulted in an overall rate of2.4 x 10-61 train-mile
for 1991.
TABLE 4.1. U.S. train accidents by traclc class, Year 1991
Track Total Collisions Derail- Other Track Equipment Human Other

Class Accidents ments Accidents Caused Failure Factors Causes
? 235 28 166 41 100 11 83 41
I 1,291 130 936 225 517 94 510 170
2 443 32 324 87 178 58 122 85
3 342 28 234 80 84 98 71 89
4 384 29 213 142 67 102 81 134
5 90 5 63 22 14 34 18 24
6 29 9 0 20 8 18 2 1
Total 2,814 261 1,936 617 968 415 887 544
SkKHce:~I(I992)
The AAR and the RPI (Railway Progress Institute), and the Chemical Manufacturers
Association (CMA) are jointly developing a hazardous materials transportation risk
analysis model. The accident cause module will include the effects of accident type
(collision, or noncollision), train speed. track class, railroad type, train length, ear
placement within the train, tank ear type, and car protective features (TABLE 4.1).
Rail Cars Involved in a Fire

Approximately 1% of the collision and derailment accidents, in the U.S., resulted in a
fire. In reporting train fire incidents it is important to notice where the car(s) of interest
are placed within the train with respect to the fire location. In 1972, about half of the
rail cars in which fire occurred were involved in derailment or collisions affecting an
average number of 10 cars. The other half of the cars were involved in other type of
accidents in which only one ear was involved. An important consideration in the
analysis of rail tank cars involved in a fire is whether the tank car is oriented so that
liquid or vapor is released from the valves. The U.S. data show that at least 62% of all
collisions and derailments involve ear rollover. The other half of fire accidents are one-
ear accidents, with the ear standing probably upright on the track. The following data
were recommended for risk consideration:
- Fraction of U.S. rail cars upright in a fire: 0.69
- Fraction of U.S. rail cars overturned in a fire: 0.31.
A severity model for train fire duration should be based on a statistical analysis of
assumed probability distributions for fire sources, type and amounts of combustible
materials, and fire-fightiog efforts.
Among the possible fire sources, there are: car heating units, brake-shoe sparks, hot
boxes, external fire sources, spontaneous ignition.
RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPORTS BY RAIL 123
Further, the risk scenario has to consider the type of car involved (empty, loaded),
the availability of combustibles (type, amount), the probabilistic involvement of cargo
and of adjacent cars. From these data the bum rates can be calculated, and the resulting
fire duration estimated, taking also into consideration the fire-fighting efforts.
Estimation of Impact and Crush Forces

The train accident scenarios should consider the following risk components with their
respective probabilities of occurrence (Figure 4.1):
•••
Accident on rail-highway grade crossing
Derailment without collision
Collision
• with ctU'S remaining 011 trtu:k
• with del'tlilmeflt of ctU'S following the collision
0 over bridge, with car falling
- into water
- on clay, silt
- on hard soil, soft rock, or concrete
- on hard rock
- rail bed, roadbed
0 over embankment, with car falling
- into drain ditch
- on clay, silt
- on hard soil, soft rock
- on hard rock
0 into slope
- on clay, silt
- on hard soil, soft rock
- on hard rock
0 into structure
- column
- abutment
- other structures
0 rollover
... rollover followed by collision
- with locomotive
- with car
- with coupler
... rollover without collision
- on roadbed
- on earth
• Other causes
Figure 4.1. Train accident scenario used for estimating impact and crush fon:es.
124 CHAPfER4
Estimating the Mechanical Damages to Rail Containers

In order to determine the impact force magnitude, a statistical analysis for both
derailment and collision data must be performed. It is necessary to compile the collision
and derailment speed distribution considering different cases (routine, over bridges,
over embankments). The effect of having two or more cars containing hazardous
material of interest involved in an accident is, however, difficult to estimate. Impact
force models, given a derailment speed, usually calculate the effective velocity changes
as a linear distribution from 100% to zero to account for hardness of surface struck.
angle of impact, etc. A distribution must also be assumed for impact angle and
container orientation when impact occurs. Such models are considered conservative,
because they assume that the average will be one-half of the maximum available impact
velocity. Denis et al. 2, (1978) have calculated the cumulative probability distribution of
the impact velocity magnitude for large containers in U.S. train accidents.
In order to determine the conditional probability of crush, the risk analyst has to
determine the ratio of the potential crushing area (for a given rail car type) to the
estimated area over which cars could be scattered as a result of derailment or collision.
Given a train accident, the conditional probability of crush (for U.S. conditions) was
estimated as 2 x 10-3• Assuming a probability distribution for the location of the tank or
rail container, and using a simple beam model, it is then possible to determine the
accident severity.
Puncture force rrwdels usually assume the railcar coupler to be the puncture probe.
The model consists of determining the minimum railcar coupler velocity sufficient to
puncture a container as a function of the wall thickness of the container. This critical
velocity is combined with the probability distribution of the pre-accident speed, which
is assumed to be equal to the relative velocity between two cars.
TABLE 4.2. Probability of large container puncture during
rail accidents
Container Probability of Wall

Wall Thickness Puncture
(in.) (per train accident)
0.4375 7.41 x 10-4

0.50 6.90 x 10-4
0.75 5.85 x 10-4
1.00 4.90 x 10-4
1.25 4.18 x 10-4
1.50 3.37 x 10-4
1.75 2.43 x 10-4
2.00 1.52 X 10-4
2.50 3.15 x 10-5
3.00 4.70 x 10-6
4.00 5.54 X 10-8
5.00 9.14x 10-12
Source: Denis e\ aI. (1978)
The number of tank car failures from puncture can be calculated as follows:
(4.1)
where
Np =
Number of tank car failures from puncture, on a yearly basis
Mr =
Rail car mileage per year (mile/yr, or km/yr)
Fa = Accident frequency per km, or mile
Ppe =
Probability of a puncture environment
PPI =
Probability of puncture failure
The probability of large container puncture during rail accident has been estimated
by Denis et al. (1978) for U.S. railroads and is given in TABLE 4.2.
Clarke et al. 3 (1976) have analyzed the severity of transport accidents for a truck
trailer or railcar with a load of small radioactive material containers. It is worth
mentioning that the analysis should be also applicable to typical hazardous material
containers, e.g., 55-gal drums. The original data used by Clarke et al. (1976) and Denis
et al. (1978) can be validated by comparison with truck and rail data, as reported by
Fischer4 et al. (1987) and Harwood and Russel (1990). s
4.1.2. BULK TRANSPORT BY TRAIN
Certain dangerous substances, which are transported by rail in large quantities, have the
potential to cause situations presenting a major hazard of fife, explosion, or toxic
release: in particularly the possibility of killing large number of people in one event
For a risk assessment of transports by rail, a full consideration of all relevant dangerous
substances would require assessing a wide range of some hundred substances, which is
not feasible at a reasonable C,"lst The first priority is to concentrate on those substa.lces
which seem to present the greatest potential hazard, either because of the large bulk
transported (e.g., automotive gasoline, LPG), or because of the inherent hazard of the
substance (e.g., chlorine, ammonia).
The most common train accident is a derailment, in which a rail car separates from
the preceding car and leaves the track. A number of cars behind the lead car follow that
lead car. Train analyses is more complicated than truck analyses because the probability
that the derailed train segment includes the hazardous material of interest must be
considered.
In the early 1970s, Brobst6 (USAEC7 , 1972) developed a scheme for categorizing
highway and rail accident severity in terms of impact speed and fire duration. Brobst's
values for accident frequencies associated with the US accident severity subcategories
(USAEC 1972), which were based on Leimkuhler's speed distribution, have been
compiled in TABLE 3.1 (see Chapter 3.1).
Motor Gasoline
In Great Britain, motor spirit is transported in tank wagons with typical capacity either
32 or 75 tons, with nominal minimum shell thickness of 6 mm. In the period 1980-
1986, HSC 8 (1991) mention in their report that in 80 cases the spillage or leakage was
due to maloperation or failure of the tank wagon equipment and involved release of
126 CHAPTER 4
only small volumes of motor spirit. The other 4 incidents involved substantial spillage
as a result of derailment (3 events) or collision (one event). In one case 64,000 liters of
motor gasoline was lost.
Liquefied Gases
HSC (1991) mention in their reportS that, as far as they are aware, no liquefied gas tank
wagon has ever been punctured on railways in Britain. In the period 1980-1986 there
has been no sustained leak of LPG and no sustained leaks of toxic gases (ammonia,
etc.) .
Liquid Chlorine. The analysis of the risk arising from the transport of liquefied gases,
in Great Britain, has been reported by Purdy et al. 9 These authors describe the models
produced to analyze the level of societal risks arising out of the transport by rail of
chlorine.
Procedure for 1M Risk Assessment of the Rail Trtuu:portation of liquid ClIlorine in

Finltmd
The procedure comprises the following tasks:
• study of accident records
• classification of railway accidents
• determination of the frequency of key accidents
• study of vessel damlJges
• determination of leakage categories
• estimation of leakage probability in key accidents
• determination of leakage probability per transport
• intoxication risk assessment
Figure 4.2. Risk assessment approach for rail transportation of liquid chlorine in Finland.
Railroad Accident ScelUU"ios

a) Accidents occurring in railway stations were categorized as folloWS:
- collision of trains at stations
- damage during switching
- leakage in a stationary tanker
- loading and unloading incidents
b) line accidents:
- collision of trains on line
- derailment
- level crossing accidents
- leakage in a moving tanker
Figure 4.3. Classification of railroad accidents.
Another assessment has been performed on the toxication risk of the population due
to bulk transportation of liquid chlorine by rail in Finland. 10 For this study fourteen rail
accident types were selected and the leakage probability was estimated for each type
separately (Figure 4.3).
In Finland the annual bulk. transportation of chlorine was about 150,000 ton by rail
and 200 ton by road in 1971. Trains hauling loaded chlorine tankers run about 500,000
kIn annually. There is on average one chlorine rail car in every 40 goods trains or one of
100 train is carrying loaded chlorine tankers. There are two size of tankers: 19-20 t and
45-50 t load capacity. The tankers are in Finland equipped with safety walls and jump-
over restraints.
The following information was collected: statistics with indication of number of
tracks, running times, switching times and places, number of crossing trains, number of
level crossings, density of motor traffic at the vicinity of such level crossings, etc. The
classification retained for railroad accidents was as shown in Figure 4.3.
Fourteen typical cases potentially affecting the integrity of the pressure tank were
retained for this study. Chlorine leakage were divided into four categories according to
the size of the vent and the liquid flow rates were estimated at 20 °C (TABLE 4.3).
TABLE 4.3. Estimation of chlorine leakage from rail tankers
Damage Type Aperture III Typical Flow Rate

(mm) (kg 5.1)
- valve leakage 3 0.2 -0.3

- broken valve 10 2-3
- medium breaks, 30 30 (in 25 min)
punctures
- large breaks 100 300 (in 2.5 min)
Note: The maximum leakage daration is given for a 45 t tanlcer.
Co"clw;wllS
The conclusions from the study on bulk transportation of liquid chlorine by rail in Finland,
show that:
• Derailment on the rail track does not represent a big /uwlrd.
• The leakage probability was highest for station accident involving a rail tanker and a
locomotive.
• Of the damage mechanisms considered, (i.e., valve group damages), impacts of general
type were found to dominate the leakage probabilities. The puncture mode was considered
relatively improbable as the safety equipment was assumed to protect the tank ends
effectively at collision speeds up to 40 kmIh. Head puncture by buffers and couplers of
adjoining wagons is generally the most probable damage type for unshielded tankers.
• In the smallest leakage category, random valve failure playa dominant role. The leakage
probabilities calculatedfor the Finish study can be taken from the reference paper.
128 CHAPI'ER4
4.2. British Experience with Train Accidents
4.2.1. MAIN CAUSES OF SPILLAGE
An analysis of the British rail accident data regarding dangerous goods movements has
shown that there were two main causes of spillage: puncture or damage to the tank
wagon due to collision or derailment; and failure or maloperation of the tank wagon
equipment Spillage of motor spirit has resulted from both causes, but neither cause has
resulted in a sustained leak of chlorine, ammonia or LPG in GB from rail tankers.
In the British study the frequency of puncture of motor spirit tank wagons was
estimated from historical data, but leaks via tank wagon equipment were ignored. For
liquefied gases, Great Britain has experienced incident-free records over the last 16
years, thus the puncture frequencies were estimated using historical experience and
expert engineering judgment, and the equipment failure frequencies estimated using
fault tree analysis.
The leaks of liquefied gases due to tank wagon equipment failure were modeled as
two-phase releases from the appropriate valve. In the absence of any actual incident
data, the following assumptions were made:
a) For motor spirit tank wagons: it was assumed that 10% of the punctures would
result in an instantaneous release of the entire contents and the remainder would
be continuous spills from a hole with an equivalent diameter of 100 mm.
b) For liquefied gas tank wagons: it was assumed that 90% of the puncture events
would result in a crack or hole with equivalent diameter of 50 mm, and the
remainder wl)uld proceed to catastrophic failure. The likelihood and timing of
ignition of flammable releases was estimated using expert judgment based on the
historical incident data for flammable liquids.
The estimation of the hazardous consequences of spillage of dangerous substances
due to rail incidents was performed according to a methodology developed by HSFJ
HSCi .
a) It was assumed that spillage of motor spirit would result in a pool, which if
ignited would cause the deaths of all those within the areas of the resultant pool
fire.
b) It was assumed that spillage or leakage of LPG, if ignited, could result in the
following types of events: torch flame, BLEVEn fireball, vapor cloud explosion
or flash fire. The likelihood of each was determined by event tree.
c) RISKAT tool (developed by HSE) was used for assessing the risk of the
dispersion of a toxic gas for different weather conditions and scenarios. The area
within which people would be exposed to a specific toxic load was estimated and
this was related to the probability that a certain proportion of the exposed
I HSE: Health & Safety Executive, London; HSC: Health & Safety Commission, London.
iI BLEVE: Boiling Liquid Expanding Vapor Explosion.
RISK ASSESSMENT OF HAZARDOUS MATERIALS lRANSPORTS BY RAIL 129
population would be killed. Account was taken of the reduction in risk for
people being or going indoors.
The societal risk was estimated from the frequency of a particular event and the
number of people likely to be within the area where fatal consequences may be ex-
pected. The variation in trackside popUlation density along the route was considered,
and the risk to passengers in other trains along these lines was also estimated, this for
each group of dangerous products and main transport route. Finally, the national
societal risk from each group of substances was then estimated for both off-rail
residents and passengers, using these population densities, passengers movements, and
the total yearly movement of each substances. The aggregation of off-rail and passen-
ger risk is a complex process, and cannot always be produced by simple addition.
TABLE 4.4. Freight train closing speeds in collisions with rolling

stock (sample size = 20)
Closing Speed No. of Proportion of

(mph) Collisions Collisions
1-10 1 5
11-20 8 40
21-30 4 20
31-40 3 15
41-50 3 15
51-60 1 5
Note: 1 mph (mde per hour) =1.6 kmJhr.
TABLE 4.5. Derailment speeds of freight trains (sample size 300)
Proportion of On plain truck Not plain

derailment truck
1-10 15 54
11-20 20 20
21-30 6 6
31-40 18 4
41-50 26 4
51-60 9 4
61-70 3 4
71-80 2 4
81-90 1 0
130 CHAPTER 4
4.2.2. ACCIDENT SPEED OF FREIGHT TRAINS
Hazard assessment of the transport of hazardous materials sometimes creates a require-

ment for data on the speed of trains at the moment of collision or derailment. Data are
presented here on the distribution of accident speed of freight trains in Great Britain. 11
The data include head-on, front-rear and rear-front collisions with passenger trains
(55%), freight trains (25%), and miscellaneous rolling stock (20%). These data are
considered to be biased sample, because reports on such accidents tend to deal with
those accidents in which there are extensive casualties and/or property damage. The
data are taken from a study of 300 derailments. Results from this study are shown in
TABLE 4.4 and TABLE 4.5.
4.2.3. 'RESULTS FROM THE HSC-STUDY
Mbtor Spirit Puncture Frequency

Motor spirit is transported in tank wagons with typical capacity either 32 or 75 t with
nominal minimum tank shell thickness of 6 mm. Puncture frequency for GB was
determined to be 6.3 X 10-8 per tank wagon km. The upper and lower bound frequency
values for a Poisson distribution. The symmetrical 90% confidence limits are given in
TABLE 4.6.
TABLE 4.6. Puncture frequency for tank wagons transporting motor spirit
Upper Bound Puncture Frequency 1.4 X 10-7 per tank wagon kIn
Lower Bound Puncture Frequency 2.2 x 10-8 per tank wagon kIn
Mean Puncture Frequency Value 6.. 2 x 10-8 per tank wagon kIn
Liquefied Gases Puncture Frequency

The puncture frequency was estimated by dividing the puncture probability by the
distance traveled by all laden tank wagons in the period of investigation (TABLE 4.7).
TABLE 4.7. Puncture frequency for tank wagons transported liquefied gases
Ammonia and LPG Puncture Frequency 2_5 X 10-9 per tank wagon kIn
Chlorine Puncture Frequency 0.9 x 10-9 per tank wagon kIn
Tank Wagon Equipment Failure Frequency

The incidents of concern are failure or maloperation of the equipment, valves, gaskets,
etc. on the tank wagon. The frequency of sustained small leaks per typical 10 he journey
was estimated to be as indicated in TABLE 4.8.
TABLE 4.8. Failure frequency of tank wagons
Chlorine Small Leak Frequency 3.1 X 10-8 per journey

Ammonia Small Leak: Frequency 1.3 x 10-8 per journey
LPG Small Leak Frequency 8.3 x 10-9 per journey
Using 1985 movements data the predicted frequencies of puncture and tank wagon
equipment failures per year, for Great Britain, were estimated to take the values of
TABLE 4.9.
TABLE 4.9. Estimated puncture and tank wagon equipment failures, valid for Great
Britain
Incldent Frequency (per year)
Puncture Equipment
Failure Small
Leaks
Ammonia 3.4 x 10-3 5.9 x 10-5
Chlorine 2.8 x 10-4 7.3 x 10-5
LPG 3.5 X 10-3 1.1 X 10-4
All Liquefied Gases 7.2 x 10-3 2.4 X 10-4
Motor Spirit 630 x 10-3 not evaluated as

consequences
judged small
(AfterHSC,I99l)
Ignition Probabilities for LPG and Motor Spirit

There have been no incidents in Britain involving a sustained large leak of LPG from a
rail tank. wagon. Consequently it was assumed that, should a leak occur, the immediate
ignition probabilities will be the same as those for motor spirit (TABLE 4.10).
TABLE 4.10. Ignition probabilities for LPG and motor spirit, valid for Great Britain
Dangerous Ignition Type Small Spill Large Spill

Substance
LPG Immediate 0.1 0.2
LPG Delayed 0 0.5
LPG None 0.9 0.3
Motor Spirit Immediate 0.1 t 0.2 t
Motor Spirit Delayed 0 0.1

Motor Spirit None 0.9' 0.7
.. ...
Note. The IgnItion probablbtles mdlcated by T were denved from actual mCldents, the remmrung
..
values were estimated statistically (after HSC, 1991).
For large spills of LPG where there is no immediate ignition there still remains a
chance that the flammable gas will encounter an ignition source and ignite before
dispersing down to the lower flammability limit. It was assumed that 70% of all large
spills would ignite giving an ignition probability factor of 0.70. Given that the assumed
132 CHAPTER 4
probability of immediate ignition is 0.2, then the probability of delayed ignition is 0.5.
In the case of small spills of LPG occurring on the railway we may expect the spill to
disperse rapidly, thus without causing delayed ignition.
Event Modeling for a Release of Motor Spirit

The assumptions for a release of motor spirit were as follows:
a) Instantaneous release of the entire tank contents
b) Continuous spill from a puncture with an assumed equivalent diameter of 100
mm (the typical size of hole made by most likely puncture tools, and
coincidentally, the size of the tanker outlet pipework).
~ maximum mass flow rate can be modeled as:
G = Cd A ~p(~ -P2) (kgS·l) (4.2)

2
and assuming that the pressure difference (~ - ~ ) is equal to P g h the head of motor
spirit above the leak,
G = Cd Ap 12g h= 25 kgs· l (for a 100 mm hole) (4.3)
where
Cd = coefficient of discharge (0.6)
p = liquid density (725 kgm·~
A = hole area (m2)
h = height of liquid above discharge point (maximum h = 2.6 m, from barrel
diameter of rail tanker).
The formation of the pool was calculated using the computer code SPREAD (and
SPll..L),l2 which uses the Shaw and Briscoe l3 equations for pool spreading. The
regression rate of the liquid due to combustion was taken from Mizner and Eyrel4. The
liquid was assumed to spread across rough sandy soil, and to have a minimum depth of
25 mm (Opschoor ls, 1978). It was estimated, that a continuous release of 25 kgs· l from
a 32 t tank wagon would give rise to a pool with an estimated radius of 24 m; and a 75 t
tank, a pool radius of 37 m.. The tanks would take about 20 minutes and 50 minutes
respectively to empty. In both cases, if immediate ignition occurred, the effective size of
the pool fire would be 12 m. The pool radii for instantaneous releases would be similar.
The consequences for the exposed population can be derived from the assumptions
that fatalities are limited to those people within the area of the pool. The effect of
thermal radiation from pool fires was considered, but it was concluded that people
exposed would almost certainly be able to escape or find shelter without sustaining fatal
injuries.
TABLE 4.11. Number of fatalities after a pool fire event, as a function of pool area and
population density
Pool Fire Event Continuous Release Instantaneous Release
32t 7St 32t 7St

Immediate, 0 0 0 906 Po
on-track
Delayed, 0 906 Po
on-track
Immediate, 415 Po 415 Po 1810 Po 4162 Po
off-track
Delayed, 1810 Po 4185 Po 1810Po 4347 Po
off-track
(AfterHSC,l991)
TABLE 4.12. Predicted national risks due to rail transport of dangerous goods (for GB)
Rail Transport Estimated National Frequency per Year (x 10·')
for N greater than
1 3 10 30 100 300 1000
Motor
Spirit
Off-rail 22,923 10,908 1,850 0 0 0 0
Passenger - - - - - - -
Total 22,923 IO,90t 1,850 0 0 0 0
LPG
Off-rail 1,020 687 522 235 82 24 0
Passenger 368 368 363 328 209 0 0
Total 1,299 1,059 923 653 365 24 0
Chlorine
Off-rail 362 197 174 87 82 20 13
Passenger 567 567 567 509 114 0 0
Total 600 591 587 576 243 66 14
Ammonia
Off-rail 1,596 1,090 910 896 304 33 0
Passenger 480 480 480 474 378 270 0
Total 1,892 1,398 1,288 1270 679 326 0
National (for four
Total substances )
Off-rail 24,902 12,882 3,456 1,218 468 76 13
Passenger 1,415 1,415 1,410 1,311 701 270 0
Total 26,714 13,596 4,648 2,499 1,287 416 14
(After HSC, 1991)
134 CHAPTER 4
The number of people (N) likely to be killed (assuming no people within 25 m of

the railway track) can be estimated from the product of the pool area (m2) and the
density of the population, Po (people/m2), wherePo is the population density per square
meter. The results for Great Britain are shown in TABLE 4.11. The national risks for
rail transportation of hazardous materials in Great Britain have been estimated to be as
shown in TABLE 4.12.
Individual Risks (GB)

In the British HSC study the individual risk has been calculated as follows:
- Risk of death to a particular individual living 50 meters from the route which
carries 55% of the total chlorine traffic by rail is 2.4 x 10-7 per year.
- This risk falls to 5 x 10-8 per year for an individual 1 km from the route.
The highest individual risk from a representative marshalling yard is 5 x 10-7 for
nearest resident, and 2.7 x 10-7 at 500 m.
These risks can be considered as negligible. They are specific to the particular route
or marshalling yard, and clearly other routes with different dangerous goods traffic
would give rise to different levels of individual risks. The above mentioned figures are
representative for the rail traffic conditions prevailing in Great Britain.
4.3. Information Concerning the Swiss Federal Rail Network
4.3.1. CONFIGURATION OF THE SWISS RAIL NETWORK
The Swiss rail network is owned principally by the Swiss Federal Railway Company
(SBB, Schweizerische Bundesbahnen), but there are also some private railway com-
panies owning and/or operating more particularly railroads with special narrow size
tracks or cable trains.
Infonnation on Swiss Railroads (Source CIA/USA)

4,418 km total; 3,073 km are government owned and 1,345 km are non-government owned;
the government network consists of 2,999 km 1.435-meter standard gauge, and 74 km
1.000-meter narrow gauge track; 1,432 km double track, 99% electrified; the non-
government network consists of 510 km 1.435-meter standord gauge, and 835 km 1.000-
meter gauge, 100% electrified.
The main rail network operated by SBB has the following characteristics (SBB I6,
1992):
• Total SBn rail network length: 2,973 km; of which
- 270 tunnels (211 km), and
- 4,114 bridges (58 km)
• Total installed rail length (i.e., length of all rail tracks of the network and
railroad stations): 7,391 kIn.
• Number of railroad stations, main and intermediate stations: 814 stations, of
which
16 are switching yards, and
- 28 are main switching freight yards centers (i.e., railroad stations in charge
of assembling freight trains)
• Number of road-crossing: 6,414, of which
- 3,261 are railway tunnels,
989 are railway bridges, and
- 2,164 are level crossings
The SBB owns a total of 2,156 locomotives, 4,496 railway carriages, and 20,236
freight cars. They rent 6,750 private freight cars. The maximum freight car load is 28
tons with two axles, and 60 tons with 4 axles. The dangerous goods traffic is principally
taking place on the whole rail network.
4.3.2. SWISS RAILROAD STATISTICS
The SBB statistics 16 mention for the year 1992 the following:
• Passenger traffic
- Number of passengers transported: 260 million
Number of passenger' kIn: 11,034 million
Number of train· kIn: 28 million
• Goods traffic
Tonnage transported (tons): 51 million
Number of tons' kIn: 8,162 million
- Number of train· kIn: 28 million
• Dangerous goods according to RlDIRSD
Tonnage transported (20% of the total goods traffic
of 50 million tons): 10 million
of which:
- transported in freight trains: 6 million
- transported in single freight wagon: 4 million
The total number of trains involved in daily movements is shown below:
• Passengers trains: 4200 trains
• Goods trains: 2500 trains, of which:
- 25 trains transporting only petroleum products.
It must be noted, that a long-distance goods train includes, on average, a total of 20
wagons, each weighing 1,600 brut tons (1,100 net tons).
136 CHAPfER4
4.3.3. TRAFFIC STRUCTURE
According to the same SBB statistics 16, the following rail traffic (in million net tons)
has taken place in 1990, totaling 10.1 million net tons:
- Inland traffic: 2.5
- Export traffic: 0.4
- Import traffic: 5.2
- Transit traffic: 2.0
The repartition of the goods was according to the following products groups:
- Petroleum products: 62 %
- Chemical raw materials: 25 %
- Gases: 4%
- Fertilizers: 3%
- Pharmaceuticals/dyes: 3%
- Others 3%
4.4. BUWAL's Methodology for Assessing the Risks of Hazardous Materials

Transportation by Rail in Switzerland
Based on rule 5 paragraph 2d of the Ordinance of April 1, 1991 concerning the "Protec-
tion Against Incidents" (StOrfallverordnung, StFV)17 the Swiss railroad company SBB
has to estimate the probability of incidents with severe injuries to the population or
severe damages to the environment.
Recently the Swiss Federal Office for Environment, Forestry and Landscape,
BUWALi has issued a "Handbook No. III, to the Regulations Concerning Incidents,
StFV - Guiding Principles for Traffic Ways" (December, 1992).17 This handbook is a
complement to the Ordinance of April 1, 1991:
We shall describe in this section the BUWAL procedure used for assessing the risk
of railroad transportation of dangerous goods in Switzerland, and discuss the main
topics to be covered by the "detailed risk assessment report" addressed to the competent
authorities (i.e., in this case BUWAL).
4.4.1. TECHNICAL ASPECTS OF THE DETAILED RISK ASSESSMENT
The general technical approach of the BUWAL methodology will be described below,
but some administrative and juridical details from the original report have however
been omitted here.
, BVWAL: Bundesamt fiir Umwelt, Wald und Landschaft, Bern.

RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPORTS BY RAn. 137
Environmental Aspects
The rail tracks considered must be subdivided into short segments (minimum 1 km in
length) with homogeneous properties, the length of which is determined by the
technical and constructional aspects, the surrounding environment, the traffic and the
safety measures related to a given segment. Among the constructional and technical
aspects of the rail network, consideration must be given to the location of a rail track,
type of track, and railway stations. Among the elements composing the track type we
find the track width and the number of tracks, railway platforms, bridges, viaducts,
galleries, and tunnels. Regarding the railway stations and switchillg yards, the length
between the entrance and exit of a switch, the surface covered and the approximate
movements of trains have to be mentioned. The safety aspects of the segment have to be
considered too.
The rail tankers, tank containers, and combined railIroad tankers ("Huckepack", i.e.,
transportation system involving the loading of road trucks/tankers on a wagon) have to
be assessed regarding number and capacity (in tons).
Along the rail track we draw a corridor subdivided into a close and a far range. The
close range has a width of 200 m on both sides of the track and the far range reaches
1,500 m on both sides. The data related to the surrounding environment along this
corridor must be compiled and comprise the following elements:
• groundwater resources and surface waters, as well as other object to be protected
• population density
Population Density: The population density, expressed as the number of inhabitants

per square kilometer (Inb.lkm2), is to be estimated in the close and far range of the
railway installation. In cases where no population density data are available, the
following ~tandard values corresponding to Swiss conditions (Year 1992) can be
assumed:
• large urban population density: > 5000 Inh.lkm2
• small urban population density: > 2000 - 5000 Inb.lkm2
• population density in villages: 100 - 2000 Inb.lkm2
• scarce population or no habitation: < 100 Inh.lkm2 •
Groundwater Resources: All the Swiss cantons have established g~graphic maps
showing the classification of the protected groundwater zones S (SI, Sll and SIll), as
well as zones of surface waters A, B, C, D. In the close-range, all protected groundwater
zones S must be mentioned. An estimation is to be made of the unsaturated zone
thickness (distance between soil surface and water table) and of the flow direction, and
quantity of ground waters. For the protected zones S one has to mention the planed
water poises and pumping capacity.
The permeability of the ground surface is to be roughly estimated, the value being
required for estimating the probability. The distance from the railway installation to the
water resources must be recorded as well as the usage (drinking water, water reserve,
fishing area).
138 CHAPTER 4
In the far-range, one is only to indicate protected groundwater zones S when they
might be indirectly contaminated through waters located at proximity of the railway
installations, or if the waters are acting as receiving waters of a drainage segment The
protected waters zones A may be indicated as well.
Surface Waters: In the close-range all surface waters (flowing waters and standing
waters) and their distance to the railway installations are to be recorded, if they are
located along the rail tracks, or if the waters are acting as receiving waters of the
drainage system of the railway installation (e.g., in tunnels).
In the far-range, the surface waters are to be recorded, if they can be polluted
through waters, which lie in the proximity of the railway installation, or are acting as
receiving waters of a main draining segment under consideration. For the flowing
waters the yearly average discharge quantity (m3/s) is to be estimated; whereas for the
standing waters it is the water volume (m3). These data can be taken from the
"Hydrological Year-Book of Switzerland". The intended use of these waters is to be
described: direct use as drinking water resources (i.e., direct drawing out), or indirect
use for drinking water production (i.e., natural or artificial feeding of a groundwater
table)
Special objects located in the close range, that need to be protected, are listed below.
These include for instance:
• Schools, hospitals, shopping centers, sport yards, camping, exhibition areas
• Protected objects or landscape zones of national relevance
• Inventoried fruit plantations
• Industrial plants with special danger potential
• Driuking water supplies
• Important traffic ways
Traffic Composition
The traffic composition regarding dangerous goods (net tons per year) is to be estimated
for each track segment. We distinguish between normal dangerous goods trains (miXed
with passenger cars) and special -dangerous goods trains (containing exclusively
dangerous goods), container traffic, and rail/road traffic combination ("Huckepack"
transport system).
Actually it is not possible to distribute the total quantity of dangerous goods
according to the RSD danger classes, so we have to use product categories instead.
Regarding the conventional goods traffic the Swiss average values can be taken from
TABLE 4.13.
The product categories 1231130 (Mineral Oil and Petroleum Products) comprise
only daf1gerous goods and represent in general the largest percentage quantity of
dangerous goods transported. When the quantities belonging to this product group are
known, then it is possible to distribute the rest of the dangerous goods transports
between the other product categories according to the ratio of the percentages for the
RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPORTS BY RAn. l39
Swiss' average. The same consideration holds true for container and "Huckepack" traffic
(product categories 201 and 202).
TABLE 4.13. Distribution of dangerous goods according to product categories
Product Category Type of Dangerous in % of the Total

Goods Quantity of
DangeroWJ Goods
123/130 Petroleum products 62
and tar; Products
derived from coal;
and natural gas
141 Chemical raw 25
products
122 Gases 4
60 Fertilizers 3
142 Phannaceutical 3
products; Dyes; etc.
- Remaining products 3
Where possible actual quantities should be used and in case of necessity the
estimates should be corrected to reflect particular conditions (e.g., harbor, strategic
petroleum products storage, main transit axis or regional railway lines etc.).
For every dangerous goods transported by special train the average quantities (net
tons per train) and the number of special tft!.:lSports per year must be indicated.
Accidents History
Based on available incident protocols, the incidents/accidents must be evaluated for the
last ten years' period, and classified by accident type (i.e., derailment, collision, etc.)
and by accident location (i.e., open-air track segments, tunnels, railway station, and
shunting yards). Accidents involving dangerous goods, that resulted in leakage, fire, or
release of toxic gases or water-polluting liquids, must be mentioned with all details.
Basically, the rail accident rates must be calculated on the basis of the number of
derailments and collisions per traveled train·km, for open tracks, tunnels, and railway
stations. When data are missing, the following accident rates can be assumed to be
prevailing for Swiss conditions (year 1993): 0.12 x 1O-6/trains·kID for open tracks and
tunnels, and 0.6 x 10-6 Itrains·kID for railway stations. For trains using special narrow
tracks higher accident rates have to be assumed: i.e., 0.3 x 10-6 Itrains·kID for open
tracks and tunnels, and 1.5 x 1O-6/trains·kID for railway stations.
140 CHAPTER 4
4.4.2. ESTIMATION OF THE LIKELIHOOD OF AN INCIDENT WITH SEVERE

CONSEQUENCES FOR THE POPULATION OR THE ENVIRONMENT
Description of the Methodfor Assessing the Risksfrom Rail Transportation

The purpose of this estimation is to identify the rail track segments with comparatively
higher probabilitiesi of risk, possibly resulting in severe consequences for the popu-
lation or the environment, for further detailed investigation. The method described
below allows to assess in a coarse way the probability of severe damages to the
population, surface waters, and groundwater resources, for each rail track section, on
the basis of incident/accident scenarios. The methodology proposed by BUWAL was
developed after taking into considerations inland and foreign experience and
knowledge. IS, 19,20
The competent Swiss Authorities have fixed the following limits of acceptance
concerning the "incidents with severe consequences":
• Population: 10 or more people killed. ii
• Groundwater: any excess of the tolerance value, as defined in the Swiss Food.
Ordinance, either for mineral oil products, or for liquids that are potentially very
dangerous to water in the surrounding of a drinking water well located in a
protected zone S.
• Surface waters (rivers): the contaminationiii of a river over a length exceeding
10 km or more.
• Surface waters (lakes and ponds): the contamination of a lake iii over a surface of
1 km or more iv.
Representative Incident Scenarios

The representative incident scenarios which have to be assumed for different kinds of
dangerous goods causing similar effects are as follows:
• Severe damages to the population: Fire, explosion and release of toxic gases.
• Groundwater resources: Release of mineral oil products and liquids that are
potentially very toxic to groundwater and surface water resources.
For each representative scenario we have to define and select a reference product
corresponding to a product category of substances presenting similar toxic effects.
, Here the term "probability" is not defined, as usual, in the probabilistic sense. but has to be understood as
meaning "frequency".
U This defmition includes as well the population in the surrounding of railroad installations. as the passengers
in the train and in the rail way stations.

.. Here, the contamination (pollution) is defined as a toxic load of mineral oil products. respectively liquids
that are potentially very dangerous to water. in excess of 15 g1m 2 of surface area, respectively 5 glm2 for a
river (including the bank zone).
'" One admits. consequently. that rivers less than 10 km in length and lakes with a surface area less than I
km2 will not be contaminated seriously.
RISK ASSESSMENT OF HAZARDOUS MA1ERIALS lRANSPORTS BY RAIL 141
Frequency of Incidents with Severe Consequences

The frequency of representative incidents scenarios with severe consequences has been
defmed as follow:
• For the population: the sum of the frequencies of the incident scenarios for fire,
explosion and release of toxic gases, which lead to severe damages.
• For groundwater resources, respectively for surface waters: the respective sums
of the frequencies for the representative scenarios involving the release of
mineral oil products, respectively liquids which are potentially very dangerous to
water, and who can lead to severe damages.
The frequency of the representative incident scenarios involving severe damages per
kilometer and year is to be calculated for each rail section as indicated below:
F. = YDG·AR·PD·RR/·RSS (4.4)
where
Fs : Frequency of a representative incident scenario with severe damages
[(km year)"!]
YDG : Yearly average number of Goods trains transporting Dangerous
goods of concern, according to the RID list [(train km)-!]
AR : Accident Rate [(train km)"!]
PD : Probability for a railcar transporting a given dangerous good to
derail, or to be Damaged [-]
RRI : Release Rate; and in case of fue and explosion the Ignition rate [-]
RSS : Ratio of representative incident Scenario with Severe damages [-]
The frequency of representative incident scenarios imolving a mix of goods trains
and of "whole goods trains" (tran~Qrting exclusively ci3ngerous goods) is calculated
by taking the sum of the frequencies estimated for each categories separately. For
"whole goods trains", the YDG value corresponds to the number of "whole goods
trains" and PD is equal to 1.
The explanation given below concerning rDG and PD are only pertaining to goods
trains.
Average Yearly Number of Goods Trains Transporting Representative Dangerous

Goods (YDG)
For each representative scenario we have to estimate the number of wagons transporting
representative dangerous goods by dividing the yearly transported quantity of such
dangerous goods by the average wagon load. In the case that the resulting number of
wagons is smaller than the number of goods trains, we assume YDG to be equal to the
number of wagons transporting representative dangerous goods. Else, depending on the
specific track segment, we have to distribute the number of wagons transporting repre-
sentative dangerous goods on the yearly number of goods trains.
142 CHAPfER4
Probability of Derailment, or Damage,for Wagons Loaded with Representative

Dangerous Goods (PD)
For Railcars Transporting Dangerous Goods: the frequency of derailment, or of being
damaged, is estimated by assllming that four railcars on average are involved per rail
accident.
PD=4(noolno) (4.5)
where
PD : Probability of derailment, or damage
nDG : Average number of wagons transporting relevant dangerous goods
in a goods train
no : Average number of wagons in a goods train
In the cases where nDG is large as compared to no, and the wagons transporting the
dangerous goods build a block (group), then the factor 4 is to be correspondingly
reduced.
Rate for a Relevant Release and Ignition (RRJ)

We assume for a given representative incident scenario that all the relevant substances
are to be transported in approximately the same quantities and type of containers, thus
we shall assume the release rate, or in case of fIre and explosion, the ignition rate to be
uniform on open track segments and in railway stations. For tunnels we shall assume a
fIve times higher release rate, in order to cope with special particularities.
The values for relevant release and ignition rates were collected from foreign
information sources and from Swiss data as well.
Ratio of Representative Incident Scenarios Involving Severe Damages (RSS)

The factor RSS is standing for the probability of a severe damage under the assumptions
that a relevant release has already taken place, or that in case of fife and explosion an
ignition has occurred.
RSS Value Concerning the Population. This is dependent upon the density of the
passengers and on the position of the wagon after the accident (i.e., on or outside the
rails)
RSS Value Concerning Groundwater Resources. This was obtained after considering
drainage behavior of the reference substance, soil permeability, the distance from the
floor and from the protected zone to the railway system.
RSS value concerning surface water. This was obtained after considering the drainage
behavior and the possible position of the wagon when the incident occurred (i.e., on
track, or beside the track; respectively on which side of the track). Most important are
however the distance from the surface waters to the railway installation, and the
discharge quantity of the riverllake. The possibility of an infIltration into a protected
groundwater zone S was also taken under consideration.

SCENARIOS INVOLVING THE POPULATION
Representative Scenario "FIRE"
Reference Substance. GASOLINE was chosen as representative flammable product for

the representative scenario "FIRE". In order to estimate the value of YDG, one has to
take 65% of the dangerous goods of the product group "Mineral oil and petroleum
products", and 15% of the dangerous goods of the product group "Chemical ground
products". It will be assumed that the freight transported per rail tanker is:
• for "Petroleum products": 55 t, and
• for "Chemical raw products": 33 t.
RRI: A rate of 0.05 was assumed for a relevant release of gasoline (quantity> ca. 1.5
t), and a probability value of 0.33 for an ignition following such a release. As a result
RRI is taking a value of 0.1. For tunnels the value of RRI(tJUlMI) was assumed to be 0.5.
RSS: the corresponding RSS-factors are listed in TABLE 4.14.

TABLE 4.14. Determination of the RSS-factors for the scenario "Fire"
Open Rail Section and Railway Stations Tunnel

Population density in the close range
(People I km 2)
>5'000 ! 2'000 - 5'000 <2'000 One track two tracks

--
RSS =0.40 ! RSS =0.20 RSS=0.05 RSS=0.05 RSS ..,0.35
Representative Scenario "Explosion"
Reference Substance. PROPANE was chosen as representative flammable/explosive

product. In order to estimate the value of YDG, one has to take 65% of the dangerous
goods of the product group "GASES", and 3% of the dangerous goods of the product
group "Chemical raw products". It will be assumed that the freight transported per rail
tanker is equal to an average of 40 t for both product groups.
RRI: a rate of 0.05 was assumed for a relevant release of propane (quantity> ca. 2 t),
and a probability value of 0.8 for an ignition following such a release. As a result RRI is
taking a value of 0.04. For tunnels the value of RRI(UUUleI) was assumed to be 0.2.
RSS: the corresponding RSS-factors are listed in TABLE 4.15.
144 CHAPTER 4
TABLE 4.15. Determination of the RSS-factors for the scenario ''Explosion''
Open Rail Segments and Railway Stations Tunnels

(people 1 lcm2)
>5'000 2'000 - 5'000 <2'000 One track Two tracks

RSS=O.50 RSS=0.30 RSS=0.15 RSS = 0.05 RSS=O.40
Representative Scenario "Release of Toxic Gases"
Reference Substance. CHLORINE was chosen as representative substance for the

scenario involving a leak and release of toxic gases. In order to estimate the value of
YDG, one has to take 5% of the dangerous goods of the product group "Chemical raw
products". It will be assumed that the average freight transported per rail tanker is equal
to 33 t.
RRI: A rate of 0.03 was assumed for a relevant release of chlorine (quantity> ca. 2 t).
As a result RRI has a value of 0.03. For tunnels the value of RRI(1IuInd1 was assumed to be
0.15.
RSS: The corresponding RSS-factors are listed in TABLE 4.16.
TABLE 4.16. Determination of the RSS-Factor for the Scenario ''Release of Toxic Gases"
Open Rail Segments and Railway Stations Tunnel

(People/km2)
>5'000 2'000 - 5'000 <2'000 One track Two tracks

RSS=0.60 RSS=O.40 RSS=0.30 RSS=O.05 RSS=O.50

SCENARIO "GROUNDWATER RESOURCES"
The probability of an incident resulting in severe damages to groundwater resources has

only to be assessed for protected groundwater zones, in which important wells for
municipal drinking water supply are present or have been planed. Severe damage to
groundwater is only possible when the well is located close enough to the railway
system, or when the groundwater flow direction is taking place from the railway system
to the groundwater zone.
Representative Scenario "Release of Mineral Oil Products
Reference Substance. FUEL OIL (for domestic heating) was chosen as representative
product. In order to estimate the value of YDG, one has to take 100% of the dangerous
goods of the product group "Mineral Oil Products". It will be assumed that the freight
transported per rail tanker is equal to an average of 55 t.
TABLE 4.17. Detennination of the RSS factor for the scenario "Release of
Petroleum Products"
Soil Permeability i
Thickness of small medium high

Soil Cover k< 10.5 m/s 10.5 < k < 10-3 k> 10-3
(fine sand, silt, (Clayey gravel (gravel)
clay stone, Sand )
<2m RSS=O.l RSS=0.3 RSS =0.8

2m-10m RSS =0.01 RSS=O.l RSS=O.4
>lOm RSS =0.01 RSS = 0.01 RSS=0.05
RRI for flowing waters: A rate of 0.25 was assumed for a relevant release of fuel oil
(quantity> ca. 19 t). As a result RRI has a value of 0.25.
RSS: The protected groundwater zones S have, in Switzerland, been chosen in such a
way that incident with petroleum products taking place outside the protected zone will
not lead to severe damages to the groundwater wells. As a consequence the probability
of an incident leading to severe damages to the groundwater resources will have to be
assessed only for protected groun.::1water zones S when the rail track is crossing the zont"
or is located very close.
The corresponding RSS-factors are listed in TABLE 4.17. In cases where particularly
efficient measures for groundwater protection have been taken, the RSS-factor is to be
adjusted in consequence.
Representative Scenario "Release of Strongly Water-Polluting Liquids "
Reference Substance. "TETRACHLORETHEN" was chosen as representative product.

In order to estimate the value of YDG, one has to take 8% of the dangerous goods of the
product group "Chemical Raw Materials". It will be assumed that the freight
transported per rail tanker is equal to an average of 33 t.
RRI: A rate of 0.3 was assumed for a relevant release of tetrachlorethane (quantity> ca.
2 t). As a result RRI has a value of 0.3.
RSS: The value of the RSS-factor is shown in TABLE 4.18.
, The soil permeability is given by the coefficient k. This coefficient is taking the values shown in TABLE
4.17.
146 CHAPfER4
TABLE 4.18. Detennination of the RSS factor for the scenario "Release of Strongly
Waler- Polluting Liquids"
Soil Permeability
Thickness of small medium high
Soil Cover k< 10.5 mls lO's < k < 10.3 k> 10.3
(fine sand, silt, (Clayey gravel (gravel)
clay stone, Sand )
Distance between protected zones and railway installation < 50 m
<2m RSS =0.2 RSS= 0.5 RSS= 1.0

2m-10m RSS=0.05 RSS=0.2 RSS=0.8
>lOm RSS=O.Ol RSS =0.05 RSS =0.5
Distance between protected zones and railway installation 50 m to 200 m
<2m RSS=O.Ol RSS=0.05 RSS=O.l

2m-10m RSS=O.Ol RSS=O.Ol RSS =0.05
>lOm RSS=O.Ol RSS=O.Ol RSS=O.Ol

SCENARIO "SURFACE WATERS"
Representative Scenario "Release Mineral Oil Products"
Reference Substance. "FUEL OIL" (for domestic heating) was chosen as represent-
ative mineral oil product. In order to estimate the value of YDG, one has to take 100%
of the dangerous goods of the product group "Mineral oil and Petroleum Products". It
will be assumed that the freight transported per rail tanker is equal to an average of 55 t.
TABLE 4.19. Determination of the RSS-Factors for the scenario "Release of Petroleum
Products Into Lakes"
Distance from the Road

<50m
I 50m-2oom
RSS=O.05 I RSS = 0.01

With confirmed infiltration
RSS=O.3 I RSS=O.l
RRI (jor flowing waters): A rate of 0.24 was assumed for a relevant release of fuel oil
(quantity> ca. 10 t). As a result RRI(Flowing Waters) has a value of 0.25.
RRI (for lakes): A rate of 0.15 was assumed for a relevant release (quantity> ca. 50 t). As
a result RRI(l-aus) has a value of 0.15.
RSS: The value of the RSS-factor for lakes is shown in TABLE 4.19, and for flowing
waters in TABLE 4.20.
TABLE 4.20. Detennination of the RSS-factors for the scenario "Release of Petro-
leum Products into Flowing Waters"
Discharge Rate Distance from Railway Imtallation

(m3/s )
<50m 50m-200m
10 - 75 RSS=0.6 RSS = 0.2

75 - 125 RSS=0.3 RSS=O.1
> 125 RSS=0.2 RSS = 0.05
With confirmed infiltration
10 - 75 RSS=0.6 RSS = 0.3

75 - 125 RSS= 0.4 RSS=0.2
> 125 RSS = 0.4 RSS = 0.2
Assumptionsfor the Representative Incident Scenario "Release of Liquids Strongly

Noxious to Groundwater Resources"
Reference Substance: "TETRACHLORETHEN" was chosen as representative product.

In order to estimate the value of YGD, one has to take 8% of the dangerous goods of the
product group "Chemical Raw Products". It will be assumed that the freight transported
per rail tanker is equal to an average of 33 t.
RRI (jor flowing waters): A rate of 0.3 was assumed for a relevant release (quantity> ca.
2 t). As a result RRI(F/OWing Waters) has a value of 0.3.
RRI (jor lakes): A rate of 0.15 was assumed for a relevant release (quantity> ca. 5 t). As
a result RRI(Lake.) has a value of 0.15.
RSS: The value of the RSS-factor is shown in the TABLE 4.20 and TABLE 4.19 for
"Petroleum Products".
4.5. Probability of Incidents on the Swiss Rail Network
Based on rule 5 paragraph 2d of the Ordinance of April 1, 1991 concerning the

"Protection Against Incidents" (StOrfallverordnung, StFVfl the SBB have to estimate
the probability of incidents with severe injuries to the population or severe damages to
the environment. Some selected topics and results from the short risk assessment report
148 CHAPTER 4
of SBB I6 to BUWAL are briefly summarized below. The data are valid for the Swiss
rail network in general, but detailed risk assessment studies have also been prepared by
SBB along the same line, for each single rail track segments.
4.5.1. SWISS RAILROAD ACCIDENT STATISTICS
For the period 1979 to 1989 the SBB have registered around 460 incidents involving
dangerous goros movements. This figure includes aU incident registered by SBB, that is
also the leak of a fuel oil tank, as well as incident with SBB owned road tankers. In
70% of all the cases, the product involved was a petroleum product, and in 30% it was
another chemical product. 13% of the incidents resulted in a fire involving a petroleum
product. Most of these incidents did not endanger people or the environment, and the
resulting damages remained reasonable. In 3% of all the incidents, there were people
injured slightly, and in less than 2% a temporary pollution of the waters had to be taken
into consideration.
It must be noted that during the period 1990-1995 some severe rail accidents,
involving dangerous goods transport as well as passengers have taken place in
Switzerland (and also in other countries). This means that the probabilistic figures
prepared by SBB at the request of BUWAL have changed. Such probabilistic data
should any how be reconsidered every year as a function of the number of traffic
accidents/incidents.
4.5.2. ESTIMATION OF INCIDENTS FREQUENCY ON THE SBB RAIL

NETWORK
Incidents with personal injuries (dead, evacuated persons) through the. transport of
dangerous goods are rare on the Swiss rail network. Thus their freqt.·ency cannot
reliably be determined only on the basis of the Swiss railroad statistics, because these
are in such a case insufficient due to the small sample size. Therefore, in the case of
SBB, the estimation of the incident frequency with personal injuries is prepared
principally on the basis of international statistics related to accidents involving
dangerous goods transportation by railroad.
Regarding incidents with environmental damages, international statistics do not
offer an appropriate support since corresponding data are hardly available. On the
contrary several incidents have happened in Switzerland in the past, where large
quantities of dangerous/toxic liquids have run out. These form a certain statistical basis
for the estimation of incident frequency with environmental damages. Indeed for most
incidents which led to a pollution of surface waters, groundwater, or soil, or had the
potential for a pollution, some quantitative statements about the extent of' the resulting
consequences are totally missing. Therefore the estimation of the extent of environ-
mental damages took place for the most part with the help of analytical model repre-
sentation.
The incident frequencies presented in the following pages represent values summed
over the entire SBB rail network, respectively the shares of the different track types.
They are valid for the present technical and operational factors of the SBB network
(reference year 1990). With regard to the calculation of incident frequencies for
local/regional rail tracks, the data must be put in relation to some characteristic values
like transport capability, traffic rise, population density, as well as the local situation
concerning surface waters and groundwater.
4.5.3. DEFINITION OF INDICATORS FOR ASSESSING THE EFFECTS FROM

INCIDENTS
We show below (Figure 4.4), as example, the scales used by SBB for connecting the
different kinds of damages (Le., number of fatalities, people evacuated, contaminated
surface areas) with the event severity scale (defining acceptable and severe damages).
Figure 4.4 shows the definition adopted for the damage indicators used for assessing
the effects of an incident and the definition of classes measuring the extent of effects. A
numerical representation of the annual incident frequencies for the entire rail network is
given for all indicators and effects severity classes. These values are briefly interpreted
and commented below and in the next section.
The SBB have calculated that, on the Swiss rail network, one may expect a
dangerous goods accident once in every 40 years;, that will cause fatalities. Accidents
with severe damages, i.e., such one occasioning more than 10 fatalities, shall take place
once in 150 years. Incidents with more than 100 fatalities are essentially more seldom to
expect. For all three damage severity classes about 75% of the incidents are concerned
with fatalities occurring inside, and 25% outside the SBB railroad installations.
Once in every 8 years one has to reckon with an evacuation of people in the vicinity,
i.e., outside of the railroad installations. Once in every 40 years we may expect an
accident that will cause the evacuation of more than 1,000 people. Catastrophic
-\l~cidents that will cause the evacuation of more than 10,000 people should occur less
ClaD once in 300 years.
It is difficult to describe the concept of contamination exactly. Here, it was assumed
that we have to do with contamination, if at least short-term damages occur to fauna or
flora, or the utilization of the affected object by people is no more possible, at least
temporarily, or its use is very limited. The duration of the environmental damage,
however, is not considered explicitly for the assessment.
Contamination of groundwater catchment areas following an incidental leakage of a
dangerous good can be expected to take place once every 100 years. It was assumed that
in approximately 70% of all cases it was concerning a catchment with a pumping
capacity of less than 500 Umin, and in ca. 30% such one between 500 and 5,000 Umin.
Incidents causing severe damages to groundwater catchment delivering more than 5,000
Umin are expected to take place only once every thousand years. The latter value does
include the case where several catchment are contaminated at the same time.
IThis value corresponds to the average recurrence period. and is mathematically equal to the reciprocal
value of the corresponding frequency (e.g .• 1111 = 11 0.025 = 40 [years».
150 CHAPTER 4
Damage Severity Scale (0 • 1.0)

o 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Damage
Fatalities
1 10 100 Number of Fatalities
I I
Evacuations
1'000 10'000
I
Number of Persons Evacuated
I
I I
Surface Waters
0.1 1
I I
Groundwater Catchment
500 5'000
I I I
Figure 4.4. Example of a scale system used for assigning single damage classes indica-
tors to an event severity scale.
Regarding surface waters, large areas can be contaminated as a result of the

particular dispersion mechanism. One has to expect that every 8 years a contamination
involving 1,000 to 100,000 m2, and once every 80 years a contamination of 100,000 m2
to 1 km2 may take place. Catastrophic incidents with severe damages, involving more
than 1 km2 are seldom and should not happen more than once every thousand years.
Soil contamination between 100 and 1,000 m2 are expected to take place every two
years and such ones involving more than 1,000 m2 once every 5 to 6 years. The square
meters of surface area, indicated here, consider only a superficial contamination of the
soil. Here, only the surface contamination relevant to the vegetation was considered.
The possibly larger surface area of the soil polluted at a greater depth is covered
through the "groundwater" indicator. For this reason, cases of contaminated surface area
exceeding 10,000 tn2 can only occur under very special conditions, that is in the
presence of an inclined ground impermeable to liquid wetting. The corresponding
RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPOKI'S BY RAIL 151
frequencies are an order of magnitude smaller and consequently not relevant. The
damage potential of accidents involving dangerous goods is, for the soil, comparatively
small. As opposed to the other indicators, the frequency of incidents that result in severe
soil damage are negligibly small, and thus can be omitted.
TABLE 4.21. Frequency of incidents classified by damage type and effects severity class for selected
indicators
Damage Type Indicator Units Effects Severity Frequency

Class per Year
1-9 0.018
Direct injuries to Number of fatalities fatalities 10-99 0.0060
individual outside SBB facilities
~loo 0.0004
Number of fatalities fatalities 1-9 0.0135
inside SBB facilities 10-99 0.0045
1-999 0.099
Indirect injuries Number of people 1,000-9,999 0.026
to individual people evacuated eVllCWlled
~l,ooo 0.003
Polluted catchment 1-499 0.0066
Pollution of Groundwaler (with licensed ground- IImin 500-4,999 0.0027
waler delivery rate) ~5,ooo 0.0002
0.001-0.1 0.13
Pollution of Surface Surface area. of
km2 0.1-1.0 0.013
Waters contaminated witters
~1.0 0.0002
ooceevery
100-1,000 2 years
Damages to Soil Surface area of polluted m2
(ecosystem, fauna, flora) soil ooceevery
~l,ooo
5-6 years
Source: SBB short report WIthin the framework of the Incident Ordinance (1992)
4.5.4. DISTRIBUTION OF INCIDENT FREQUENCIES ACCORDING TO RAIL

TRACK TYPE
Accidents involving 1 to 9 fatalities are to be expected principally in railway stations.

Accidents with more than 10 fatalities are taking place primarily in tunnels. Due to
partial obstruction of the tunnel by the railroad wagons toxic vapors and combustion
gases are prevented to escape to open air. If at the time of the incident/accident a
passenger train is also present in the tunnel, then the passengers of this train are
possibly in danger. Moreover, the intervention of rescue troops in tunnels is strongly
complicated, because the long approach ways, the tunnel entrance that are not easily
152 CHAPTER 4
accessible, and the narrow passages between railroad material and tunnel walls
(TABLE 4.22).
TABLE 4.22. Distribution of fatality incidents by track segment, inside SBB railroad installations
(F: Frequency per year)
Type of Rail Road Number of Fatalities

Installation Inside SBB Railroad Installations
1· 9 10·99 ~ 100
F = 0.0141 year F = 0.00451 year F = 0.00031 year
Railway station 45% 23% 10%
Shunting yard 5% 2% -
Open track 20% 15% 10%
Tunnels 30% 60% 80%

Source: SBB short report within the framework of the Incident Ordinance (1992)
TABLE 4.23. Distribution of fatality incidents by track segment, outside SBB railroad installations
(F: Frequency per year)
Type of Rail Road Number of Fatalities

Installation Outside SBB Railroad Installations
1·9 10·99 ~100
F = 0.0041 year F = 0.00151 year F = 0.00011 year

Railway station '40% 50% 10%

Source: SBB short report wIthin the framework of the incIdent OrdInance (1992)
Fatalities occurring outside SBB facilities (TABLE 4.23) are due to dangerous
goods accidents taking place mainly at railway stations. The reason is primarily due the
higher population density in the vicinity of a railway station, as compared to the
population density along the open railroad track. In the vicinity of shunting yards, the
probability of accidents involving many fatalities outside SBB railroad installations is
comparatively seldom. Indeed the velocity of the wagons during shunting is small com-
pared to usual railroad traffic speeds. Therefore the impact energy during collision or
derailment is generally to small in order to create severe accidents. Dangerous goods
accidents in tunnels are occasioning fatalities outside SBB facilities only in special
situations. Such accident type represent only a small share.
Most accidents necessitating evacuation of people can be expected to take place at
railway stations or on open railroad tracks. Tunnels and shunting yards contribute each
by about 10%, Other SBB statistics are given in TABLES 4.24 through TABLES 4.26.
RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPORTS BY RAn.. 153
TABLE 4.24. Distribution of incidents by track segment. 00 SBB railroad insta1latioos. leading to
an evacuatioo of people (F: Frequency pel' year)
Type of Rail Road Number of Evacuations

Installation on SBB Rail Network
1-999 1,000 - 9,999 ~ 10,000
F = 0.099/ year F = 0.026/ year F = 0.003/ year

Tunnels 9% 4% 30%
· . the framework of the Incident Ordinance (1992)
Source: SBB short report within
TABLE 4.25. Distribution of incidents by track segment, 00 SBB railroad insta1latioos.leading to a

contaminatioo of grOlmdwater carchment area (F: Frequency pel' year)
Type of Rail Road Contamination of Groundwater Catchment Area

Installation along SBB RaIlroad Installations
1- 499l1min SOO - 4,999l1min ~ 5,000 llmin
F = 0.0066/ year F = 0.00l7/ year F = 0.0001/ year
Shunting yard 14% 14% 14%
1---
f---
Tunnels 1% 1% 1%
· . the framework of the InCident Ordinance (1992)
Source: SBB short report Within
TABLE 4.26. Distributioo of incidents by track segment, on SBB railroad installations. leading to
a contamination of surface waters (F: Frequency per year)
Type of Rail Road Contamination of Surface Waters

Installation along SBB Railroad Installations
0.001 - 8.1 lun2 0.1-1.0km2 ~ 1.0tm2
F = 0.0066/ year F = 0.00l7 / year F = 0.0001/ year
Shunting yard 15% 6% 3%
Tunnels 2% 2% 2%
· . the framework of tbe InCident Ordinance (1992)
Source: SBB short report Within
154 CHAPTER 4
As already mentioned, the statistics and frequency data are no more up to date siBce
the publication of the SBB report. The risk assessment process, like a rolling-ff1reCast,
should be reviewed every year and revised accident frequency data prepared m order to
keep up to date with new developments concerning equipment modernization and safety
aspects on the rail network and in the tunnels. Railroad traffie frequency is also
changing over the years, and new rail tracks or switching yards are being implemented.
During the last years, there has been problems due to intensive rain-storm occasioning
earth movements that dragged away the tracks, or earth avalanche covering up the
tracks with mud, gravel and rocks. The risk of such event, which may have significant
consequences, like derailment, should be considered in future railroad risk assessment
studies. Recent earthquake in the USA and Japan have also damaged highways, rail
tracks and bridges of modem construction type. In Switzerland, such incident might
lead to land slide as well.
Summary
This chapter introduces the methodological framework for performing a risk analysis of
hazardous materials transportation by rail in the content of regional risk assessment and
safety management Experience with train accidents in a number of countries is
presented (e.g., USA, UK, Switzerland). Details are given on the Swiss methodology
for assessing the risks of hazardous materials transportation by rail: technical aspects of
the detailed risk assessment, estimation of the probability of an incident with severe
consequences for the population or the environment, various assumptions concerning
the representative incident scenarios involving the population (e.g. scenario for fire,
explosion, release of toxic gases, etc.), or the representative incident scenario for
groundwater resources, surface watt:rs, etc. The information and risk estimates pre-
sented in the last part of this chapter were compiled by SBB in accordance with the
requirements of the existing Swiss legislation on risk assessment for rail transportation
of dangerous goods.
1 FRA (Federal Railroad Administration), Accidentsllncidents Bulletin No. 160: Calendar Year 1991,
FedernI Railroad Administration, Office of Safety, Washington, D.C., (July 1992).
2 Denis, A.W.; Fowler, G.F., Railroad Tank Car Safety Assessment, Report RA-12-4-58, Association a
American Railroads, Chicago, (1990).
3 Clarke, R.K., et al., Severities of Transportation Accidents. Report SLA-74-0001, Sandia National Lab0-
ratories, (1976).
4 Fischer, L.E., et ai., Shipping Container Response to Severe Highway and Railway Accident Conditions,
Report NUREG/CR-4829, Lawrence Livermore National Labontory, (1987).
5 Harwood, D.W.; Russel, E.R., Present Practices of Highway Transportation of Hazardous Materials, Report
FHWA-RD-89~13, U.S. Department of Transportation, (1989).
6 Brobst, W.A., Transportation accidents: how probable 1, Nuclear News 16(5), (1973), pp. 48-54.
'USAEC (U.S. Atomic Energy Commission). Environmental Survey of Transportation of Radioactive
Materials to and from Nuclear Power Plants, Report W ASH-1238, (December 1972).
8 Health & Safety Commission (HSC): Major hazard aspects of the transport of dangerous substances, Report
and appendices, HMSO, London, (1991).
9 Purdy, G.; Campbell, H.S.; Grint, G.C.; Smith, L.M., Analysis of the risk arising from the transport a
liquefied gases in great Britain, J. Hazard. Mater. 20 (1988), pp. 335-355.
10 Lautkasi, R.; Mankamo, T., Chlorine transportation risk assessment, Technical Research Center of Finland,
Nuclear Engineering Laboratory, Report 27. Available from Valtion Painatuskeskus, PI.. 516, 00101
Helsinki 10, Finland, (1977).
11 Davies, P.A.; Lees, F.P., Accident speed of freight trains, J. Hazard. Mater. 28 (1991), pp. 367-370.
12 Prince, A. I., A users manual to SPIlL, UKAEA Reliability Directorate, Culcheth, SRD Report R2IO,
(1981).
13 Shaw, P.; Briscoe, F., Evaporation from spill of hazardous liquids on land and water, UKAEA Reliability
Directorate,Culcheth. SRD ReportRl00, (1978).
14 Mimer, G.A.; Eyre, I.A., Large-scale LNG and LPG pool fues, in The Assessment of Major Hazards,
Rugby, Institution of Chemical Engineers, (1982), pp. 147-164.
15 Opschoor, G., Theoretical study of the evaporation of liquids on land, Chapter 5 Ev~on, TNO Report
No. 78~8345, Nederlandse Organisatie voor Toegepast-Natuurwetenschappelijk ~derzoek Apeldoom,
The Netherlan<!s, (1978).
16 SBB: Schweizerische Bundesbahnen, Kurzbericht dec SBB im Rahmen der SWrfallverordnung, Teil I:
Netzweiteaspekte, SBB-Generaldirektion, Bern, (12. MiIrz, 1992).
17 BUWAL, Handbucb m zur SWrfallveroIdnung, Ricbtlinien fUr Verkehrswege, Dezember 1992, BWKlesamt
fUr Umwelt, Wald und Landschaft, Bern.
18 Health & Safety Commission (HSC): Major hazard aspects of the transport of dangerous substances, Report
and appendices. HMSO, London, (1991).
19 Inter-Agency Program on the Assessment and Management of Health and Environmental Risks from
Energy and Other Complex Industrial Systems: Manual for the classification and prioritization of risks
from major accidents in process and related industries, International Atomic Energy Agency, Vienna,
(1991).
20 US Federal Emergency Management Agency, US Dep. of Transportation, US Environmental Protection
Agency: Handbook of Chemical Hazard Analysis Procedures, Washington DC, (1989).
21 Verordnung tiber den Schutz vor SWrflillen (SWrfallverordnung, StFV, vom 27. Februar 1991).
CHAPTER 5
TRANSPORTATION OF DANGEROUS GOODS BY SHIP
5.1. Risk Resulting from Ship and Barge Transport on Lakes and Waterways
Ship and barge transport on lakes and waterways present some risk to the environment,
people and to fixed installations and buildings. The greatest risk arises from explosion
of a flamrTUlble gas {propanelbutane, gasoline, solvents) cloud escaping from tanker-
ship damaged by a collision. Non-combustible gases, such as chlorine or ammonia, can
escape from a container as the result of mechanical or corrosion damage and form a
toxic gas cloud drifting in direction of populated areas along the lake or river. Another
risk may be due to water intrusion into a damaged barge and subsequent leaching of
some water soluble toxic substance, which may pollute lakes and rivers, thus presenting
a danger to ground-water and aquatic life.
Water Pollution Risk

The presence of water on a barge may be the result of fire fighting or of a leak or breach
of containment integrity. Water entering the barge may also chemically react with some
product in bulk, or in textile/paper bags and produce noxious fumes which will drift
away and may present a danger to the surrounding population. Humidity or water may
also react with a product in bulk, which as a result of swelling may mechanically
damage the barge (as it has already been the case with barges transporting cereals on the
Rhine).
Contamination by chemicals in liquid form may be due to a container leaking. The
spill on water may take fire and eventually cause an explosion. Finally a barge may
after a collision capsize, thus loosing bulk material or containers. Other incidents due to
handling may take place during loading and unloading of LPG, gasoline, solvents or
other chemicals and lead to fire and explosion. Finally pollution of the waterway may
be the result of tank cleaning operations conducted in a improper manner.
Risk of Explosion Due to the Transportation of Flammable Products

Among the potential hazards due to maritime or fluvial transport, the explosion of a
flammable cloud formed by accidental release of liquefied gas on the site of the
accident or after drifting is of particular concern. Such an explosion can produce an
over pressure wave in the air, which can cause damage to property or people or affect
the resistance of any kind of building or process plant in the surrounding. In general, the
risk of explosion due to transportation of dangerous materials can be estimated on the
basis of probabilistic considerations. Overfilling of tanks combined with elevated
ambient temperature can produce an increased venting due to the partial pressure of
solvent present in the tank, thus resulting in an increased flammability risk. Welding
158 CHAPI'ER5
operation on tanks or smoking near-by may result in fire and explosion. A cause of
explosion is quite often due to the presence of static electricity. The electrical charges
are created during the pumping of flammable liquids, as a result of a missing earthing or
the use of unsuitable piping equipment in conjunction with a to high a pumping
velocity.
5.2. Techniques for Risk Assessment of Ships
5.2.1. OBJECTIVES
The objectives of an overall risk assessment of the major hazard aspects of the bulk
transport of dangerous substances by sea, in the ports and around the coasts, or on
inland waterways (Le., rivers and lakes) should be as follows:
1. To analyze the risks of major hazard accident arising in a given country as a
result of the transport of dangerous substances in bulk by ship or barge. The
risks to be assessed should include both individual and societal risks of death
due to fire, explosion, or toxic release. For shipments on inland waterways, the
analysis should include statements on the possible pollution of water surfaces by
such shipment, in case of fire or shipwreck.
2. To identify the main components of these risks.
3. To develop a methodology for such risk assessment adapted to the local
conditions, and suitable for evaluating remedial measures.
4. To evaluate selected remedial measures using cost-benefit analysis.
The main categories of dangerous materials which are transported in bulk include:
Crude oil
- Flammable and toxic liquefied gases
- Flammable liquid petroleum products
- Flammmable liquid chemicals
- Hazardous solids (e.g., ammonium nitrate)
- Solid and liquid chemicals potentially hazardous to water and aquatic life.
Concerning the transport of hazardous materials by ship or barge on inland
waterways, substances known to be potentially toxic to water and aquatic life should be
included in the study, since they can pollute a river, like the Rhine from Basle to
Rotterdam, as the toxic load moves down the river.
Many shipments are in tank containers, Ro-Ro vehicles, drums, and containers
enclosing smaller packages. Chemicals which are transported in small shipment on
parcel chemical carriers should be included in the study, provided that they are pumped
and not lifted ashore. Bagged ammonium nitrate, should be included because its
explosive effects are similar to those of the same material in bulk. The analysis of
TRANSPORTATION OF DANGEROUS GOODS BY SHIP 159
explosions of ammonium nitrite shows, however, that these are extremely unlikely
events.
5.2.2. METHODOLOGY FOR ASSESSING THE RISK OF INDIVIDUAL PORTS
A paper on risk assessment of shipping operations has been recently published by the
Health and Safety Commission (HSC) in Great Britain! and the methodology followed
is presented below. Three of the 88 terminals in 42 ports in Great Britain were analyzed
in detail and a simplified method was developed for the others. 2 "Major hazard
accidents" are accidents which threaten people on-shore or in other vessels in the port,
or crew members while the ship is berthed in port. Only spills or explosions of LPG are
likely to reach the population at risk. LPG is usually imported by both large fully-
refrigerated and small semi-refrigerated gas carriers. The small vessels have cylindrical
pressure tanks, but carry the cargo fully refrigerated. Their cargo capacity is typically
3,700 m 3 • The large gas carriers have prismatic tanks and a total capacity of 23,000 m3 •
The cargo is carried fully refrigerated at atmospheric pressure. An articulated arm with
quick-release and emergency shut-down is used for ship-to-shore transfer. Petroleum
products are imported in small tankers using flexible hoses for transfer.
The risk analysis of individual ports follows the classical fonn of quantified risk
analysis, consisting of the following steps:
• System definition, in which information on the ports and waterways and its
hazardous trades is collected.
• . Hazard identification, in which events are identified which may lead to the
release of hazardous cargo.
• Frequency estimatico, in which the frequency (i.e., likelihood per year) of each
event is estimated, u!tually from historical accident data.
• Consequence modeling, in which the consequences (i.e., fatalities) of each event
are estimated, usually by a mixture of accident experience and theoretical
models.
• Risk representation, in which the frequencies and consequences of each event are
combined to give the overall individual and societal risks.
• Assessment of the tolerability of the risks using agreed individual and societal
risk criteria.
The modeling of the consequences can be perfonned, for example, using the
computer program SAFETI (Ale and Whitehouse3, 1986), which allows efficient calcul-
ation of the effects of population distribution, meteorological conditions, ignition
source location, and release type, duration and location.
The risk presentation should follow the following fonn:
Individual Risk. The likelihood of death per year for an individual continuously at a
given location, assuming an average vulnerability to the hazards and allowing for
realistic probabilities of being sheltered indoors.
160 CHAPTER 5
F-N Curves. They are a measure of societal risk showing the relationship between
numbers of fatalities and frequency (i.e., likelihood per year) of accidents. The FN
curves are actually cumulative plots, showing the frequency F of accidents involving N
or more fatalities.
Annual Fatality rates. They are simpler measures of societal risk, equal to the expected
long term average number of fatalities per year (i.e., expectation values). These give no
indication of the numbers of fatalities in individual accidents.
Mean Fatalities per Event. They represent the average number of fatalities in given
accident types. These supply the missing information in the fatality rates. They are
equal to the fatality rate divided by the event frequency.
Information on Hazardous Bulk

Approaches to port authorities often reveal a surprising lack of detailed knowledge
about specific materials passing through their ports. Information on the movements of
bulk hazardous cargoes is sometimes difficult to obtain, because data are only available
on bulk shipments, without hazardous materials being identified separately.
In Switzerland, the situation in the ports of both Basle is complicated, due to the
goods being unloaded from ship onto trucks, road and rail tankers, and the statistics
prepared by the different parties, are sometimes difficult to reconcile.
Terminals which handle many different materials are sometimes unable or unwilling
for commercial reasons to provide information on individual materials, so they are
condensed into broad product groups. Quite often does the grouping correspond to the
requirements of the custom and is not performed according to danger classes.
Estimation of Accident Rates

Hazard identification involves identifying all relevant types of accidents for the
analysis, based on statistics from historical accident records and from previous risk
analyses. It is useful to chose a hazard classification which could fit all known
scenarios. Many types of accidents can be excluded as they have no potential to cause
fatalities among the public.
One of the approach for estimating the accident rates consists in using the mean
accident rate for most parts of the route on the waterway, and using the actual number
of accidents for those segments that have a significantly higher number of accidents. A
better approach is based on a technique developed by Hauer4 (1986). In this approach
Hauer uses a Poisson distribution to assess the accident rate of a large number of
seemingly equal locations (i.e., route segments with equal length). The test he proposes
gives a readily applicable estimate of the accident rate T for a location with n reported
number of accidents, Tn. It is assumed that a reported number of accidents may
occasionally be high, while the underlying accident rate is actually low, or the other
way round. The estimation of accident rate as function of the location within the
waterway is especially important because of the high nonlinear behavior of the
probability of death as function of space (for relatively small radius). It must be
remembered that the location of an accident is not at the center of the waterway but may
be spread anywhere over the entire width. According to this description effect models
should incorporate a correction factor or a different probability function.
The analytical process of estimating waterway vessel casualty probabilitiess
involves the following approach:
• Define waterway and subzones
• Analyze subzone historical vessel casualties
• Estimate subzone vessel traffic by type and size
• Quantify subzone waterway characteristics
• Calculate national average casualty rates by casualty type, vessel type, and size
(in the case of rivers, e.g., the Rhine, include traffic and accident data for the
whole river length, i.e., from Holland to the lake of Constance (Bodensee)
• Make regression analyses of subzones casualty rates with waterway characteris-
tics
• Develop subzone navigational risk adjustment values
• Finally calculate subzone specific vessel casualty probabilities by casualty type,
vessel type, and size.
The waterway specific characteristics may be described as waterway physical mea-
surements as well as meteorological and hydrologic conditions. Among the physical
measurements, we have channel width, depth, route alignment in terms of number of
turns, average or sum of total degrees of course changes, number of obstructions (e.g.,
bridges, anchorage, crossing lanes, etc.). total waterway length and total area of water
,surface in the subzones. These attribute should be measured for each river segment
Among the meteorological and hydrologic factors to be included are wind speed,
visibility, current velocity, water volume in m3/s. Having these information on hand it is
then possible to develop a linear multiple regression model for estimating the
probability of vessel casualties. The most significant variables to be selected for the
stepwise linear regression are: the waterway/channel characteristics, a local traffic
density factor, and two subzone type dummy variables (like constricted waterway,
harbor area, etc.).
It has been proven difficult to find a relationship between the meteorological or
hydrological conditions and the casualty probabilities, although visibility, wind speed,
and current velocity play a role regarding the navigational risk. In fact actual
meteorological conditions are more important than annual average values.
Risk Assessment Method

The type of releases to be considered are: hot releases of liquefied gas, flash fires and
vapor cloud explosions; cold releases of liquefied gas or flammable liquid, and
explosion of flammable liquid. Explosions result either from ignition source, fire or
directly from collisions between ferries or ship and tankers. Pool fires result from spills
162 CHAPTER 5
of flammable liquid around the jetty, tanker and ferry bow. Fire may spread from the
tanker to other vessel or ferry boats or harbor installations.
• Estimate the frequencies for the occurrence of the different scenarios, as explained
above.
• Assume different local meteorological conditions (a combination of wind speed
and atmospheric stability).
• Estimate the population density, using two distributions: one for day-time and one
for night-time. The data may be compiled from census data and from information on
the local industries in the region of interest.
• Establish the ignition sources distributions in order to model the probabilities of
ignition from single people, motor vehicles, industries, trains, ships, loading/filling
station, etc. based on judgment.
• Using special models like, for instance, SAFETI 3 (Software for the Assessment of
Flammable, Explosive and Toxic Impacts), calculate the spreading of the pool on
waterlland and the evaporation rate as well as the dispersion of the gas cloud with
respect to time. 6
• Calculate the possibility of ignition during the dispersion phase of the cloud, using
the lower flammability limit (LFL), and assuming a flash fire upon ignition in 90%
of the cases and an explosion in 10% of the time (vapor cloud explosion is very
unlikely). Different possible time steps for the occurrence of ignition must be
considered. Assume 100% of death within the reach of the flash fire for people
outdoors, and 30% for people indoors.
• Calculate the fatalities for each failure case and each of the atmospheric parameter
combinations, wind direction, population distribution (at day- and night-time),
ignition sources, time steps, and the proposed two ignited outcomes (flash fire and
vapor cloud explosion). The fatalities are expressed both as probabilities at each
point on a 100 m grid and as the total expected number of deaths.
• Compute the measures of individual and societal risk, using the probabilities of
each combination and the failure case frequencies. Usually the catastrophic
consequences are restricted to the vicinity of the harbor installations, but in some
instances catastrophic gas releases may travel over 1 krn before reaching their lower
flammable limit (with the wind prevailing in this direction), and so their frequency
is relatively low.
Based on the experience from different harbors, the largest contributions to the
societal risks arise from tanker transfer spills, which are ignited by passing motor crafts
or cars and from collisions between tankers and ferries or other ship. Sensitivity test can
be used to assess the effects of the proposed remedial actions. Finally using cost-benefit
analysis, it is possible to compare the costs of implementing remedial measures for
reducing accidents. The benefits of remedial measures are expressed in terms of the
risk-factored costs of accidents avoided.
Ship Accident Scenarios

The types of ships which are used to carry the bulk hazardous cargoes must be defined.
Among sea-going ship we have:
- Tankers - i.e., vessels carrying crude oil, liquid petroleum products or liquid
chemicals in bulk. These cargoes are mostly carried at around atmospheric
pressure, with the tanks vented to the atmosphere.
Gas carriers - Le., vessels carrying liquefied gases in bulk. They are of the
following types: pressurized ships, refrigerated and semi-refrigerated ships, and
ammonium nitrate ships.
Among the ship accidents, we have to consider collisions between ferries and tankers,
tankers and other vessels, and tankers with passenger ship. The types of accident which
pose a fatality risk are mentioned below:
• Collision - where two vessels under way run into each other.
• Striking of a vessel berthed at the jetty by another passing vessel.
• Impact - where a vessel runs into a dock wall or jetty.
• Grounding - where a vessel runs onto the shore or submerged rocks (wrack!
stranding)
• Fire or explosion on the tanker
• Cargo transfer failure
In addition to these main causes, there are other minor causes such as:
foundering/capsize, structural failure of the cargo tank, spontaneous failure of the cargo
tank, refrigeration failure, domino accident.
A major hazard accident usually begins with an accidental release of hazardous
cargo. Two types of releases are distinguished: cold releases due to mechanical impact,
and hot releases due to the tank failing in a fire; but also including explosions within the
tank.
The main groups of people at risk are:
- Residents and workers along the Rhine harbor: only spills' burning or explosions
of LPG, gasoline or other flammable product are likely to reach the population at
risk.
- Ship's crew and jetty staff: are vulnerable while the ship is berthed at the jetty.
- People on other vessels: it includes people on ferries, and on passenger ships
moving on the Rhine.
5.2.3. FREQUENCY ESTIMATION FOR SHIP TRANSPORTS
The frequencies which are required for a study are likelihood per year of an accidental
release or explosion of hazardous cargo on. Tanker explosion frequencies can be
estimated directly from historical data. The frequencies of cargo release should be
estimated for collision, striking, impacts and gas tanker fires as the product of two
164 CHAPTER 5
components: accident frequency and conditional probability of a release of cargo given

the accident scenario.
Calculation of Striking Frequencies

The striking frequency for a ship can be estimated according to:
SF - YF (5.1)
- PM·PY
where
SF : Striking frequency (per passing movement)
PM : Passing movement (per year)
PY : Period of time (number of years)
Gas carrier Vessel struck Cargo tal Angle of Penetration Result of ProbabiHty
tanks
'struck at in way of loaded collision of tank accident
jetty cargo (obUque)
0.15 (Yes) Release 0.019
0.38 (Yes)
-
0.85 (No) No spill
0.5 (Yes)
0.65 (Yes) 0.62 (No) No spill
0.5 (No) No spill 0.981
0.35 (No) No spill
-
Figure 5.1. Event tree for gas carrier striking.
Figure 5.1 shows, as an example, an event tree for a gas carrier striking, and the
associated event probabilities.
Expressing the frequency in this way allows it to be compared with striking
frequencies in other ports or fluvial segments with different traffic levels. The striking
frequency for a ship oflength equal to the jetty head can be estimated as:
_ SL TB·YF
~ - SF· JL' 24.365 (5.2)
where
F. : striking frequency for a particular ship (per year)
SF : striking frequency per passing movement
SL : ship length (m)
JL : jetty length (m)
TB : time at berth per visit (br)

VF : visit frequency (per year)
The above striking frequency is based on the vessel being struck while at the jetty,
whether loaded or not. Using a similar technique, frequencies can be estimated for all
causes of release of flammable liquids from vessels at the jetty. The corresponding
event tree can be described according to Figure 5.1.
Estimating Explosion Risk

Fire and explosion frequencies depend more on the ship carrying hazardous cargo than
on the port it is in. Thus world-wide accident data for the appropriate cargo and ship
type can be used. Tanker explosion (including those with total lost) can be derived from
a special analysis of Lloyd's Casualty Returns for the period 1977-1986 (TABLE 5.1).
The number of cargoes loaded and discharged were estimated from the number of
ships in service and an assumed number of cargoes delivered per ship year. These were
allocated into the material groups in proportion to the number of total losses other than
cargo explosion in the group. Dividing the total losses in port due to cargo explosions
by the estimated number of cargoes loaded and discharged gives the explosion
frequencies for each cargo type.
TABLE 5.1. Accident frequencies by cargo group for the period 1977-1986
Cargo group Transfer Accident Explosion

(per cargo (per visit)
transferred)
Crude oil 1.9 x 10-4 1.3 x ur'
Low flash products 1.8 x 10-4 3.5 X 10-6
High flash products 1.8 x 10-4 1.5 X 10-6
Flammable liquefied gases 7.6 x lO-s -
Toxic liquefied gases 7_6 x lO-s 0
Low flash chemicals 1.5 x 10-4 1.3 x lO-s
High flash chemicals 1.5 x 10-4 0
Ammonium nitrate 0 7.0x 10-8
Source: HSC I ,1991.
5.2.4. PROBABILISTIC APPROACH
EDF (Electricite de France) has developed a method for estimating the risk resulting
from ship transport of dangerous cargo in the vicinity of industrial plants located along
a river or lake, based on probabilistic and deterministic approach. 7 The methodology is
described below.
166 CHAYfER5
The most general expression for the annual probability P j (year l ) of a fluvial or sea
accident, which can cause damage to people, property, or more specifically, to a
chemical or nuclear plant, is written, for a given product, i, and accident scenario s, as:
(5.3)
where
Pa : rate of occurrence of any accident on a lake or on a waterway per ship/barge
carrying dangerous materials and per kilometer of river bank or shore
p. : conditional probability of occurrence of an accident on a lake or on a fluvial
waterway, which could cause an explosion
p. : conditional probability of explosion
Pk : a weighting coefficient, associated with the geographical location studied and
the nature of the dangerous substance
F, : yearly traffic in the neighborhood of sites expressed in terms of the number
of ships carrying the dangerous substance, i, over a year
P, : probability of occurrence of the scenario, s
P!fW : probability of occurrence of given meteorological conditions on the site, as
defined by atmospheric stability and wind speed (sw)
Ej : factor associated with meteorological conditions (direction and frequency of
winds, conditions of atmospheric diffusion), which exists only in the case of
gaseous cloud dispersion
L, : length (km) of the river bank, or shore, or ship trajectory on a lake to be
taken into consideration:
(5.4)
Do : nummum distance (km) between a popUlated area (urban region), or

industrial area (chemical or nuclear plant) and the ship/barge.
Dj : safe distance (km); beyond it the effects of the explosion no longer affect
safety. This distance depends on the standard accident scenario, s, taken into
account, and on the kind of effects considered (effects on people, housing,
industrial buildings). The type of building must be considered too, since the
resistance of the structures to over pressure chock waves depend on the
construction type and niaterial used (bricks, steel, reinforced concrete, etc.).
The safe distance also takes account of possible spreading of the gaseous
cloud.
Gas Carrier Fire Scenario

The frequency of cargo release from a gas· carrier in a fire can be estimated as the
product of : frequency of frrelexplosion on the ship and probability of cargo release
given that a fire/explosion incident has occurred.
It has been estimated that nearly 100 such incidents may have occured with no cargo
release so far. The probability can be obtained by assuming that the number of releases
follows a Poisson distribution. Then with 50% confidence, the true mean number of
releases will not be greater than 0.7 (i.e., the probability of release is < 7 x lO-3). As fire
scenario one can assume a hot rupture of a tank of flammable liquefied gas. Hot rupture
is assumed to occur in 10% of hot releases. The other 90% are assumed to be leaks
through the safety valve which are consumed by the fire with no major hazard.
Accident Rate
The accident rate per dangerous ship can be estimated from statistical data. As an
example we indicate typical data (Hendrickx et al. 7 , 1983) for French coastal waters
and sea passage (estimated for the year 1983).
Coastal waters: P a = 7.6 X 10-8 accident. (dangerous shipr!. (k:m of coa~t\ine)"!
In sea passage: P a = 2.9 X 10-6 accident. (dangerous ship)"!. (Ian of channel bank)"!.
These averages are weighted by the geographic location of the accident: PI allows
for this adjustment and was varying for the French study from 0.2 to 2.8.
The accident frequencies, in Great Britain, have been reported by the National Port
Council (NPC 8, 1976) and are reproduced in TABLE 5.2
TABLE 5.2. Ship accident frequencies in Great-Britain for the year 1976.
River type Collision Grounding Striking Impact

(per encounter) (per passing) (per visit)
Open sea port 5.0 x 10-4 6.5 x 10-5 4.0 X 10-6 2.2 X 10-3
Wide estuary 4.0 x 10-5 8.0 X 10-6 4.0 X 10-6 2.2 X 10-3
Wide river 1.2 x 10-4 5.0 x 10-5 9.0 X 10-6 2.1 X 10-3
Narrow river 5.0 x 10-4 1.6 x 10-5 4.2 x 10-5 6.5 X 10-3
Similar accident frequencies for barge and ship traffic on the Rhine in the Basle area
must be calculated.
Probability o/Threshold Event

Probability of threshold event Pth is the productP"p". From the various incident on
fluvial waterways and harbors or lake, some only are a potential cause of an explosion.
The values P" and p. must be estimated or calculated from statistical data. The initial
causes can be defined as:
• Direct grounding
• Collision between ship
• Collision with other obstacles, such as bridges or pontoons, walls of a lock, etc.
168 CHAPrER5
• Propulsion failure. combined with drift and grounding or collision

• Direct explosion
TABLE 5.3 shows the probability of an initial event causing an explosion on a
tanker ship transporting a dangerous cargo.
TABLE 5.3. Probability of an initial event causing an explosion on a tanker ship ttansporting a
dangerous cargo (aft« Hendrickx, 1983)
Initial Event Probability Probability Probability Traffic Type

Causing an h, Pe, hxPe,
Explosion for French on a world of threshold
waters basis event
(1970-1979) (1965-1974)
Direct 0.20 0.008 0.002 Coastal Traffic

Grounding
Collision 0.24 0.10 0.024 Coastal Sea

Passage
Propulsion 0.04 0.08 0.0004 Open Sea,

failure + Drift Seaways
+ Grounding
Direct 0.08 1 0.08 Coastal Sea
Explosion Passage
Definition of a Standa ...d Tanker Accident

The use of the parameter L/ in above formula requires the calculation of a s~fe distance
D,. and thus. a realistic determination of the volume of gas released and of the spreading
of the cloud. It should be noted that for the transport of LPG tankers with a double hull
are required. which makes them less brittle than oil tankers in case of a collision.
A representative basic accident for a tanker can be defmed on the basis of three
terms:
• the size of the breach/leak and its location: Breakage of the hull breach
corresponding to the depth of penetration of the colliding body (other ship. rock.
steel superstructures (in a harbor. of a pontoon. lock. etc.).
• the maximum volume (100% of a tank) which can be released (due to a breakage
of the hullship ripped open).
• the possibility that the cloud formed will drift away prior to ignition.
In order to calculate the eventuality of a gas release. the following assumptions must
be made:
• The gas carrier must be struck in way of a cargo tank, since not all the length of
the ship is occupied by cargo tanks. Assume a probability of 0.65 (depending on
vessel type and tank location and spacing).
• The tank of the gas carrier must be loaded: assume a probability equal to 0.5.
• The collision must be at an oblique angle, with a probability estimated to 0.38,
and,
• The striking ship or barge must have sufficient energy to brea(;h the cargo tank.
This energy is proportional to the striking ship displacement and the square of its
speed, and is also affected by the strength of the other tanker structures, jetty, etc. In a
British study, the probability for a ship to have enough energy as been estimated to be
on average 0.31 for large gas carriers and 0.85 for small gas carriers. These values
should be reconsidered for the conditions prevailing on the Rhine in the region of Basle
(i.e., vessel type and size. speed, etc.).
The major hazards from bulk substances are shown in TABLE 5.4.
TABLE 5.4. Hazards from bulk transport
Material Outcome
Toxic liquefied gas Gas cloud
Rammable liquefied gas Vapor cloud explosion (VeE)
Rash fire
Fireball
Rammable liquid Explosion (of vapor)
Pool fIR'
Bulk solid Explosion
Occurrence of a Breach with Discharge of a Fluid

The following assumptions must be made for calculation:
• assume liquid drainage (pessimistic hypothesis),
• assume isothermal flow in the case of cryogenic liquids, or
• assume adiabatic conditions inside the tank, and thermodynamic equilibrium of
the liquid and vapor phase, in the case of pressurized hydrocarbons
The corresponding formula (which can be found elsewhere in the open literature)
must be used for assessing the consequences from a tank rupture (i.e., fire, explosion,
pollution, etc.).
Formation and Vaporization of a Liquid Layer on Water

Calculate the total rate of vaporization and the radius of the liquid spill as a function of
time, taking into consideration that:
170 CHAPfER5
• the rate of flow from the breach and the volume discharge depend on the
position of the breach relative to the water line.
• the substance discharged forms a liquid layer on the water surface, and is being
vaporized due to heat supply from the environment.
• first a circular expansion, and then regression, of the layer, whose thickness
varies in space and over time up to a limit beyond which the effect of gravity is
taken as negligible.
• the vaporization caused by the supply of heat from ambient air, solar radiation,
and from the water (mainly for sea or lake water).
• the vaporization is further influenced by: convection phenomena, phase changes
(freezing of the water due to the evaporation process), conduction (the thickness
of ice varies with space and time).
Drift and Dispersion of the Gaseous Cloud

Because the source has a flow rate and a spill radius varying with time, use:
• a Gaussian puff dispersion model, to calculate the contour of the flammable part
of the cloud and the volume of hydrocarbons within the flammability limits.
(Actual codes do include gravity slumping effects).
TABLE 5.5. Probability of occurrence of a gas cloud explosion (for dangerous

goods transported by tanker ship)
Scenario: Explosion of a Gas Cloud

Gas/Liquid Max. Volume Drift Probability of
Release Released Occurrence, p.
Low '" 5% ofa tank No 0.12
Yes 0.121
High '" 65% of a tank No 0.56

Yes 0.07
Serious 100% of a tank No 0.22

Yes 0.03
Note: The probabilities of occurrence mentioned here apply to French waters only,
and are given as.example.
Explosion of the Gaseous Cloud

In order to estimate the consequence of an unconfined gas cloud explosion we use the
concept of TNT equivalent. At the 95% confidence level the total explosion yield is
10%, which corresponds to a theoretical TNT equivalent of 1 kg of TNT per kg of
released hydrocarbons. This total explosion yield can be interpreted as the product of a
vaporization yield (i.e., flammable mass I released mass) and an explosion yield.
TRANSPORTATION OF DANGEROUS GOODS BY SIllP 171
Assuming a vaporization yield of 20%, the empirical TNT equivalent of flammable

hydrocarbons is then equal to 5 kg of TNT per 1 kg of flammable hydrocarbons. In
order to estimate the mechanical damages as a function of distance, multiply this last
value with the flammable mass within flammability limits to obtain the practical TNT
equivalent. Knowing the mechanical resistance of buildings and installations, it is
possible to determine the safe distance, Vi' using the TNT equivalent concept (fABLE
5.5).
5.2.5. FAILURE CASES AND RELEASE SCENARIOS
It is difficult to assess rigorously events which involve extremely complex physical

phenomena, such as two-phase flow, chemical kinetics, turbulence, atmospheric
dispersion, etc. As a result pessimistic hypothesis must be used to cope with the
uncertainties involved. A striking could produce a cold rupture of a tank followed by
instantaneous release of the content, or a range of hole sizes and locations, which may
result in a leak through a hole, whose size is assumed to be that of a loading pipe
connection, located at the bottom of a prismatic tank or at the mid-height of a
cylindrical or spherical tank.
TABLE 5.6. Release scenarios for tanker-ship
Case 1: Cold Release of Liquefied Gas
t.I Refrigerated Catastrophic Contents release onto water over a period of 5 min.
Tanks cold rupture Predict spill vapor dispersion.
cold leak Flow is at first driven by hydrostatic head, followed by

heating of the cargo by the sea water. A constant flow
rate equal to 75% of the initial hydrostatically-driven
release is assumed. TIle entire contents of the tank are
released at this rate. Predict spill vapor dispersion.
1.2 Pressurized Catastrophic Assume an isentropic hemispherical expansion of the gas.

Tanks cold rupture Initial dilution of the gas must be determined. Pool
formation due to rain-out must be estimated. Predict spill
vapor dispersion.
Cold leak from Assume continuous release of the liquefied gas is driven
pressurized tank by the internal pressure. Calculate th~ steady-state
or loading arm flashing release rate. The momentum of the jet is assumed
to be dissipated by the ship secondary containment
(double hulled walls for gas tankers). Predict spill vapor
dispersion.
172 CHAPTER 5
TABLE 5.6. (Continued)
Case 2: Hot Release of Liquefied Flammable Gas
2.1 Refrigerated Hot rupture of a) To be modeled eitber as a fireball, on tbe basis that
Tanks tank containing pressure could build up inside the tank before it failed in
refrigerated a fire, or as a pool fire. Determine the thermal radiation
liquefied gas from the mass involved.
b) Pool fife: Calculate the spreading of the pool on the

water and the evaporation rate. Then determine the
thermal radiation from the pool.
2.2 Pressurized Hot rupture of Assume a fireball and determine the thermal radiation
Tanks tank containing from the mass involved.
pressurized
flammable
liquefied gas
TABLE 5.6. (Continued)
Case 3: Cold Release of Flammable Liquid
3.1 Atmospheric Damaged Release will be driven by hydrostatic head. A constant

Tanks tanker flow rate equal to 75% of the initial hydrostatically-
driven release is assumed. The entire contents of the tank
are released at this rate.
3.2 Full-bore Full-bore trans- The liquid flow will emerge at the pumping rate.
Leak fer spill during
loading or un-
loading opera-
tion
3.3 Pipe Leak Leak during For leak during transfer, assume an outflow at 10% of the
loading or un- pumping rate.
loading opera-
tion
For case 3.3: Consider spreading of the liquid onto water

under gravity, and assume 5 mm liquid thickness. Pool is
assumed to drift for up to 20 min. Assess the possibility
of ignition at each time step (during release and pool
spreading). Use lower flammability limits. In 90% of the
occasion an ignited cloud will form a flash fire.
It is assumed that 90% of cold releases are due to leaks and 10% to ruptures.
Experience has shown that using two different hole sizes for modeling would be
preferable, since it is now believed that a cold rupture of a gas carrier tank is very
unlikely. if a fire, however, causes rupture of a tank containing pressurized flammable
liquefied gas, this must be modeled as a fireball. The corresponding formula, which can
be found in literature, must be used. A model like SAFETI9 (which is basically an
improved version of the standard box model of Cox and Carpenter lO (1980) may be
1RANSPORTATION OF DANGEROUS GOODS BY SHIP 173
used to estimate the effects and consequences of (dense) gas release, flash fire, fireball,
or vapor cloud explosion. Possible scenarios assumptions are listed in TABLE 5.6
below.
Cold release of flammable liquid from a damaged tanker will be driven by
hydrostatic head. A constant flow rate equal to 75% of the initial rate may be assumed
in this case. This flow rate will continue until hydrostatic equilibrium is reached.
During transfer operation, the flow is assumed to emerge from a ruptured pipe at full
bore (loading arm) at the pumping rate. For a leak during transfer, assume a rate of 10%
of the pumping rate.
The release of liquid is considered to spread under gravity until it reaches an average
thickness of 5 mm. Then the spill will drift away under the influence of wind and water
currents.
5.2.6. POLLUTION OF WATERWAYS BY DANGEROUS CARGO
Solid Materials Stored in Containers. Such containers are usually boxes, textile or
craft-paper bags, barrels, or steel drums. The storage method usually consists of piling
or storing these containers. Solid bulk materials are stored in piles, or in bins, silos or
bunkers located on the ship. Such solid materials present less problems.
TABLE 5.7. Scenario for water pollution by soluble (toxic) solid substances
FirelExplosion TextileJPaper Use of Foam!· Product in Consequences on

on SbiplBarge Bags Damaged Water Contact with Aquatic Life
Water
Yes No Yes/No No No effect

Yes Yes Ves No No effect
Yes Yes Yes Yes See cases below.
Water Penetra- Bulk material Yes Product is non- No effect on aquatic

tion of Barge. wet or flooded toxic and insolu- life
with water. ble in water.
Water Penetra- Textile or paper Yes Product is non- No effect on aquatic

tion of Barge. bags wet or toxic and insolu- life.
flooded with ble in water.
water. Product is toxic No effect on aquatic
but insoluble in life.
water.
Product is non- Water pollution with
toxic but soluble no effect on aquatic
in water. life.
Product is toxic
and soluble in Water pollution with
water. effects on aquatic life.
174 CHAPTER 5
Bulk Storage of Liquid Materials. This type of storage requires the use of reservoirs and
tanks. Liquid materials stored in containers are generally packed in barrels, kegs, drums,
cans, or glass or ceramic containers. These are beeing replaced, nowadays by plastic
containers. Cans are usually stacked on palettes.
TABLE 5.8. Scenario for water poUution by soluble (toxic) liquid substances
FlrelExplosion Contalnerffan Use of Foam!- Containers or Consequences on

on SbiplBarae kDamqed Water Tank Leaking Aquatic Life
Yes No Yes/No No Noeffed.

Yes Yes Yes No Noeffed.
Yes Yes Yes Yes Water Pollution Even-
tual effects on aqWllic
life.
A probabilistic approach similar to the one recommended by the French, for

estimating tanker explosion, could be used to assess the transport risk by barge/ship
with regards to the pollution of water ways as a result of accidental release (or leaching)
of toxic substances. In this case Eq. 5.3 should be modified accordingly, as shown
below:
(5.5)
in which
p. : rate of occurrence of any accident on a lake or on a river per ship carrying
dangerous materials and per kilometer of river bank
P_ : conditional probability of occurrence of an accident on a lake or on a river,
which could cause a pollution of the waterway by a water soluble (toxic)
substance
P,oI : conditional probability of the material container being damaged, thus
releasing water-soluble substances (liquid or solid)
p. : a weighting coefficient associated with the geographical location studied and
the nature of the dangerous substance (liquid, solid solubility, toxicity class)
F, : yearly traffic on the waterway of concern, expressed in terms of the number
of ships carrying the dangerous substance, i, over a year
P, : probability of occurrence of the scenario s.
The scenarios resulting in water pollution of waterways as a result of transporting
dangerous goods by barge/ship are summarized in TABLE 5.7 and TABLE 5.8.
Accidental pollution of waterways during loading and unloading at a tanker terminal
or as a result of ship cleaning operations in the harbor have not been taken into
consideration in the scenarios mentioned below.
5.3. Risk Management for Waterways and Ship Cargo
Setting limits for acceptable risk, assessing existing risk levels and predicting the cost
and benefits of proposed safety measures are all meant to assist the risk management
process. If local factors have led to a relatively high accident frequency, improvements
to the infrastructure (waterways, bridges, locks) or the local traffic situation is the
logical way to look for a solution. But local improvement may not suffice: a reduction
in accident frequency of say 20% as a result of structural improvements or traffic
control is quite an achievement, but is sometimes not enough to reduce risks that
increase on a logarithmic scale. There are not too many alternative waterways which
would allow to choose another route and shifting to train or truck transport may not
improve safety. One has to look for a more generic solution at national or even
international level. It must be kept in mind that structural improvement to vessels will
normally only be feasible for newly built ships. From the current developments it can be
anticipated that in future all hazardous cargo will be transported in double hulled ships
(at least for ocean going ship). Furthennore the improvement of human and
organizational factors will playa significant role. It is important to define and establish
unifonn transport risk limits for all transport modes, i.e., not only for ship transporta-
tion. Once these limits are set it is possible to guarantee some generally accepted safety
level.
5.4. Swiss Waterways and the Ports of Bo'" Basle
5.4.1. mSTORICAL BACKGROUND AND ORGANIZATION
Navigation on the Rhine river all die 'year round has only become a practical possibility
in this century and necessitated regulation of the fairway on the upper Rhine. The first
towed barge arrived in Basle on June, 1904. Cargo-handling was makeshift but work
was soon started on the construction of proper dock facilities.
Constructional phases of the ports in the canton of Basle-Town:
SUohan (left bank) 1906-1911.
- Kleinhiiningen (right bank): Klybeck- and Rhine quays, Basins I 1919-1926;
Basin II: 1936-1942;
Constructional phases of the ports in the canton Basle-Country:
- Birsfelden and Au (left bank): 1937-1940 (extended after 1950).
The following information describes briefly the situation:
- Inland waterways
65 kID; Rhine (Basle to Rheinfelden, Schaftbausen to Bodensee);
12 navigable lakes
- Ports
Basle (river ports of both Basle)
176 CHAPTER 5
- Merchant marine
23 ships (1,000 GRT or over) totaling 337,455 GRT/592,213 DWT; bulk
10, cargo 4, chemical tanker 5, oil tanker 1, roll-onlroll-off cargo 2,
specialized tanker 1.
Ownership of the Port Areas

The cantons of Basle-Town and Basle-Country are the respective owners of the port
areas. They lease the land with the requisite infrastructure (quays, roads, railroad tracks,
water and power supplies, lighting, etc.) to private firms which erect and operate cargo-
handling and storage facilities at their own expense.
Organization of the Ports of Both Basle

Pursuant to an agreement from 1946 between the two cantons the ports form an
operating unit under the name "Ports of both Basle" and come under the administration
of the Rhine Shipping Office at Basle, together with the port administration of Basle-
Country. Based on an agreement from 1958 between the cantons Argovia, Basle-Town
and Basle-Country regulations concerning navigation on the Rhine between Basle and
Rheinfelden also come under the authority of this office.
Railroads in the ports of both Basle are operated by Swiss Federal Railways on the
basis of an agreement between the cantons of Basle-Town and Basle-Country.
The ports of both Basle are not free-ports but inland harbors with facilities for
bonded-storage.
5.4.2. SHIP TRAFFIC IN THE PORTS OF BOTH BASLE
The ship traffic amounts to a total of 8,530,910 tons for the year 1992, respectively
8,353,401 in i991. The distribution of the goods was as indicated in TABLE 5.9 and
TABLE 5.10. The number of containers handled in 1991 amounted to 24,500 units
TEU (water side), respectively 60,000 units TEU (raiVroadiwater).
The transport capacity for Switzerland represents approximately 5 x 109 tons.kIn (or
ca. 60% of the goods transported by SBB-rail in 1991, which amounted to ca. 8.1 x 109
tons.km). This represents also 15% of the total Swiss foreign trade, which is handled by
ship transport on the Rhine.
The Swiss BUWAL (Federal Office for Environment, Forestry and Landscape) has
not requested a special risk assessment, investigation for the ports of both Basle, nor
have they proposed a particular methodology, but instead they request a short report
based on general risk considerations.
TABLE 5.9. Goods traffic by product category, at the ports of both Basle
Goods Traffic at Year 1991 Year 1992

Ute Ports of BoUt (tons) (tons)
Baste
Inflammable Liquids 3,878,971 4,381,828
Metals 1,266,753 1,222,169
Cereals 280,661 267,930
Coal 340,358 121,550
Scrap iron 190,355 272,131
Sand and Gravel 664,386 571,247
Other Goods 1,129,132 1,101,990

Source: Rheinschiffahrtsdirektion, Basel
TABLE 5.10. Dangerous Goods traffic by product category, at

both Basle ports
, Product Type Year 1992

(tons)
Coal 108,923
Brown coal 5,399
Crude oil , 20
Motor fuel and fuel oil 4,636,768
Other mineral oil products 15,978
Fertilizers 173,691
Chemical raw materials 43,484
Other raw chemicals derived 514

from petrol and coal
Other chemical products 78,347
Summary
The chapter introduces the methodological framework for performing a risk analysis of
hazardous materials transportation by ship in the context of regional risk assessment
and safety management Techniques for risk assessment of ship are outlined within the
methodology for assessing the risk of individual ports. A computational framework for
the accident frequency estimation is included. The method developed by EdF identifies
178 CHAPTER 5
the annual probability of a fluvial or sea accident with possible damages to people, or to
a chemical, or nuclear plant. Various calculation procedures for the pollution of water-
ways by dangerous cargo, and the risk management for waterways and ship cargo are
considered in this chapter, too. Examples and information concerning the ports of both
Basle (Switzerland) are also included.
1 Health and Safety Commission (HSC). Major Hazard Aspects of the Transport of Dangerous Substances,
Advisory Committee on Dangerous Substances, (1991).
2 Spouge, 1.R., The Use of risk assessment for ships carrying hazardous cargo in port areas, Conference on
Safety at Sea and in the Air - Taking Stocks Together, Royal Aeronautical Society, London, (November
1990).
3 Ale, B.I.M.; Whitehouse, R.J .• SAFETl: - Computer based system for risk analysis of process plants. in
Heavy Gas and Risk Assessment m,. Hartwig (ed.). D. Reidel, Dordrecht, The Netherlands. 5 (1986).
4 Hauer, Accident Anal. & Prev. 18 (1986), pp. 1-12.
S Spouge. I.R.. Techniques for risk assessment of ships carrying hazardous cargo in port areas. in F.
Saccomanno and K. Cassidy (eds.), Transportation of Dangerous Goods: Assessing the Risk, JnstilUte for
Risk Research (1RR). (1993). pp. 154-182.
6 Opschoor. G .• Methods for the Calculation of the Physical Effects of the Escape of Dangerous Materials.
TNO. The Netherlands Organisation of Applied Scientific Research, (1979).
7 Hendrich. S.; Lannoy, A .• Probabilistic and deterministic safety study on the sea-borne traffic of liquefied
gases in the vicinity of a nuclear site. in Loss Prevention and Safety Promotion in the Process industries.
IChemE 4th International Symposium on. 12-16 September. Harrogate. England. Vol. I - Safety in Opera-
tions and Processes, EFCE Publication Series No. 33 Vol. I. (1983). pp. DI-09.
8 National Ports Council (NPC. 1976). Analysis of Marine Incidents in Ports and Harbours. ISBN 0 90
1058939, (September 1976).
9 SAFETl (Software for the Assessment of Flammable. Explosive and Toxic Impacts): Ale. B.I.M.;
Whitehouse, R.J .• Computer based system for risk analysis of process plants. in S. Hartwig. D. Reidel
(eds.), Heavy Gas and Risk assessment m. (1986), Donlrecht. The Netherlands.
10 Cox. R.A.; CarPenter. R.J., Further development of a dense vapour cloud model for hazard analysis, in S.
Hartwig and D. Reidel (eds.). Heavy Gas and Risk Assessment. Dordrecht, The Netherlands, (1980).
CHAPTER 6
PIPELINES AND STORAGE TANKS FOR GASES AND PETROLEUM

PRODUCTS
The natural gas consumption for the year 1992 was worldwide around 2'150 Mrd m 3
(1 Mrd = 1 billion = 109), which corresponds to a primary energy consumption of 23%.
In Western Europe the natural gas consumption for the same period reached 320 Mrd
m 3 (i.e., 14% of the world consumption). This corresponds to primary energy consump-
tion of 17%. The need were covered to about two third by supplies from Western
Europe own production and to one third by supplies from Russia, Algeria, and Libya.
The most important production site in Western Europe is Norway (North-Sea), with
2'000 Mrd m 3 of assessed reserves. The supply from this region amounts to 30 Mrd m 3
per year, of which one part is distributed by pipeline to Great-Britain and the other to
Continental Europe, corresponding to 9% of the consumption of Western Europe. The
second largest producer are the Netherlands Groningen), with assessed reserves of
1,950 Mrd m3 natural gas. The natural gas is exported from Holland by "Gasunie" and
amounts to 40 Mrd m 3 per year. The natural gas pipeline network in continental Europe
has a length of over 600'000 kID and shows many ramifications, allowing a fine distri-
bution to most important consumers (Figure 6.1).
A significant fraction of hazardous materials is transported by pipeline, in particular

natural gas, LPG, ethylene, ammonia and chlorine.
6.1. Transport Pipelines in Switzerland
Since 1977, the "Swiss Gas" company is importing natural gas from the North-Sea via
the German supplier "Ruhrgas" and since 1974 from the Netherlands through
"Gasunie". Actually Switzerland is inter-connected to ten different pipelines entering
the country in the North, East, South and West direction, thus ensuring a safe and
regular supply. Switzerland is also a producer of natural gas, since a gas field was
discovered in 1990 in Finsterwald (TABLE 6.1).
In Switzerland the natural gas network, including distribution network, total 12'143
kIn, of which 4'548 kID form a high pressure pipework and 7'595 kID a low pressure
distribution network.
The main Swiss natural gas transport pipelines totalize a length of 1'886 kID. This
figure includes the length of the transit pipeline from the Netherlands to Italy, which has
on the Swiss territory a length of 165 km. Part of this pipeline has been installed on the
ground of the lake of Lugano (Agno - Capolago), other parts (32 kID) are also installed
in tunnels in the Alps region. The length of the primary Swiss Gas network amounts to
244 kID and that of the regional distribution network 1'477 kID (Figure 6.2).
182 CHAPTER 6
TABLE 6.1. Natural gas import and consumption in Switzerland
Natural Gas Import and Consumption in Switzerland for the Year 1992
GWb TJ
Swiss Production 29 110

(Finsterwald)
Natural Gas Import Swiss Gas 17729 63'820

Gasverbund Mittelland AG 3'011 10'840
Gasverbund Ostschweiz AG 1'903 6'850
Gasnat S.A. 2'065 7'440
Others 142 510
Total Natural Gas Imports 24'850 89'460
Sub-total Natural Gas 24'879 89'570

Supply
Town Gas production 62 220
Total Gas Supply 24'941 89'790
Market Share of Swiss 71.% 71.2%

Gas
The diameters and lengths of the main pipelines installed on the Swiss territory are
summarized in TABLE 6.2 (gas pipelines) and TABLE 6.3 (oil pipelines).
TABLE 6.2. Diameters of gas pipelines on the Swiss territory (after "Eidg. Rohrleitungsinspektorat",
Report SVDB No. 12411992)
Diameter 36" 34" 16" 14" 12 J14 " 10,J14" 8"·" 65/8 ' 4 112 • .3/8"
2
(Zoll)
Diameter 914.4 863.6 406.4 355.6 323.8 273.0 219.1 168.3 114.3 60.3
(mm)
Length (Ian) 93.8 70.7 254.3 48.9 228.4 484.4 252.7 113.2 64.3 0.8
The total installed pipelines lengths are: 1'611.5 kIn for gas pipelines, and 249.6 kIn
for oil pipelines. Depending on the size of the natural gas pipeline, the nominal
pressures are approx. 50, 64, 70 and 80 bar, respectively. Pipeline are usually buried
about one meter below ground. For most pipelines the variation of temperature is smalL
Internal corrosion depends on the material transported, but in many cases it is very low.
PWEL~ AND STORAGE TANKS 183
Figure 6. J. European natural gas transportation network.
TABLE 6.3. Diameters of oil pipelines on the Swiss territory (after "Eidg. Rohrleitwlgsinspektoru",
Report SVDB No. 12411992)
Diameter 22 20 18 16 123/4" loJ/4" - - 6S/8" 4 112"

(Zoll)
Diameter 555.8 508.0 457.2 406.4 323.8 273.0 209 193 168.3 114.3
(mm)
Length (Ian) 121.0 3.5 0.5 25.5 70.8 11.2 4.1 10.4 0.7 1.9
184 CHAPTER 6
Legend: ANatural gas users

•
---
Operating Plant
•
(i)
Propane/air mixing Plant (Town gas)
Natural gas occurrence (Finsterwald)
High-pressure pipeline (1.0. > 500 mm)
High-pressure pipeline (1.0. 350 - 500 mm)
High-pressure pipeline (1.0. < 350 mm)
Local distribution network
Projected or in construction
Figure 6.2. Overview of the gas pipeline network in Switzerland.
Gas pipelines and petroleum products pipelines are, in Switzerland, subjected to an

official control by the Swiss Federal Pipelines Inspectorship ("Eidg. Rohrleitungs-
inspektorat", Ziirich), and are regularly inspected for corrosion, both external and
internal. According to this Federal Office, there has, in the past up to 1994, never been
any accident involving fire or explosion of a pipeline in Switzerland. Most of the
damages to pipelines are due to mechanical action, as a result of earth-moving
equipment, other incidents are the result of mechanical failure or corrosion either to
pipelines (9) or at the gas station (1). Interval shutoff valves and check valves are
PIPELINES AND STORAGE TANKS 185
usually installed between two pipelines segments at a distance varying from 12 to 15

km (20 km in the Alps). Usually the gas pressure is around 60 bar and the gas bas a
density of 42 kglm3.
TABLE 6.4. Number of incidents/damages to gas pipelines announced to the Swiss Federal Pipelines
Inspectorship
Indclentl Number of Inddentsldamages to Gas PlpeHnes

Damage Year 1988 Year 1989 Year 1990 Year 1991 Year 1992
Mechanical 5x (of which - - - -
Damage 3 x Gas out-
(Natural Event) flow)
Mechanical - - 2x 6x 2x(of
Damage which I x
(Due to Third oil outflow)
Party)
Mechanical Failure - I x (Gas Ix (alGas Ix (alGas -

outflowal station) station)
station)
Damage to
Electrical Control
2x - - - -
Cables
Corrosion 9x Ix (Gas out- - - Ix

flowal
Station)
Other Causes - 4x 3x - -
TABLE 6.5. Crude oil and petroleum products pipeline network
ProdudlOll Pipeline Network
Oleodotto del Reno (oil)

Ol&lduc du Rh6ne (oil)
Ol6oduc Collombey-Chavallon (oil)
Ol&lduc du Jura Neuchltelois (oil)
AG~ (Pipeline for Petroleum Products, MW'8Z)
SAPRO (product pipeline)
Heavy oil pipeline, Comaux
Army pipelines (Oil, jet kerosine, heavy gas oil)
CARBURA pipeline (gasoline)
Connection to the Airport Zilrich-Kloten
IGESA (heavy gasoil pipeline. La Chaux-de-Fonds)
Every five years the pipelines are controlled internally with special equipment
recording any corrosion damage and wall thickness variation. At special occasions, the
186 CHAP1ER6
gas contained in a pipeline segment has to be vented to atmosphere (which takes a few
hours). A special computer program is used for calculating the venting time. Due to the
good conditions of these pipelines, which are recent, leakage problems are negligible.
TABLE 6.4 gives an indication of the small number of incidents, involving pipelines or
gas stations, that occurred in Switzerland between 1988 and 1993.
TABLE 6.6. Companies operating gas pipe lines in Switzerland
Operating Company Gas Pipeline Network I Segments

(from - to)
• Gtuverbund Ostschweiz AG • Degenau (Gossau) - Bischofszell
• Bilrglen - Ennetaach
• Balzers - Trilbbach (Swiss border)
• RUdlingen - Lottstetten
• RUmlang - Airport Kloten
• Obringen (Winterthur)-
Bischofszell (parallel pipeline in
planing, SOO mm 0)
• Gtuverb"nd Mittelllnd AG • LaufenthaIlThiersstein and Canton
Jura
• Pratteln - Arlesheim
• Bleienbach - Oberbuchsiten
• Villmergen - Lenzburg
• Connetion to the paper plant
Oftringen
• Derendingen - Biberist
• Naturgasleitung GW Bern
• Stiidtische Werke Luffl"O • Pipeline chiasso - Lugano

• Pipeline Trilbbach - Chur
• Trtuuilgtu AG • Wallbach - Dllniken (parallel

pipeline in planing, 900 mm 0)
• ErdgasllUlorpng Bind". • Trilbbach - Chur
RMilltlIl AG
• ErdgtU Zell1TtJbchwBir. AG • Ettiswil - Sursee
• MaIters - Hilnenberg
• Inwil - Peden (Root)
• GtwuItSA • Geneva SouthlLathoy (F) - Vernier
• La Cure - Gland
.U~SA • Ruswil - Altavilla
The geographic location of the pipelines and the revisions/modifications which were
made during the last years were reported in the yearly report of the Swiss Federal
Pipelines Inspectorshipl, and are mentioned in TABLE 6.5 - TABLE 6.6.
Beside the above mentioned pipelines, there are many other technical installations
which are not listed here, but include distribution stations, compressor/pumping
stations, gas and product storage, etc.
TABLE 6.7. Gas pipeline
Gas Pipeline Network
• Gasverbund Mittelland
• Gasverbund Ostschweiz
• Gas pipeline Holland-ltalien (Transit)
• Gazoduc Suisse Romand
• Erdgasleitung Ost
• Erdgasleitung West
• Erdgasleitung Rhonetal
• Erdgasleitung Zentralschweiz
• Gazoduc Orbe-Millchi (parallel pipeline to A1tavilla, 600 mm 121, in planing)
• Gazoduc Giand-Noville, VD, in planing (500 mm 121)
• Gazoduc A1tavilla - La Chaux-de Fonds - Les Verriaes
• Erdgalsleitung Finsterwald - Entlebuch
• Erdgasleitung Lichtenstein
• Erdgasleitungen der Gas-, Wasser- und Fernwllrmeversorgung der Stadt Bern
• Transitgasleitung
• Erdgasleitung Chiasso - Lugano
• Erdgasleitung Trilbbach - Chur
6.2. Risk from Pipelines
The risks from pipeline operation is not quite negligible as can be seen from. During the
perio11972 through 1994, according to the records taken from the data base PC-Fact;,
there were approximately 90 accidents taking place in the world, involving mostly
pressurized natural gas pipelines resulting in 516 fatalities and 1386 injuries (at the
minimum). These figures do not include Russia and Asia.
In Russia. a train induced a pipeline explosion in late '80s, causing the death of ca.
800 children. In Seoul (South Korea) a severe pipeline explosion took place in 1995.
Finally, a recent gas pipeline explosion in the USA (1994) endangered 100 people and
killed one person. Of these reported accidents, ca. 63 were explosions, the rest being
gas release, or gas release with fire. The main causes of fire and explosion were, among
others, due to: collision of tank vehicle (in one case), rupture of pipeline by bulldozer,
rupture of corroded pipeline by overstress, gas explosion in pump station. gas explosion
in building, gas leakage.
6.2.1. EXPERIENCE AND SITUATION IN THE USA
In the USA, pipeline transport is regulated by the DOT (Department of Transportation,

USA) through the Natural Gas Pipeline Safety Act 1968 and related regulations for
other hazardous materials. A code of practice for pipelines is provided by the Model
Code of Safe Practice, Pt. 6 Petroleum Pipelines Safety Code by the 1P (1967). In the
USA a pipeline code is provided by the ASME Code for Pressure Piping, Section B31.4
Liquid Petroleum Transportation Piping Systems (1974) and Section B31.8 Gas Trans-
188 CHAPTER 6
Transmission and Distribution Piping Systems (1975). There exist also several API
codes relevant to pipelines.
TABlE 6.8. Overview of pipeline accidents (including gas distribution). involving 10 or more
fatalities or injuries
Year Country Chemicals Fatalities I.quries Summary
1994 USA Natural gas 1 80-100 Explosion

1993 YU NaturaIgas >51 15 Heavy explosion
1989 USA Natural gas II 3 Sea vessel
1988 USA Ammonium 2 350 Fire and explosion in a
perchlorate rocket fuel plant
1987 Italy Natural gas 3 30 Gas explosion (in
house)
1987 Germany Gasoline 5 40 Collision of tank vehicle
1987 GB Natural gas 1 20 Gas explosion (in
house)
1985 France Natural gas 2 29 Explosion
1985 France Natural gas 5 38 Gas explosion (in
house)
1985 Belgium Natural gas 1 10 Gas explosion (in
house)
1985 Pakistan Natural gas 16 8 Explosion & fue
1985 USA Natural gas 6 14 Explosion & fife
1985 PL Natural gas 6 30 Gas explosion (in
house)
1984 USA Natural gas 5 23 Pipeline rupture
1984 France Natural gas 3 14 Explosion & fue
1984 Mexico Natural gas 9 37 Explosion
1984 BR Natural gas 61 25 Explosion & fue
1981 USA Natural gas 2 15 Gas explosion (in
hou,e)
1982 Italy Natural gas 6 10 Gas explosion (in
house)
1980 USA Natural gas 1 37 Explosion & fue
1979 PL Natural gas 45 70 Gas explosion (in
house)
1978 France Natural gas 12 45 Gas explosion (in
house)
1978 Mexico Natural gas 52 several Explosion
1978 Ireland Natural gas/- 7 40 Oil weIl on fue
crude oil
1978 USA Natural gas 6 47 Explosion
1976 Germany Natural gas 1 22 Gas explosion (in
house)
1976 Italy Natural gas 3 II Gas explosion (in
house)
1976 USA Natural gas 2 14 Explosion
Source: PC-Facts.
Information on the failure rate of pipelines is available in the USA from a number
of sources (Figure 6.3). These include the annual reports on natural gas pipelines by the
DOT, and those on oil industry pipelines by the oil industries international study group
for conservation, CONCA WE.
The main causes of failure of pipelines in the U.S. is interference by external action.
Damages can occur from earth-moving and excavating equipment. Some major acci-
dents arising from the pipeline transport of hazardous chemicals are the fire at Hearne
(fexas, 1972); the explosion at Natchitoches (Louisiana, 1965); at Port Hudson (Mis-
souri, 1970); the toxic release at McPhearson (Kansas, 1973). These has been described
and documented elsewhere as case histories.
Pipeline Surveys
• A survey on the U.S. Safety of Interstate Natural Gas Pipelines by the Federal Power
Commission (FCP, 1966) gives information on failures in these pipelines over the period
19501965. The length of transmission pipelines covered was 204,730 miles in 1964. The
number offailures was 1058 (of which 121 compressor or metering station failures; 1115
field testing failures). Employees suffered 35 deaths and 135 injuries and non-employee 29
deaths and 87 injuries. Of the 64 deaths, 46 arose from failures in lines an 17 from failures
in compressors or metering facilities. The major source of failures were: earth-moving
equipment (279 cases); corrosion (193),· weld failure (190).
• In the year 1974, there were 335,000 miles of pipeline installed and for the period 1970-
74 there were 2093 failures (Annual Report on the Administration of the Natural Gas
Pipeline Safety Act of 1968 by the DOT). From these data the failure rate for natural gas
pipelines was 1.25 x 10-3 per mile per year.
• The annual report HSpillages from oil industry cross-country pipelines in Western
Europe H, Statistical summary of reported incidents for 1976 by CONCA WE in 1977 (9m),
states that between 1972 and 1976 the total length of oil industry pipelines increased from
15,800 km to 18,100 /em and that in the period 1972-76 there were 93 spillages. Assuming
an average of 17,700 /em pipeline over this period, the failure rate for such pipelines was
1.05 x 10-3 per km per year.
• Further information on gas pipelines operated by the British Gas Corporation (BGe) has
been reported by Knowles, Tweedle and van der Post 3 (1977), for the period 1970-77:
there had been 309 repairable incidents, of which 73 were defined as Hlost gasH incidents.
For chlorine pipeline a failure rate of 5 x 10-3 per mile per year has been estimated for
1974.
Figure 6.3. Information on pipeline surveys.
Hydrocarbon Pipelines
The majority of hydrocarbon pipelines carry flammable gases or liquids such as natural
gas, ethylene or LPG. Two hazards which may result from a break in a line carrying
flammable gas are afire or an unconfined vapor cloud explosion. If line is ruptured and
the gas escaping is ignited so that there is a flame shooting up in the air, the heat
radiated from the flame may be considerable.
In many locations the hazard is, however, slight. The greater hazard is likely to be
that of an unconfined vapor cloud explosion (UVCE) resulting from an escape of gas or
liquid from the line. This hazard is presented particularly by propane which is one of
190 CHAPTER 6
the main flammable liquids transported by pipeline. If liquid propane at 16°C escapes to
the atmosphere, about 33% flashes off as vapor and the temperature of the remaining
liquid falls below -42°C. An example of hazard survey of a hydrocarbon pipeline
(Figure 6.4) is described in "A Safety Evaluation of the Proposed St. Fergus to Moss
Morran Natural Gas Liquids and St. Fergus to Boddam Gas Pipeline", by the HSE
(Health and Safety Executive, London, 1978).
The "St. Fergus pjpeliMs" (GB)

The "St. Fergus pipelines" (dia. = 406 mm) has a maximum operating prellure of 70 bar
and a minimum operating pressure of 18 bar, with a typical fluid temperature of LOOC. The
NGL (Natural Gas Liquids) would contain 50% ethane and 30% propane. The emission
rates from a rupture for the NGL pipeline were estimated as follow:
• Small orifices through which the maximum rate of emillion as liquid would be less than
100 kg/s, the maximum pump delivery rate would deliver NGL as liquid.
• Large orifices which could deliver more than 100 kg/s would give an initial liquid flow,
followed by choked two-phase flow as the pressure fell locally to the saturated vapor
pressure. At this lower flow the pressure would rise again and reversion to liquid flow
would occur. There would follow an oscillation between liquid flow and two-phase flow.
Estimates were made for the first 5 min. after the rupture.
• The dispersion of the gas cloud from the pipeline was calculated assuming a neutral
buoyancy gas and using the Pasquill equations. The effect of possible heavy gas behavior
was considered.
• The heat radiated from a flame ignited at the pipeline was calculated by the methods
given in API RP 521 (1969). Three heat radiation levels (1 kW/m2, 8 kW/"r and 16 kW/m2 )
were considered and the respective distances calculated for different release rates.
• The quanti~1 of vapor in concentration above the lower flammability Wl'S also estimated
for different Pasquill stability categories.
For a population 250 m away from the pipeline, the frequency of interaction for all releases
up to 100 kg/s was estimated as 4.3xl (J5 / year.
Figure 6.4. The "St Fergus pipelines" (G8).

The failure rates of comparable pipelines (in the USA, Europe, and UK) and the
factors influencing such failures are discussed in detail in that report. The incident rate
rr
of the pipelines was 2.32 x 1 per Ian per year, estimated on the basis of the relevant
data for the European petroleum pipeline system (CONCAWE 1977-9n7) and for the
UK gas pipeline system (provided by BGC).
A major unconfined vapor cloud explosion occurred at Port Hudson in 1970 (750 bblof
propane (ca..l20 m3) were released by a line break and gave an explosion equivalent to
about 50 t of TNT).
Ammonia Pipelines
In the USA, farmers are using ammonia as a fertilizer and there are some very
extensive liquid anhydrous ammonia pipeline systems installed. Description of the Mid-
American pipeline System are given by Rohleder4 (1969) and by Luddekes (1975), and
of the Gulf Central Pipeline System by Inkofer6 (1969).
There is a hazard of brittle fracture, therefore the composition of the anhydrous
ammonia must be closely controlled to prevent stress corrosion cracking. The ammonia
has to contain 0.2% water. Cases where brittle fracture has propagated along a pipeline
at sound velocity are known.
Example of the GulfCelltral Pipeline

The Gulf Central pipeline has a main diameter of lOin (254 mm) and a length of 548 miles
(ca. 882 km). The operating pressure is 100 bar. The temperature range is 2°C - 24°C.
Such pipeline is equipped with interval shutoff valves and check valves every 10 miles (16
km). At rather greater intervals there are pump stations with instrumentation which detects
any sudden changes of pressure caused by a line break and stops the pumps and closes
automatic shutoff valves.
• In the event of a large break due to a major pipeline rupture there would be an initial
pressure discharge of ammonia as the pipeline pressure vented. The line pressure would
then fall to that corresponding to vapor pressure of ammonia at the line temperature (e.g.
7.378 bar at 15.5°C).
• Flashing of ammonia would cool the remaining liquid and spasrrwdic ejection of liquid
and vapor would result. The proportion of ammonia flashing off from saturated liquid at
15.5°C and 7.378 bar is approximately 10%. The remaining liquid would be cooled to -
33.3°C and would form a pool. The evaporation from this pool would be in the range 1-7
IW and thus fairly slow.
• The quantity of ammonia which could escape at 15.5°C from a 16 km section amounts to
591 tons. Assuming 90% of this form a pool at -33.3°C , the resulting 708 m3 could form a
pool of 1,161 m2 with a depth of 61 cm, or one of 13,935 m2 with a depth of 5.08 em. The
vapor from an ammonia line break is initially cold and heavier than air. The vapor cloud
can travel in lethal concentration for several miles.
The basic procedure to contain a spillage is to dike around it. The line is excavated and the
rupture repaired. An incident involving a vapor leak on the pipeline near McPhearson,
Kansas (1975) is described by Ludekke 7 (1975).
Chlorine Pipelines
Pipeline transport of chlorine ;s much less extensive than that of ammonia. The chlorine
may be transported either in the vapor or in the liquid phase. The situation which might
give rise to mixed phase conditions should be avoided. The maximum operating
pressure is not to exceed 21.7 bar. Liquid chlorine has a very high coefficient of thermal
expansion and it is necessary to protect the line against hydrostatic rupture. The pipeline
may be above or below ground. Internal corrosion is very low if the chlorine is dry.
Chlorine pipelines are dealt with in Chlorine Pipelines by the Chlorine Institute
(1971, Publication No. 60). Further discussions are given by Danielson8 (1964), and by
Nelson9 (1964).
192 CHAPTER 6
6.2.2. PIPELINES INCIDENTS IN SWITZERLAND
In Switzerland there has never been any accident with a pipeline involving damage to
people or property. The risk arising from natural gas transport pipelines is relatively
small since in case of a small leak in a weld region, the methane gas which is lighter
than air, will escape and rise straight away in the atmosphere. In case of corrosion of the
pipe wall, or as a result of mechanical damage by earth-moving equipment, a jet may
however form (due to the high pressure of 50 to 60 bar existing in the pipeline) and
come into contact with an ignition source. Two hazards which may result from a break
in a line carrying flammable gas are a fire or an unconfined vapor cloud explosion.
The first case (fire) may happen with natural gas, whereas the second case (explosion) is
more probable with propane which may become heavier than air under special circum-
stances. In case of fire, heat radiation may be considerable. An example of pipeline risk
assessment has been described earlier in this chapter. The calculation of the conse-
quences resulting from a gas pipeline rupture may be complicated and necessitates
specialized programs. In Switzerland there are pipelines transporting natural gas, jet
kerosene/aviation gasoline, petroleum products and semi-products, however there are
no chlorine or ammonia pipelines installed.
Although Switzerland was. lucky to experience only minor incidents without conse-
quence, there has been abroad severe pipeline accidents involving natural gas explo-
sions, which took place during the last five years: Russia (late '80s), USA (1994), and
Seoul (1995). Unfortunate combination of incidents, usually related to large-scale pipe-
line fire and explosions, are causing severe accidents due to heat radiation or blast
effects.
6.2.3. HAZARD CONTROL
The risk associated with accidental releases from pipelines transporting hazardous
liquids over long distances can be extremely significant Integration of risk and
reliability analysis with advanced leak detection technology through the expert system
approach, can lead to substantial risk reduction goals.
This is the main objective of a CEC-funded project in the framework of the
"Environment Programme" undertaken by a Wode Group composed of SNAM-
PROGETTI (Italy), TNO (The Netherlands), ABA (UK), and TEA.
6.3. Storage Tanks
The object of storage of liquids and gases is to smooth fluctuations in the flow in and
out a production or terminal site. By far the largest quantities of chemicals are found in
storage facilities which may be at chemical plants or in other storage terminals. In order
to satisfy the storage requirements different storage method can be considered which all
have a different economical impact, due to the different storage design. Most of the
materials held in store are flammable liquids or liquefied gases and the main concern
in the design of storage installation for such liquids is to minimize the hazard of fire.
Loss of containment from storage of toxic chemicals could in principle give rise to the
worst kind of chemical industry disaster. Much storage is located in or near urban
areas. It is generally reckoned to be relatively safe and unlikely to cause large loss of
life. Beside the risk of casualties or fatalities due to fires in a storage, the resulting
financial loss may be very high. The prevention of loss in storage is therefore very
important Good operation and maintenance depends on an effective management
system, but there is potentially a problem in situations where products such as LPG
(Liquefied propane gas) or chlorine are stored and used in non-chemical and low
technology industries (chlorine stored in public swimming pools).
6.3.1. TYPES OF STORAGE
The main types of storage for liquids and liquefied gases are:
1) Liquid at atmospheric pressure and temperature, so called atmospheric storage
(0.5 - 15 psig). Fixed tanks with cone (diameter up to 76 m) and domed roofs (0
up to 20 m), floating roof tank (0 up to 76 m, height up to 22 m or more).
2) Liquefied gas under pressure and at atmospheric temperature (pressure storage)
3) Liquefied gas under pressure and at low temperature (refrigerated pressure
storage, semi-refrigerated storage)
4) Liquefied gas at atmospheric pressure ( < 1 psig) and at low temperature (fully
refrigerated storage)
5) Gas under pressure (> 15 psig up to higher pressures).
Note: Since the codes mentioned are P.ritish codes, of which some are still using British
units, we indicate here the conversion factor: 1 bar = 14.5 psi; 1 m3 = 6.2898 US oil
barrels
Storage of gas under pressure is usually done in horizontal cylindrical pressure
vessels. The need for very large storage capacities for LNG (liquid natural gas) has led
to the development of pre-stressed concrete storage tanks which are protected by an
earthen embankment, or berm, and are internally insulated. Underground cavities may
also be used for storage of natural gas (also depleted oil or gas field).
Atmospheric storage tanks are quite often inerted with N2 in order to reduce the
risk of fire and explosion. The fluids so stored are referred to as volatile liquids,
flashing liquefied gas, semi-refrigerated liquefied gas, refrigerated liquefied gas and gas
under pressure.
Jumbo storage tanks ranging from 300,000 to 1,000,000 barrels capacity are
currently being used to store not only crude oil but also flammable intermediate and
finished products.
A large number of incidents occur of the bursting or collapse of atmospheric
storage tanks due to maloperation. Atmospheric pressure is maintained in a storage
tank by a vent to atmosphere. The tank may be subjected to pressure or vacuum which it
cannot withstand not only if this vent is blocked but also if it does not have the capacity
to handle the flows required to equalize the pressure. A hazard which is liable to arise as
194 CHAPTER 6
a result of modification to the plant or storage is the invalidation of the arrangements

for pressure relief and blowdown or instrument systems.
The repair and demolition of tanks is another operation which has given rise to a
number of accidents and requires careful precautions. The principal hazard of work in
a vessel or confined space is that arising from toxic and/or flammable gas or vapor, the
presence of asphyxiating atmosphere and the accidental inflow of process fluids and
movement of machinery such as agitators. A significant proportion of accidents in
vessels is fatal and some involve multiple deaths.
The main guide to the conduct of plant repairs and modifications is API RP 510
Inspections, rating and repair of pressure vessels in petroleum refinery service (1975).
Effect pf Fire on Storage Tanks

There are two main effects of fire on an atmospheric tank. One is that the parts of the
tank which are not cooled by the liquid inside may become hot and weaken. The other
is that the liquid in the tank is heated up and its vaporization is increased. The effect of
fire on a pressure vessel is broadly similar to that on atmospheric tanks. The vessel may
burst due to loss of strength. In such a case overtemperature is just as serious as
overpressure (e.g., see the disaster at the refinery of Feysin (F), where a spherical vessel
containing propane gas took fire and burst under the effect of heat).
The effect of fire on a refrigerated atmospheric storage tank is similar, but the
heating and vaporization of the product is much more rapid. Fire can also have a very
damaging effect on pipework, which sometimes may withstand the fire for only about
10 min. A small fire at a leaking flange can have a torch effect like in the Flixborough
disaster in Great-Britain.
The fo!lowing considerations allow a qualitative appreciation of the risk involved:
Risk Presented by Stomge Vessels

• A leak of a volatile liquid held at atlOOSpheriC temperature and pressure results only in a
slow evaporation of the liquid.
• Escape of a refrigerated liquefied gas at atlTWspheric pressure gives some initial flash-off
and then an evaporation which is relatively slow but faster than the first case.
• Loss of containment of a liquefied gas under pressure and atlTWspheric temperature,
however, causes immediate flashing off of a large proportion of the gas, followed by a
slower evaporation of any residue, and it is usually much the IOOSt serious case to
consider for risk analysis.
• It is generally considered, that there is a greater hazard in storing large quantities of
liquefied gas under pressure than at low temperature.
There are numerous standards, codes of practice, and guidelines which are
applicable to storage. 1O Hazard assessment studies have been also reported 11. 12. 13. 14 as
well as case histories. IS The definitions and codes indicated below were adopted from a
British transport compendium.
Filling Ratio
The maximum permissible working capacity of a storage vessel is often expressed in
terms of the filling ratio. Filling ratios are dealt with in BS 5355: 1976 Specification for
filling ratios and developed pressures for liquefied and pennanent gases, and in other
publications on storage (e.g., ICI/rasp, 1970 ISn4; LPGITA, 1974 LPG Code 1;
BCISC, 1975 Item 1).
The formula used to calculate the filling ratio is generally of the form:
e = rp / rw{l- f) (6.1)
where e is the filling ratio; rp the specific gravity of product at reference temperature; rw
the specific gravity of water at 15° C; andf, the free space (fractional). The reference
temperature is the maximum temperature which the product should reach in service.
Segregation and Separation Distances

The segregation and separation of materials within the storage area is largely based on
classifications of the material stored, on electrical areas, and on fire protection
measures.
Volatile liquids are classified according to their flash point into Classes A, B and C.
This classification is less suitable for liquefied gases. Petroleum products such
petroleum spirit and other flammable liquids of similar volatility are stored with
segregation of Class AlB, other liquids belong to class C. The former require full bunds,
but the latter do not. The total capacity of fixed roof tank and of floating roof tanks
within a bund should not exceed 60,000 m 3 and 120,000 m 3 respectively.
Minimum recommended separation distances for storage (min. 6-15 m up to half
diameter of larger tank) are given in the various codes and other publications. Typical
separation distarJces for petroleum products are those for Class A and Class B liquids,
excluding LFG, given in the IP Refining Safety Code (1965). The separation distances
for LPG and LFG are generally greater (pressure storage: 7.5 - 60 m; low pressure
refrigerated storage: 15 - 90 m).
Calculations on separation distances related to heat effects are usually based on
direct flame impingement and on heat radiation. It is assumed that in a storage tank fire
the tank is effectively a pool of liquid with a flame burning on this Equid. If the tank
diameter is 0, the height of the flame in still air is 20. If there is a wind, however, the
flame is distorted. At a wind speed of 2 mls it has been observed that the flame is
deformed to an angle of 45° to the horizontal and that on the downwind side of the fire
the flame hugs the ground for a distance of about 0.50. For the heat radiated from the
flame it is usually assumed for simplicity that still-air conditions prevail. The
application of such engineering calculations to the separation distances for storage tanks
has been treated by Robertson l6 . The heat radiated from the envelope of a flame burning
on a liquid pool in a storage tank is:
(6.2)
where A is the surface area of pool (m2); C the calorific value of liquid (kJ/kg); Q
the quantity of heat radiated from the flame envelope (kW); r the liquid burning rate
(mls); p the density of liquid (kglm3); and kJ a constant equal to 0.3.
196 CHAPTER 6
The values for the liquid burning rates are given as: 0.083-0.17 mm1s. Alternatively
values of 0.1,0.3, and 0.4 inlmin are also given for 1 ft,7 ft, and 35 ft diameter pools,
respectively. For most hydrocarbon flames the heat radiation from the surface of the
cylinder (with dia. = D, height = 2D) is approximately 170-240 kW/m2. The limit for
the heat received by an adjacent storage tank is assumed to be 37.8 kW/m2.
Note: I ft = 0.3048 m; I inch = 0.0254 m.
Heat Received by Pressure Vessels

Two methods of determining the heat received by pressure vessels are also presented in the
ICI LFG Code, but these are not used in the code to determine separation distances. An
alternative approach is to base separation distances on engineering principles. Two main
factors which should determine separation are (a) heat from burning liquid, and (b)
ignition of a vapor escape.
Bunds
Some types of liquid storage tank are normally surrounded by a bund and/or provided
with a pit to retain any spillage of the liquid. Bunds are made of earth or of concrete. In
general, bunds are provided for atmospheric storage tanks and for fully refrigerated
storage tanks of liquefied gas, but not for pressure or semi-refrigerated storage of
liquefied gas or for acid or alkali storage, although this generalization needs some
qualification.
Special ConsidertUions Regarding Bunds for Ammonia Storage

• It is traditional to put bunds around all types of ammonia storage, but this practice has
Cfen questioned regarding pressure storage on the grounds that in this case sudden loss of
containment tends to result in ejection of all the contents in the form of vapor, or spray
leaving no residual liquid.
• Escape from refrigerated storage, however, results in immediate flashing of only about
2 % of the liquid. Even a primitive bund should hold some 90 % of the liquid, while a tall
bund as high as the tank should retain 98 %.
• In the case of ammonia storage, it is not enough to catch the ammonia liquid released. If
the surface area of the bund is large, the evaporation of liquid ammonia is still a serious
hazard. It is therefore desirable to reduce the surface area. This may be done by the use of
controlled drainage which runs the liquid off to a pit where its evaporation area is reduced.
Foam blanketing is another possibility suggested, provided the area to be covered is not
excessive.
The object of buntiing is to retain the liquid so that it can be dealt with in a
controlled manner, by evaporation from a specially designed catchment evaporating
area, by foam blanketing or other measures. Bunds tend not to be used for pressure
vessels because these rarely fail, the emission when it does occur is mainly in vapor
spray form, and the dispersion of small leaks and spillage is hindered. Low walls may
be used to keep flammable liquids from some external source from reaching a storage
vessel.
For atmospheric storage tanks containing Class AlB flammable liquids a full bund
should be provided. The bund capacity should equal the total capacity of the tank. If
there is more than one tank inside the bund, the capacity should be that of the largest
tank after allowing for the displacement due to the other tanks. Low division walls
between tanks within a bund are recommended.
Some aspects of the storage of particular chemicals are considered in the next
section.
6.3.2. STORAGE OF SPECIAL PRODUCTS
LPG Storage
LPG is stored at petroleum storage terminals, at chemical works and at many non-
chemical factories. In addition to the codes which refer explicitly to LFG (Liquefied
Flammable Gases), methods of minimizing spillage and hazards of fires can be found in
the ICI Liquefied Flammable Gases Storage and Handling Code (the ICI LFG Code)
ICIIRoSPA 1970 ISn4).
Chlorine Storage
Storage of chlorine as a liquid is mainly either in chemical works or at installations
where it is required for water treatment such as waterworks, power stations and
swimming baths. The boiling point of chlorine is 34° C. The handling and storage of
chlorine is dealt with in the BCISC Code of Practice for Chemicals with Major
Hazards: Chlorine (the Chlorine Code, 1975 Item 1), and in the various publications of
the Chlorine Institute, including Chlorine Manual (1969 Pub. 1), Facilities and oper-
ating procedures for chlorine storage (1973 Pub. 5), and other literature (1965 Pub. 57,
1970 Pub. 9, 1974 Pub. 53, 1975 Pub. 61).
Chlorine Code
The Chlorine Code gives a well defined philosophy for the storage of chlorine. In view of
the importance of the hazard of toxic release from storage the code's statement of the
problem is of interest: "liquid stored in bulk is the potential source of the most serious
incidents. Stored liquid could be released by failure of the vessel holding it, or by failure of
pipework or equipment connected to the vessel". The risk of some damage to a storage
vessel cannot be totally eliminated. the main dangers derive from failure of pipework
whether on producing or using plants or connected to storage.
Chlorine is typically stored in vessels up to about 400 t capacity. For larger

quantities up to several thousands of tones refrigerated atmospheric storage is used. The
main hazard of chlorine in storage is a large toxic release which could cause a major
disaster. The containment of chlorine is therefore critical.
Ammonia Storage
Anhydrous ammonia is stored as a liquid mainly either in chemical works or,
particularly in the USA, in storage terminals supplying agricultural users. The handling
198 CHAPIER6
of ammonia, including its storage, has been a major preoccupation of the industry and
this is reflected in the Chemical Engineering Progress series Safety in Air and
Ammonia Plants (AlChE 1960-69 Items 12-21) and Ammonia Plant Safety (and Related
Facilities; AIChE 1970-77 Items 25-32). The work of the industry Ammonia Storage
Committee has been described by W.L. Ball (1968) 17. There is a CIA Code of Practice
for the Large Scale Storage of Fully Refrigerated Anhydrous Ammonia in the United
Kingdom (the CIA Refrigerated Ammonia Storage Code) (1975 Item 8). The boiling
point of ammonia is - 33° C.
As with chlorine, the principal hazard of ammonia in storage is a large toxic
release, since a large release could give rise to a major disaster. The containment of
ammonia is therefore critical. With ammonia also the design philosophy is that the
standards should be such that vessel failure is a rarity. A review of ammonia storage
methods has been prepared by (Hale 18, 1974). The larger refrigerated atmospheric
storage tanks hold 40,000 ton.
Hydrogen Storage
Gaseous Hydrogen. The gas is stored in pressure vessels designed to the regular codes
NFC 50 A, "Gaseous Hydrogen Systems at Consumer Sites", (1973), gives separation
distances. The separation distances for a large gaseous hydrogen storage (> 15,000 scty
are 25 ft from an ignition source and 50 ft from a large flammable liquid storage ( >
1000 US gal). Pressure relief to the open air is permitted provided this is done at a safe
height.
Liquid Hydrogen. The liquefied gas is stored in pressure vessels designed to the
regular codes. Separation distances given in NFC 50 B "Liquefied Hydrogen Systl!ms at
Consumer Sites" (1973) for a large liquid hydrogen storage (> 15000 US gal) are 50 ft
from an ignition source and 100 ft from a flammable liquid storage. Again pressure
relief to the open air at a safe height is allowed. NFC 50B states that liquid hydrogen
should preferably be stored on ground higher than the other flammable liquid storage.
This, of course, is the exact opposite of the requirements for storage of products such as
LFG.
Hydrogen Storage HaZllrti

There appears to be some disagreement on the desirability of providing a protective wall
around a liquid hydrogen storage vessel. NFC 50 B refers to reductions in separation
distances allowable if there are walls. On the other hand, Scharle 19 points out that an
unconfined hydrogen gas cloud does not apparently detonate, but a confined cloud can.
Hydrogen flames are practically invisible and may only be detected by the heat radiated.
This constitutes an additional and unusual hazard to personnel which needs to be borne in
mind in designing an installation.
i 1 scf(standard cubic feet) = 2.832 10.2 m3, and 1 US gallon = 3.785 10- 3 m3•
Summary
The risk assessment of pipelines as well as storage tanks for gases and petroleum
products is of emerging importance for regional risk assessment and safety manage-
ment. The present chapter first surveys the transport pipelines in Switzerland: gas
pipeline networks in the content of the European environment, technical characteristics
of the gas and oil pipelines, number of incidents and damages to gas pipelines. The risk
analysis of pipelines is made within the framework of experience in the USA as well as
in Switzerland. Regarding the risk of storage tanks, the present chapter mentions
international codes of practice and regulations, and presents some methods of
calculation; storage of special products and their specific risk aspects are part of the
present analysis.
200 CHAPTER 6
1 SVDB. Schweizerischer Verein fiir DruckbehiilteriiberwachlDlg ("Eidg. Rohrleitungsinspektorat".

Freiestrasse 88. 8032 ZUrich. Switzerland): Jahresbericht (1988•...• 1992).
2 PC-Facts database: AC-Laboratorium, Wimmis/Spiez. Schweiz: Contact person is Dr. P. Roder (AC-
Zentrum. Tel: 033-55 11 11).
3 Knowles. A.E.; Tweedle. F.; van der Post. lL.. The background and implications of IGE/ID/1 Edition 2.
Comm. 1044. Institution Gas Engineers. London (1977).
4 Rohleder. G.V .• Considerations in an ammonia pipeline system. in Safety in Air and Ammonia Plants. vol.
11. (1969). pp. 35.
l Ludekke. D.E .• Ammonia pipeline maintenance and repair. in Ammonia Plant Safety. vol. 17 (1975). pp.
99.
6 Inkofer. W.A .• Ammonia transport via pipeline. In Safety in Air and Ammonia Plants. vol. 11 (1969). pp.
40. (See also Chern. Engng Prog .• 65 (3). (1969). pp. 64).
7 Ludekke (1975): ibid.
8 Danielson, G.L., Handling chlorine - 1. Tank car quantities. Chern. Engng Prog., 60 (9). (1964), pp. 86.
9 Nelson, H.P., Handling Chlorine - 2. Barge and pipeline safety. Chern. Engng Prog., 60 (9), (1964), pp. 88.
10 Lees. F.P., Loss Prevention in the Process Industries, vol. 2. Butterworths, London, (1980).
11 Husa, H.W.; Bulldey, W.L.. Hazards of liquid ammonia spills from low pressure storage tanks, in Safety in
Air and Ammonia Plants, Vol. 7 (1965).
12 Siccarna, E.H.• The environmental risk arising from bulk storage of dangerous goods in port areas, in
Transport of Hazardous Cargoes, Vol. 2, paper 10 (1971).
13 Solomon, K.A.; Rubin. M.; Okrent, D., On Risk from the Storage of Hazardous Chemicals, University of
Calif., Los Angeles, Rep. UCLA-ENG-76125 (Los Angeles, Calif.), (1976).
14 Health and Safety Executive, Canvey: An investigation of Potential Hazards from Operations in the Canvey
Island IThurrock Area, HM Stationary Office, London (1978).
Il Lonsdale. H .• Ammonia tank failure. South Africa, in Ammonia Plant Safety, Vol. 17 (1975), pp. 126.
16 Robertson. R.B., Spacing in chemical plant design against loss by fire. In Process Industry Hazards, (1976),
pp.157.
17 Ball. W L., Status report of the ammonia storagl' committee. In Safety in Air and Ammonia Plants, Vol. 10,
(1968), pp. 38.
18 Hale, C.C., Amonia storage design practice. In Ammonia Plant Safety, vol. 16 (1974), pp. 23.
19 Scharle, W.J., The safe handling of liquid hydrogen. Chern. Engr.• London, Vol. 185, CE16, (1965).
CHAPTER 7
TRANSPORTATION OF DANGEROUS GOODS

- GENERAL INFORMATION AND STATISTICS
7.1. Information for Regional Planing
Risk assessment for regional planing should also comprise information and statistics on
the volume and class of dangerous goods transported or stored in large quantities in the
region. This information should be stored in a dangerous goods data base which should
include:
• Information concerning the storage location, such as: location of the storage area,
type of storagelpackaging (tanks, containers, bags), tank farm arrangement, separation
distances, oil catchment equipment, etc.
• Information on the products: volume of product stored, danger class and physical
properties of the liquids or solids.
• Information for medical treatment of patients
• Information concerning the fire fighting equipment (type/number of hydrants,
typeIquantities of extinguishing product) and information regarding the availability of
firemen brigades (on a local/regional basis).
• Regional information concerning the availability and location of special protection
equipment against petroleum products, for pollution prevention and remedial actions.
Using a Geographical Information System (GIS) it is possible to visualize the
location of dangerous goods stored in large quantities in the region. An overview of the
volume and main transport flows of dangerous goods through the region might also be
represented on a synoptic map. The map should show the main transport and distribu-
tion roads through the country, and indicate the location of petrol stations along the
highways and in the cities. Additional information concerning the presenCe of highways
restaurants, the location of schools, sport stadium. hospitals, shopping centers in the
vicinity of such critical points would be useful as well. The population density along
the mains transport axes should be estimated and reported on the digital map, using a
grid of 100 x 100~. Such information is helpful in calculating the consequences of an
accident involving fire, toxic gas dispersion, or explosion, and gives also information
on the number of people to evacuate in case of an emergency.
At present such development in Switzerland is in progress and the Cantonal Planing
Offices are in the process of preparing digitized maps for use with GIS, but it will take
a few more years until the tas is accomplished. A strong coordination between the
different Cantonal Offices would be highly desirable, especially what concerns the type
202 CHAPTER 7
and format of information to be stored in regional data bases. Remote public access to
such data bases at a National level should be favored. The public should have the right
to have access to such information. as this is the case in the USA and in other countries.
Concepts for integrated network access of risk data bases and other regional data
bases of interest, using hypertext and special query languages of the fourth generation,
are being investigated in the United States and in Germany (on behalf of the European
Community).
In the framework of the PPR&S it was decided to collect and review basic
information of legal and technical matter concerning the transportation of dangerous
goods in Switzerland, as far as available. At this point it is worth noting, that there is no
law obliging trucking companies, or the Swiss railroad companies to publish any kind
of statistical information or to report publicly any kind of incident/accidents concerning
the transportation of dangerous goods. Some basic information concerning the
transportation of dangerous goods in Switzerland has been reported in Gennan
language by Kieselbach l (1993), in the framework of the PPR&S project, and selected
topics are also presented here.
7.2. Dangerous Goods Movements in Switzerland
7.2.1. SOURCES OF STATISTICAL DATA
It is rather difficult to obtain reliable data for use in risk assessment studies, concerning
the number of road tankers accidents, or statistics regarding the type and quantity of
dangerous goods transported through Switzerland National statistical data can be
obtained, for instan-=e, from the following source:
• Swiss Federal Office of Statistics ("Bundesamt fijr Statistik, Bern").
• "Statistisches Jahrbuch der Schweiz", an agenda with statistical data concerning
Switzerland
• Swiss Railway Company ("SBB, Schweizerische Bundesbahnen")
• Petroleum Products Association ("Erdolvereinigung")
• Fuel Dealers Association ("Brennstofthandlerverband")
• Pipeline operating companies like INTERGAS, SWISSGAS etc. have statistics
concerning products/quantities distributed by pipelines
• Swiss Federal Pipelines Inspectorship (statistics concerning pipelines incidents)
Within the framework of PPR&S we have tried to obtain from different sources
various information concerning dangerous goods transportation, e.g., from the Swiss
Petroleum Products Association, and the Swiss road tankers/truckers companies, etc.,
but without success, since no Swiss law does oblige such companies to disclaim any
information concerning the type, quantity, and destination of chemical products or other
dangerous cargo (gasoline, heating oil, industrial gases, etc.) transported by them. The
companies in question are considering such "marketing information" as confidential,
TRANSPORTATION OF DANGEROUS GOODS - GENERAL INFORMATION 203
and are not prepared to divulge them to any third party. The problems we encountered
are mention below.
a) There are very little statistics on road tankers/trucks accidents that are published in a
form useful for risk analysis. Data published abroad cannot be extrapolated, as such,
to Swiss conditions, and must be considered with prudence.
b) The quantities of goods imported in transit (by rail, road, or ship) or transported
within the country are registered at the custom-houses according to a schedule of
dues developed for fiscal purpose deserving more the preparation of financial
budgets by the Swiss government. The records cannot be used easily for risk
assessment because the quantities of various products, that do not belong to the
same class of risk, have been aggregated together.
c) Some petroleum products imported by international and diplomatic organizations, as
well as jet kerosene for the international air traffic are not subject to payment of
duty, and therefore do not appear in the statistics of the Swiss custom-houses,
respectively, of the inland revenue department.
d) The ports of Basle (on the Rhine river) represent a special case, which complicates
the statistical evaluation of the distribution of the goods traffic between ship, rail
and road. Part of the goods transported by ship are unloaded in Basle to be further
transported by rail or road for inland distribution. Record keeping by the custom-
houses does not mention the final destination of the (dangerous) goods.
e) In Switzerland we have strategic tank farm'! for petroleum products (automotive
gasoline, heating oil, jet kerosene). Imported petroleum products may be stored for
some time in such storage areas located at proximity of air ports and cities, or they
may be brought directly to the final user. Redistribution may also take place from
such storage. Thus it is not easy, without precise record keeping, to trace back the
movement of dangerous goods involving petroleum products.
t) Inland distribution of petroleum products is not always done by the oil companies,
like ESSO, SHELL, etc., but frequent use is made of private road tankers compa-
nies.
Furthermore, due to the very large autonomy of the different Cantons, several
regional statistics of interest mayor may not be made available officially to the public,
due to a new legislation on data secrecy/privacy that obliges the different Federal and
Cantonal offices to handle their data on a "confidential basis". Another difficulty arises
from the fact that the denominations of the different offices may also vary considerably
from one Cantonal Office to another, which does not simplify data inquiry.
Further statistics are available at the Cantonal level: i) - for information
concerning population density distribution, usually the Communal or Cantonal
Cadastral Office; ii) - for traffic accident statistics, the Cantonal Police Authorities,
or eventually some Cantonal or Urban Office for Statistics, or Civil Engineering Office
("Tiefbauamt"). The latter data are used for planing and taking remedial actions
pertaining to roads and traffic control.
204 CHAPI'ER 7
Some data on petroleum products imports/movements are available from yearly

reports and other kinds of publications by the oil and gas companies, and similar
associations. The oil companies operating in Switzerland do not distribute all of their
products themselves, but have agreements with trucker companies. Since record
keeping of such transports is not expressly required by legislation, the companies do not
exactly know when and where which products were transported. The trucking firms as
well as most of the trading companies are not willing to discover such information to
third parties. The reasons are that administrative record keeping was not foreseen for
such a case and disclosure of marketing information is not usual in this kind of
business. Another reason is that some products are in transit through Switzerland either
by train, road, tanker ship or pipeline, others (mainly petrol, kerosene, and heating oil)
are stored in large tank-farms which are part of the national strategic storage system.
Such tank-farm areas are found in the surrounding of large cities, such as Basle,
Geneva, Chiasso, at the air ports and some other locations. Distribution of the
petroleum products can take place directly from the border or from intermediate storage
or from the strategic tank-farms. Such products movements are difficult to trace without
an efficient book-keeping system.
7.2.2. DANGEROUS GOODS CLASSES
There are actually 10,000 products that are known to be dangerous and are being
considered by the different regulations for industry, storage, and transport. Usually the
products are classified by dangerous goods classes. According to the UNO
recommendation the products may be simply listed and numbered, or according to
RID/ARD, they have to be classified on the basis of their physical and chemical
properties and grouped in different classes according to some main properties repre-
sentative of this type of dangerous goods.
RIDIARD Regulation
The RIDIARD regulation prescribes a subdivision into 9 classes (1-9) according to
TABLE 7.1. The most important classes are the classes 3, 6.1 and 8. All the RID/ARD
classes (except the classes 2 and 7) are further subdivided into three subdivisions giving
an indication of the acuity of the danger (Le., a- extremely dangerous, b- dangerous, c-
less dangerous).
Packaging Groups/Codes
In the regulations we find, according to the same scheme, the packaging groups I-llI
corresponding to the danger class (a-c). The danger classes with their corresponding
packaging groups and code a presented in TABLE 7.2.
The significance of this packaging code is that the packaging has to comply with
some min~mal exigency corresponding to the danger class. Transport of dangerous
goods by post is principally forbidden. According to international regulations, it is only
allowed to transport dangerous goods internationally, when these are properly packed in
accordance with the packaging groups and codes corresponding to the representative
danger class. The definition of the tanks and containers to be used in conjunction with
TABLE 7.1. Classification of dangerous goods according to RID/ARD
RID/ARD Product Description

Class
1a Explosive product and objects
1b Objects containing explosive products
lc Ignition products, fireworks, and similar products
2 Gases compressed, liquefied or dissolved under pressure
3 Flammable liquid products
4.1 Flammable solid materials
4.2 Self-igniting products
4.3 Products which, when in contact with water, develop flammable

gases
5.1 Oxidizing products causing ignition
5.2 Organic peroxides
6.1 Toxic products
6.2 Contagious or disgusting products
7 Radioactive productsy
8 Caustic products
9 Different dang~T<?U:S products and objects
TABLE 7.2. Danger classes with their corresponding packaging groups and codes
Danger Class Packaging Group Packaging Code

a) Extremely Dangerous I X
b) Dangerous n X,Y
c) Less Dangerous n X,Y;Z
RIDIARD have been described in detail by Kieselbach 1 and are not reproduced here.
The dangerous goods to be transported must be marked accordingly (following the main
classification scheme). The road tankers loaded with dangerous goods must be provided
with an orange colored label. The danger class is defined by a number located in the
upper part of this label and a second number (so-called UNO number) in the lower part
gives an indication of the content of the tank.
206 CHAPTER 7
7.2.3. TYPE AND QUANTITY OF GOODS TRANSPORTED IN SWTIZERLAND
An estimation of the q~tities of dangerous goods transported in Switzerland was

prepared by Kieselbach i from different statistical sources. The Swiss Railways'
Company (SBB, "Schweizerische Bundesbahn") has published the data reported in
TABLE 7.3. The distribution shown in this table is based on a quantity of 8 million tons
of problematic (dangerous) goods. In a similar statistic published by the Swiss Federal
Office of Transportation ("Eidgenossisches Bundesamt fUr Verlcehr, BAV") we found
similar data distribution, assuming 20% of dangerous goods for a total of ca. 50 million
tons (Mt) transported.
TABLE 7.3. Rail transports of dangerous goods (SBB Statistics, 1989)
Product Category Quantity Distribution

(Mt) (%)
Hydrocarbon Compounds 5.44 68

(Petrol, Heating Oil)
Raw Chemical Products 1.52 19
PhannaceuticallDyes 0.32 4
Fertilizers 0.32 4
CompressedlliqueflCd Gases 0.24 3
lifM.: the tonnage in the folWwing tables is upressed either in million tOIU
(Mt). or tlroustJ1ld toIlS (let). or toIlS (t).
The Swiss Federal Railways are not only transporting goods by train, but offer a
special customer service, SBB-CARGO, for delivering the goods by truck to the
customers' place.
According to an information published by SBB-CARGO (111993), 50 tons of goods
were transported in transit or inland, of which 7 tons were dangerous goods. The period
was not mentioned but this statistic probably refers to the year 1992 or possibly 1991.
The distribution of the dangerous goods transports between rail and road is shown in
TABLE 7.4.
TABLE 7.4. Distribution of dangerous goods' transports between rail and road
(according to SBB-CARGO. 1/1993)
Danger Category Rail Transport Road Transport

Distribution in % Distribution in %
Extremely Dangerous 43 57
Less Dangerous 38 62
The proportion representing rail transports was distributed as follows by product

category (TABLE 7.5).
TABLE 7.5. Rail transport statistics by product category
Rail Transport by Products Category Quantity (Mt)
Petroleum Products 5.4

Extremely Dangerous Products 0.4
Less Dangerous products 1.2
TABLE 7.6. Rail transport statistics by product category (according to Swiss Federal Office
of Transportation; BAV, 1990)
Rail Transport by Products Category Quantity Distribution

(Mt) (%)
Hydrocarbon Compounds (petrol, 5.9 59

Heating Oil)
Chemical Raw Products 1.5 15
PhannaceuticaU1)yes 0.2 2
Fertilizers 0.2 2
Compressedlliquefied Gases 0.2 2
The figures of TABLE 7.6 have been reported by the Swiss Federal Office of
Transportation (BAV, Schweizerisches Bundesamt fur Verkehr). The problem with this
kind of information reporting is that the content of the category/class is not defmed at
all and does not correspond to the definitions uSf_i by RID/ARD for the classification of
dangerous goods. The United Nations Organization, UNO, has published in 1989 a
statistic regarding the transportation of different kinds of goods. The data concerning
Switzerland are reproduced in TABLE 7.7.
TABLE 7.7. Dangerous goods transports in Switzerland (according to UNO-statistics)
Dangerous Goods Transportation Medium
D.G. Product Category Rail Road Ship Air Pipeline Total

Class (Mt)
11 Solid Combustibles 77 106 284 - - 572
12 Petroleum Products 2238 922 3915 10 5538 12570
12.1 Crude Oil - - - - 3013 3845

12.2 Petroleum Products, 2238 922 3915 10 2525 9125
Gas
13 Tar & Raw Products 22 11 2 - - 34
14 Chemicals 1334 2196 177 4 - 3475
15 Concrete 196 839 34 I - 1853

208 CHAPTER 7
TABLE 7.8. Statistics on dangerous goods' distribution according to different

codes (Survey established by the Swiss Society of Chemical Industry)
Type of GoodsIDistribution Quantity Quantity

(according to given code) (tons) (%)
Dangerous Goods 699'946 34.1
SDR 239'429 34.2
ADR 59'568 8.5
RSD 263'788 37.7
RID 103'098 14.7
IMDG 32'432 4.6
leAO 1'546 0.2
Total Freight 2'054'387 100.0
Another statistic compiled in 1987 comes from the Swiss Society of the Chemical
Industry (SOCI, "Schweizerische Oesellschaft der chemischen Industrie") and was
established on the basis of a questionnaire which was send to 28 member companies
(TABLE 7.8).
The distribution of the SOCI data according to RID/ARD categories is shown in
TABLE 7.9.
TABLE 7.9. Statistics on dangerous goods' distribution accord-
ing to dangerous goods classes (Survey established by SGCI. the
Swiss Society of the Chemical Industry ) in 1987
Dangerous Quantity Quantity

Goods' Class (tons) (%)
1 1'078.0 0.2
2 25'437.5 3.6
3 99'149.7 14.2
4.1 21'197.0 3.5
4.2 0.5 0.0
4.3 0.5 0.0
5.1 6'063.0 0.9
5.2 4'700.0 0.7
6.1 55'027.0 7.9
7 0.0 0.0
8 483'889.0 69.1
TOTAL 699'546.2 99.9
The SOCI statistics present a difference of 400 t. It seems probable that the transport
of petroleum products has been omitted or reporting was incomplete.
TABLE 7.10. Variarion of the national economic index (period

1987-90)
Year Index Basis Change in %

1987 179 100
1990 223.8 125
SGCI has also reported a change of the Swiss national economic index for the period
1987 to 1990 (TABLE 7.10), and it seems reasonable to admit that the quantities have
also increased proportionally to this index change.
TABLE 7.11 show an approximate distribution of the goods volume between ship
and rail, based on data published for Germany and extrapolated to Switzerland
conditions. Within Switzerland, the goods are transported only by road or rail, so that
we can assume that goods imported by sbiplbarge are to be distributed on both transport
medium. We have to consider two transportation scheme:
• Sea vessel - Sbiplbarge on the Rhine - RaiVroad distribution
• Shiplbarge on inland waterwaysllake - RaiVroad distribution
The proportion of the ship freight to be further allocated to the rail transport must be
guessed (e.g., 40-60%) and added to the amount mentioned for rail (in the second
column). The results shows up in the third to fith columns:
TABLE 7.11. Distribution of the transported tonnage between different transport medium. (Extra-
polation from German statistics to Swiss conditions, assuming a rail share of 40, SO, and 60%)
Transport Trame Swiss Share (%) Swiss Share (%) Swiss Shan' (%)
Medium Share Rail =40% or Rail = 50% or Rail = 60% cf
(%) Ship freight Ship freight Ship freight
Rail 21.3 43.9 49.5 55.1

Road 22.3 56.1 50.5 44.9
Ship (Rhine 29.3 - - -
Traffic)
Sea Vessels 27.1 - - -
RaillRoad ratio 0.955 0.783 0.980 1.227
TABLE 7.12 shows the traffic volume resulting from different rail share
assumptions (40, 50, and 60%). If we assume a 50% share (i.e., an equal distribution
between rail and road). The corresponding distribution with respect to goods classes is
shown in TABLE 7.13.
There is another possibility of estimating the distribution of goods transported by rail
and road in Switzerland which is based on the known quantities of products belonging
to "Class 3", which were transported to the end consumer (TABLE 7.13). It was
assumed that 60% of the goods are transported by road and 40% by rail. The
extrapolation based on "Class 3" results in too low proportion of dangerous goods
210 CHAPTER 7
transported by rail, which means that there must exist a different percentage distribution
between the dangerous goods.
TABLE 7.12. Distribution of dangerous products between rail and road (extrapolation)
Product Ra il Road
(acc:ording to BAV (estimate)
1990)
Catego~ Class Quantity Share Quantity Share

(tons) (%) (tons) (%)
Petroleum products 3 5.9 59.0 6.1 60.2

Raw Chemicals 2/6/8 1.5 15.0 1.5 15.3
Pharmaceutical & 3/6.1 0.2 2.0 0.2 2.0
Dyes
Fertilizers 4 0.2 2.0 0.2 2.0
Compressed 1 2 0.2 2.0 0.2 2.0
liquefied Gases
Total - 8.0 100.0 8.2 100.0
TABLE 7.13. Distribution of dangerous goods transports between

rail and road (Extrapolation based on "Dangerous Goods Class 3")
Dangerous Goods' Road Rail

Class Transport Transport
(Mt) (Mt)
2 0.2 0.1
3 7.2 4.8
4.1 0.1 0.1
5.1 0.1 0.1
6.1 0.5 0.3
8 0.4 0.3
Various Products 0.2 0.1
Dangerous Goods ofCIass 2 [3.1,3.5]
Chlorine. Within the present study it was not possible to obtain independent
information concerning the production of chlorine in Switzerland. An estimation made
by Green Peace on the basis of insider information indicates the quantities shown in
TABLE 7.14. Transport of chlorine in Switzerland is taking place in liquid form in
pressurized tankers of 22'000 I or 50'000 I content, mobile tank of 1'000 1 or in cylinders
of 50 I content.
Liquid Propane Gas. LPG for sale is actually a mixture of propane-butane (so-called
C-mixture) which is traded under the name "propane". This mixture which contains
97% propane behaves physically like pure propane. The total quantities of LPG
transported in Switzerland can be derived from the statistics of the Swiss Petroleum
Products' Association (EV, "Erdolvereinigung"). LPG is transported in rail-tanker of
105'000 I content and road tankers 35'000 I content, in tank-containers with a content of
20'000 I and in gas bottles of 2 to 60 1. The rail-tankers usually build a continuous
block of single cars transporting the gas directly from the refmery to the customer'
place.
TABLE 7.14. Estimated chlorine production in Switzerland
Location Quantity Usage

(kUyear)
Zurzach 30 - intennediate product; Sales
Monthey 25 - for own use at location
Schweizerhalle 15 - intennediate product for own use; Sales
Attisholz 9 - for own use at location
Vinyl Chloride. Vinyl chloride is used as raw product for the fabrication of PVC-
plastic. It is imported as special liquefied gas with rail-tankers of 70'000 I exclusively,
assembled in blocks, some of which are also transiting through Switzerland.
Other Industrial Gases, Air-mixtures
Air-mixtures. These include oxygen, nitrogen, and argon gases pure or in mixture with
air. These gases are either imported or produced d.irect!y in Switzerland. The companies
trading w~th these gases are usually daughter compames of foreign holdings, which are
also member of INTERGAS, located in Bern. The companies active in Switzerland are
represented inTABLE 7.15.
It was not possible to obtain any information on type and quantities of gases
transported from these companies. Transport is taking place in liquid form by
refrigerated road and rail tankers of 60'000 I, and as compressed gas in cylinders of 50
l. Such gases are inert and do not present any risk during transportation.
Hydrogen. Hydrogen is a by-product of the chlorine production from sodium chloride

electrolysis. The gas is compressed to 200 bar and is transported in rail-tankers of
30'000 I, in cylinders of 50 I, and by road in gas cylinders bundles.
Acetylene. It is produced industrially, at several place in Switzerland, from calcium

carbide. The gas cannot be compressed and must be dissolved under pressure in a
solvent (acetone). Transportation is taking place in cylinder bundles.
212 CHAPTER 7
TABLE 7.15. Suppliers of industrial gases, associated with IN1ERGAS
Mother Companies Abroad Daughter Companies

in Switzerland
LINDE (0) PAN-Gas (CH)
AIR LIQUIDE (F) CARBA AG (CH)
AGA(S) AGA(CH)
Mi:?SSER GRIESHEIM (D) Sauerstoffwerk Lenzburg (CH)
Dangerous Goods of Class 3

Petroleum Products. The largest amount of goods transported belong to class 3
(petroleum products). Official statistics, from which a distribution according to region,
product class and type of transport could be deduced, are not available for Switzerland.
The quantities transported were, in the following, derived from statistics of the year
1991 (TABLE 7.16).
TABLE 7.16. Type of petroleum products imported to Switzerland in 1991
Products Imported Quantity Flash Point UN·No.

to Switzerland Classification
(Year 1991) (1,000 tons) ("C)
LiquidlRefinery gases 178.8 <·20 Class 2

Light petrol L VN 362.2 -30 1203
Super petrol 1'640.8 -30 1203
Unleaded petrol 95 2'215.2 -30 1203
Jet kerosene 1'077.8 47 1202
Diesel 1'135.2 60 1202
Heating oil (extra light) 5'572.1 60 1202
Heating oil (medium) 216.0 >60 1202
Heating oil (heavy) 407.1 >70 1202
or free
Bitume 315.9 free
Petrol cocks 71.2 free
Lubricating oil/grease 74.7 free
Aviation gasoline 4.9 -30 1203
Lamp kerosene 4.6 47 1202
Heavy petrol 4.2 5 1203
Petrol for technical 5'021.0 >70 1202
applications or free
Paraffin/wax 11.8
White Spirit 12.4 39 1202
The distribution of the imported petroleum products between the different transport
medium is shown in TABLE 7.17. The transport capacity of the firms belonging to the
fuel dealers as well as the type of road tankers are indicated in TABLE 7.18. The
market share of members of the Petroleum Products' Association have been estimated
and are shown in TABLE 7.19.
TABLE 7.17. Distribution of imported pettoleum products between the
different transport medium (1991)
Transport Type Quantity Quantity

(1,000 tons) (%)
Ship 3759.5 29.88
Rail 2'136.6 16.98
Road 973.7 7.74
Air l.0 O.ol
Pipeline 5710.4 45.39
Total 12'58l.2 100.00
TABLE 7.18. Transport capacity of the fl11l1S belonging to the Association of Fuel Dealers
Product Number 2-Axle 4-AxIe Trailers P1atfonn Transport

of Firms Trucks Trucks Trailers Capacity
10' t.kmJFlrm
Heating oil 103 113 21 52 140.1
-
Heating oil 114 251 114 194 IS 2'188.51
andPettol
TABLE 7.19. Market share (%) of the members of the Pettoleum Products
Association (Estimate)
Finm Petrol I Diesel Heating Oil EL

Quality extra light
(%) (%)
Shell 25 25
Esso 12 15-20
BP 15 10
Migrol 10 10-15
Tamoil 6 10-15
Agip 4 -
Avia 5 -
Aral 2 -
Mobil 2 -
OK Coop 2 -
Total: 83 70-85
I Only heating oil transports.

214 CHAFfER 7
Comments concerning TABLE 7.19 (see previous page)

The second column of TABLE 7.19 includes petrol and diesel. Distribution of these
products to the customers takes place from the main strategic storage locations
close to the border, from regional intermediate storage locations, and from the
refineries as well. Four petroleum companies are distributing ca. 7'668 let of
petroleum products in Switzerland, according to their own information.
TABLE 7.20. Distribution of the pettoleum products (in 1,000 tons) according to the
different Cantons
Canton Number of Petrol Petrol I Total

Inhabitants Heating 011
(x 1'000) (1,000 tons) (1,000 tons) (1,000 tons)
AG 497.5 282 570 852

AI 13.8 8 16 24
AR 51.8 29 59 89
BE 949.4 539 1,088 1,626
BL 230.8 131 264 395
BS 192.5 109 221 330
FR 208.4 118 239 357
GE 380.0 216 435 651
GL 37.9 22 43 65
GR 180.1 102 206 308
JU 65.8 37 75 113
LV 32Q.6 182 367 549
NE 161.2 91 185 276
NW 32.9 19 38 56
OW 29.3 17 34 50
SG 421.7 239 483 722
SH 71.9 41 82 123
SO 226.3 128 259 388
SZ 110.8 63 127 190
TG 205.0 116 235 351
T1 289.7 164 332 496
UR 34.0 19 39 58
VD 587.5 333 673 1,006
VS 254.9 145 292 437
ZG 85.4 48 98 146
ZH 1,157.0 566 1,325 1,982
Total 6,796.2 3,856 7,785 11,641
TABLE 7.21. Geographic dislribution of petroleum products imports (according to Swiss Petroleum
Association, 1991).
Petroleum Products
(1,000 Ions)
R a I I S hlp Pipeline Road Total
LOCATION 8.0ill Petrol 8.0111 Petrol 8.0111 Petrol 8.0ill Petrol

Diesel Diesel Diesel Diesel
North 423 26 · · · 232 311 992
Basel 789 390 1006 453 · - - 448 3086
MUlhausen 11 - - - · · · - -
Schaffhausen 28 - - - - · · · -
Kreuzlingen ~ - - · · - · - -
Bern 41 51 - - - - - - .
Luzem 24 30 - - - - - - -
A1tdorf 1 3 - - - - - - .
North-West 261 23 . - 2 - 510 242 1038
Val10rbe 47 46 - - - - · - 93
Fribourg 40 20 - - - - - - 60
AigIe 40 84 - - 52 30 - · 206
West! Geneva 55 72 . · 531 109 - 170 937
South 1 - - · - - 54 131 186
Tessin 24 23 - - - · 64 12 123
Poschiavo - . - - - - 2 - 2
Simplon 2 - - - - - - - 2
East 87 20 - - - - 71 ~ 267
ZUrich 157 94 - - - - - - 251
WiVSG 22 40 - - - - - - 62
Sennwald 50 - - - - - - · 50
Erlen 4 11 - - - · · - 15
Landquart 3 18 - - - - - - 21
Bund 5 3 - - - - - - 8
TOTAL 2199 951 1006 453 585 139 933 1414 7659
Note: The rubrique H.OiI (Heatmg Oil) and DIesel also UlCludes kerosene.
There are actually four main petroleum and products pipelines in operation in
Switzerland (see also Chapter 6). These are:
216 CHAPIER 7
• Pipeline "Oleoduc du Rhone" carrying crude-oil from Genua to the refinery of

CoUombey (RSO, "Raffinerie du Sud-Ouest" in the RhOne valley).
• Pipeline "Oleodotto del Reno", bifurcation of Sennwald, carrying petroleum
products.
• Pipeline "SAPRO" from Marseille to Geneva, carrying petroleum products.
• Pipeline "Oleoduc du Jura", bifurcation of the SEPL-pipeline, carrying crude-oil
to the refinery of Cressier (near Neuchatel).
A fifth pipeline is carrying heavy fuel oil from Collombey to the thermal power
station ofVouvry (in the RhOne valley). .
An estimation of the distribution and consumption of petroleum products by region
is shown in TABLE 7.20 and TABLE 7.21.
It seems to be a slight discrepancy between the data supplied by the above men-
tioned unnamed companies and the total quantities of imported petroleum products
(petrol, gasoil, and heating oil) announced by the Swiss Petroleum Association. A
comparison of the imported quantities with the inland consumption is shown TABLE
7.22.
TABLE 7.22. Comparison between imported quantities and domestic sales
Product Petrol Heating Sub-total Crude Oil Total

OiIlGasoil
(1,000 tons) (1,000 tons) (1,000 tons) (1,000 tons) (1,000 tons)
Imported 2'810 4'707 7'517 4'554 12'071
Sales 3'856 7'807 1l'663 - 11'663
Source: StatIStIC from the SWISS P~troleum Assoelatwn
Product of Class 6.1, 8 and 9

Toxic Products. For these products, there was no statistic available besides scarce
information concerning rail transportation. One very toxic representative compound of
this class 6.1 is tetraethyl-lead, a petrol additive, which was transported earlier in large
quantities. Actually the most dangerous products to be found here are pesticides and
herbicides, which are used in large quantities for a agricultural purposes. No data was
available regarding the transported quantities. Only small quantities are sold on the
domestic market and the rest of the production is exported to third countries.
7.2.4. STATISTICS CONCERNING PETROLEUM PRODUCTS
The statistics presented in this section were prepared by the Swiss Petroleum
Association, EV ("Erdolvereinigung"). The original report covers, however, a larger
period than this is the case for the data presented here.
Swiss Market. The. sales of petroleum products in Switzerland (TABLE 7.23) totalized
12,732 million tons for the year 1994, of which motor fuel 47.4%, heating oil 45.6%,
and other products 7.0% (including refmery own consumption).
TABLE 7.23. Sales of petroelum products in Switzerland (1994)
Petroleum Product Swiss Market Sales

Year 1994 (tons x 10')
Petrol 1 29.1 %
Diesel ~ 6,033 8.8%
Jet kerosine J 9.5%
Heating oil 5,815 45.7%
Other Products 0,883 5.2%
Refinery own consumption 1.7 %
Total Sales 12,732 100.00%
Swiss Statistics Concerning /mportationlExportation of Petroleum Products

The type and quantities of petroleum products imported to and exported from Switzer-
land during the period 1992-1993 are mentioned in TABLE 7.24 and TABLE 7.25.
TABLE 7.24. Swiss import slalistics for petroleum products
Product 1 Year 1991 Year 1993 Year 1994

(toIlS) (toIlS) (toIlS)
Automotive gasoline (leaded) 1,063,449 624,776 487,015

Automotive gasoline (unleaded) 1,917,833 1,967,342 2,084,009
Light gasoline 44,308 27,506 33,244
Jet kerosine I motor kerosine 65,713 47,678 30,411
Diesel oil 661,942 624,953 645,391
Domestic heating oil 4,129,556 3,062,629 3,480,395
Bitumes 153,596 143,538 139,677
Petrol coke 49,122 63,550 69,927
Lubricating oils I Grease 71,141 67,(1)6 70,836
LPG 33,573 27,022 31,807
Other products (incl. MTBE I 248,964 316,742 168,826
spikes)
Total finished products 8,439,197 6,973,342 7,241,538

Other products, as defined under 1 4,124,040 4,499,304 4,766,577
Sum total 12,563,237 11,472,646 12,008,115
Source: Bundesamt fUr Energlewirtschaft
~: I Without the products not liable to custom duty, such as Samnau, TEE, DiplOllUllic and
international organizations, and without the jet kerosine for the international air traffic (not liable to
duty), and without the importations which do not neccessitate dispatching way-bills.
218 CHAYI'ER7
TABLE 7.25. Swiss export statistics for petroleum products
Product Year 1992 Year 1993 Year 1994

(toDS) (toDS) (toDS)
Automotive gasoline 55 1,125 36
Diesel oil - - 385
Jet kerosine 9 11 2,487
Heating oil 370,238 573,995 551,551
Bitume 1,363 330 296
Lubricating oil I Grease 11,205 7.527 12,269
LPG 30,792 37,850 47,221
Other products 1,635 1.550 1,592
Total 415,297 622,388 615,837
Transportation and Distribution of Petroleum Products in Switzerland

Importation of Petroleum Products. For the year 1994, the total crude oil and petro-
leum importations totalized 12,875 million tons (including 0.867 million tons of
imported products not liable to custom duty). The transport market share between rail,
road, ship, and pipelines is shown in TABLE 7.26.
TABLE 7.26. Importation of petroleum products (distributed by transport medium)'
Imports by Year 1991 Year 1992 Year 1993 Year 1994

Trsp.
tonsJ: 10' (%) tons J: 10' (%) tons J: 10' (%) tonsJ: 10' (%)
medium
Ship 3,759,526 29.9 4,113,110 32.7 3,883,039 33.9 3,759 31.3

Railroad 2,136,624 17.0 2,161,047 17.2 1,231,185 10.7 1,749,586 14.6
Road 973,664 7.7 977,813 7.8 846,702 7.4 852,733 7.1
Aircargo 979 0.0 1,042 0.0 48 0.0 28 0.0
Pipelines 5,710,412 45.4 5,310,225 42.3 5,511,672 48.0 5,646,543 47.0
Total 12,581,205 100.0 12,563,237 100.0 11,472,646 100.0 12,008,115 100.0
~: 'Without the products not liable to custom duty, such as Samnau, TEE, Diplomatic and
international organizations, and without the jet kerosine for the international air traffic (not liable to duty),
and without the importations which do not neccessitale dispatching way-bills.
Petroleum Products in Transit by Pipelines. The pipeline connecting the area of Genua
!Marseille to Switzerland transported a total of 5,890 million tons crude oil, spikes
(semi-products), and finished products into Switzerland (including finished products
not liable to custom duty).
The pipeline connecting Central Europe to Italy transited 7,450 million tons crude
oil and finished products through the Canton Graubilnden. TABLE 7.27 shows the
tonnage of petroleum products transported by pipelines in Switzerland, including toll-
free products, aviation gasoline, and jet kerosine.
During the year 1994 there was no damage or incident to the pipeline network
covering the Swiss territory.
TABLE 7.27. Quantities ofPeIroleum products transported by pipelines (year 1994)
Pipeline Petroleum product Quantity Total Average

Denomination (tons) (tons) Flow Rate
(tonslhr)
Oleodotto del Reno 7,717,316 1,052

(Transit pipeline) Crude oil 6,725,295
Domestic heating oil 333,765
Virgin naphta 392,153
Sennwald Fuel oil I diesel oil 266,103
OIeoduc du RhOne 1,797,561 327

Crude oil 1,740,811
Fuel oil (untreated) 51,550
Virgin naphta 5,200
Oleoduc du Jura Crude oil 3,008,799 3,008,799 488

Neucbitelo1s
Sappro 817,551 292

Automotive gasoline 168,645
Jet kerosine 251,446
Fuel oil I diesel oil 397,460
Vouvry 37,687 37,687 33
Ship Transports on the Rhine River. The petroleum products imported to Switzerland
by ship transport on the Rhine river (TABLE 7.28), amounted to 4,161 million tons for
the year 1994 (inclunding products not liable to custom duty).
TABLE 7.28. Petroleum products imported to Switzerland by ship
Product Imported by Ship Year 1992 Year 1993 Year 1994

(on the Rhine) (tons) (tons) (tons)
Automotive gasoline 1,562,257 1,708,355 1,602,978
Other motor fuels 107,030 121,957 130,674
Heating oil (light) 2,704,441 2,223,822 2,389,370
Heating oil (mediumJheavy) 8,080 8,781 5,418
Lubricating oil/Grease 8,027 9,161 9,672
Petroleum coke and other products 7,837 10,316 23,319
Total 4,397,672 4,082,392 4,161,431
220 CHAYfER7
TABLE 7.29. Inland sales of petroleum products by categories
Petroleum Product Year 1992 Year 1993 Year 1994

(tons) (tons) (tons)
Automotive gasoline 3,994,318 3,704,768 3,703,074

Jet kerosine 1,137,114 1,176,931 1,207,776
I
Diesel oil 1,099,659 1,060,367 1,122,686
Domestinc heating oil 6,073,315 5,532,892 5,815,058
LPG 174,274 180,479 172,159
2
Light gasoline 45,005 27,506 32,244
Bitume 285,681 265,913 278,982
petrol coke 48,168 62,639 68,935
Lubricating oilslgreace 70,971 67,819 62,993
Aircraft gasoline 4,766 4,227 4,331
Lamp kerosine 3,366 1.493 1,989
Special boiling point spirit 3,676 3,591 3,619
Naphta (for technical means) 5,164 5,368 8,677
ParaffmIWax 11,171 11,444 10,560
White spirit 13,265 11,796 12,551
Other products 665,507 642,275 658,040
Refmeries own consumption 171,410 195,401 225,389
Sum total 13,141,323 12,312,634 12,732,023
..
Source: BWldesar.tt filr Stattsttk, Bern
~: I incIudinll 18,062 tons of diesel oil stored in the Rhine ports of the Basel area
2 of which 4,031 tons were delivered to gasworks.
TABLE 7.30. Structure of the swiss petroleum products sales
Product Year 1983 Year 1991 Year 1992 Year 1993 Year 1994
% % % % %
Automotive gasoline 24.6 29.7 30.4 30.1 29.1

Jet kerosine 6.3 8.3 8.6 9.6 9.5
Diesel oil 6.4 8.8 8.4 8.6 8.8
Domestic heating oil 54.8 46.3 46.2 44.9 45.7
Other products 6.4 5.5 5.1 5.2 5.2
Refineries own consumption 1.5 1.4 1.3 1.6 1.7
Total 100.0 100.0 100.0 100.0 100.0

..
Source: BWldesamt fiIr Statisttk, Bern
With a share of 56.1% (54.7% in 1993) of the total goods transported by ship, the
petroleum products imports represent quantitatively the largest amount of the products
arriving at the ports of both Basel (located on Swiss and German territory as well). The
export by ship of petroleum products (mainly heavy heating oil) represented 48.6% of
the ship transports in the Basel area. The distribution of petroleum product sales by
categories is given in TABLE 7.30.
Railroad Transports. Inland railroad transports of petroleum products from the ports
of Basel amounted to 28% (1,165.08 million tons) of the total quantities imported by
ship for the year 1994 (against 34% in 1993).
Importation of Petroleum Products by Road. They are relatively unimportant and

amounted to 0,853 million tons, mainly imported from Italy.
Service Stations. At the end of the year 1994 the number of service stations delivering
automotive petrol and diesel oil amounted in Switzerland to 3841 fuel stations. The
sales of the self-service fuel stations represented 89.8% of the total sales of petrol and
diesel oil, on average, 1,295 million liters per station. The fuel stations along the
national highways sold, on average, 5,560 million liters per station.
Inland Sales of Petroleum Products

The inland sales of petroleum products by categories is shown in TABLE 7.29 for the
period 1992-1994.
7.3. Traffic Accidents Statistics Involving Dangerous Goods
7.3.1. TRAFFIC ACCIDENTS STATISTICS (SWITZERLAND)
In Switzerland the reporting of accidents by the local police or the highway police
should definitively be done according to an accident protocol, which still remains to be
dermed and agreed on. Nowadays both statistical data concerning volume and type of
dangerous goods in transit and the number, type and consequences of transport accident
are not easily available. The statistics do not allow conclusions on accident rate per
dangerous goods category to be drawn. It is also not possible to establish a flow chart of
the chemicals and other dangerous goods transported within or through Switzerland
There is a necessity for implementing a data base concerned with dangerous goods
transport incidents, as this is done in the U.s. or in other countries. The actual
legislation does not oblige the transport firms and trucking companies to announce
dangerous transports to the authorities nor are they at present obliged to supply
information on goods movements and volume transported
The reporting of accidents should also attempt to break the causes of an accident
down into two main categories, human error or mechanical failure. Among the types of
accidents we find rollover, leak, spills, fire, (explosion), equipment failure. Rollover
and spills are usually quite frequent. Liquids transported in bulk require special driver
skills because of the high center of gravity and liquid movements which can easily
cause rollover.
222 CHAPI'ER 7
The Public Safety Service for the control of dangerous goods transports (Cantonal
Administration! Cantonal Police) should ensure that:
• Material safety data sheets for all products are carried in the cabin of the trucks.
• Detailed maps of all consignee loading and unloading areas and emergency
equipment location are Available at the Public Safety Service.
The Public Safety Service should also assist in the development of the trucking
companies' policy and procedure manual. Vital elements in preventing motor vehicle
accidents are the driver's skill and knowledge, both of which can be improved through
appropriate training. Driver attitude is affected by the company's philosophy. In view of
increasing concurrence in the transportation business, the pressure on the drivers from
the side of the companies is increasing. As a result violation of speed limits, rest period,
etc. are a frequent cause of accident due to fatigue, speed excess or inattention.
7.3.2. TRAFFIC ACCIDENTS STATISTICS (WEST GERMANY)
In view of the lack of availability of adequate Swiss statistics concerning transport

accidents, which could be used for risk assessment studies, we have summarized below
information from German statistics, which show some general trends. One difficulty in
extrapolating statistics from one country to another arises from the different road
construction types, speed limitations, driving habits, and regulations concerning driver's
rest time, etc., which may show large variations between the different countries and
even within the european countries.
The Ministry of Traffic ("Bundesanstalt fUr Strassenwesen") in West Germany, is
regularly publishing the results of their investigations concerning traffic accidents
involving the tmnsportation of dangerous goods (TABLE 7.32). One of their latest
publication covers the period 1982 - 1991. The total dangerous goods transported by
rail, road and ship amounted to 359 Mio tons for the year 1990 (of which 74% were
transported by road, 12% by rail and 14% by ship).2 The transport yield was 17 Mrd
ton-km for road transport, 8.7 for rail, and 37.6 for ship transports. A comparison of
accident rates shows that the absolute accident rates are the lowest for rail transport.
TABLE 7.31. Road tanker accidents in Western Germany, for the period 1989-1991
Accidents with Release of Toxic

Products to tbeEnvironment
OfIlclal Survey Prepared by: Year 1989 Year 1990 Year 1991
"Bundesanstalt fiir Strassenwesen, (0)" 49 28 37
Accidents involving transport of goods 60 69 66
toxic to water (including loading I
unloading)
OfflCial statistics of road accidents - - 34

rtiv - THESEUS - 22 21
TABIE 7.32. Transport of dangerous goods (TOO) in Western Germany. and related
accident statistics3
Designation Ran Road Ship (Inland) Total
TDG-Distribution 12.0% 74.0% 14.0% 100%

share (1990) (359 Mio. t)
TDG -Transport 8.7 17.0 11.9 37.6

yield (Mrd t-km)
TDG toxicto
water
(average for 1982-
1991)
Number of 24.0 306.0 82.0 412.0
Accidents
Volume released 137.0 411.0 386.0 934.0
to environment
(m3)
Risk Index
Accidental Risk 2.8 18.0 6.9 -
(A.R./Mrd t-km)
Index-value: 1 6.5 2.5 -
Rail =I
Damage Risk 15.8 24.2 32.5 -
(m3/Mrd ton-km)
Index-value: 1 1.5 2.1 -
Rail =I
The share of road-trucks accidents is 74%. Regarding the accidents of road tankers
and ships, 85% of the total load was released to the environment. The volume released
to environment per accident was the lowest for road transport (1.3 m3). The quantities
involved in ship and rail accidents were much larger (4.7 m3 per ship accident and 5.7
m3 per rail accident). A comparison of accidents, respectively volume released to
environment based on the transport yield in Mrd ton-kID, one obtains two risk indices,
showing that rail transportation is more secure than the other two.
The transport of dangerous goods by road (TABLE 7.31) resulted in 113 accidents
involving only road tankers for the period 1975-1976,4 and 248 dangerous goods
transports for the period 1982-1984. 5 Reports differentiating the accident according to
parcel transport and road tanker transport were also reported for different periods:
• Period 1987-1989 (reported in 1990): Accidents involving parcel transporta-
tion6
• Period 1990 (Reported in 1991): Accidents involving parcel transportation7
• Period 1988-1990 (reported in 1992): Accidents involving water toxic pro-
ducts. 8
• Period 1989-1991 (reported in 1993): Accidents involving road tankers9.
224 CHAPTER 7
The differences between the results of the different surveys under consideration are
due to different definitions of the accidents, transports, and product involved. It was
also established that not every accident is recorded in the same way, nor that every
accident is recorded at all by the different statistical services in charge of producing
official surveys. For the period 1989-1991, there were 131 accidents recorded
(involving product release or damaged valves/piping). The accident frequency has
dropped as compared to the year 1980, but the number of participants involved in an
accident bas remained nearly constant
TABLE 7.33 shows the distribution of accidents in West-Germany according to the
location where the accident occured (i.e., inside a locality, on country roads, on high-
ways).
TABlE 7.33. Accident distribution in West..<Jermany for the period 1989·1991
Place of Accident Locality Country Highways

roads
Accident Distribution 12% 43% 45%
Some considerations were also given to the accident circumstances and these are
summarized below:
Day and Time. 120 of 131 accidents occurred from Monday to Friday and 101 of 131
happened at day-time.
Type of Accidents. 60 of 131 accidents were due to driver's own fault (of which 53%
were &elf-accidents), 33 of 131 were due to side collisions or due to front-to-rear
collisio.ls (ca. 50%).
Kind of Accident. It includes deviation of the vehicle to the left side or right side of
the roadllane (70 of 131 accidents). The next most common cause was driving onto the
next vehicle (front-end collision type). Usually this kind of accident occurred on
Highway, and was occasioned by a foreign vehicle colliding into the rear of a road-
tanker.
Responsibility. The driver responsible for the accident was, in most cases, the tanker-
driver (70%). This is due to the high rate of self-accidents (53%). In the case of
accidents involving two or more vehicles the tanker-driver was only found responsible
in one third of the cases.
Cause of Accident. A major cause of accidents was inadequate speed. 27 of 65 tankers

were damaged on the side. This resulted in tum-over of the tanker (on side or roof) in
55% of the cases. The trailer was also quite often involved (14 of 20). The average age
of the drivers involved in accidents was 38 years.
smaller on country roads, and the least severe in localities. In 74 accidents involving
dangerous goods, there were altogether 166 people injured or killed (table below):
TABLE 7.34. Damage to people resulting from accidents involving the transportation
of dangerous goods
Number of People Killed I Injured

Accident Killed Heavily Slightly Combined
Location Injured Injured
Highways 10 17 44 71
Country 3 33+ 27 63
Roads
Localities 0 3 29++ 32
TOTAL 13 53 100 166
Note: 'II' 19 persons were heavily injured after a collision with an omnibus;
'II' 'II' 23 persons were slightly injured after a collision with a tram.
The dangerous goods, as such, were the direct cause of two people's fatality and of
one person being injured (TABLE 7.34).
Regarding the distribution of the transported goods according to the RID danger
classes (TABLE 7.35), 103 (82%) of the 126 trucks had an accident involving danger-
ous goods of class 3 (flammable products).
TABLE 7.35. Accidents classified by location and dangerous goods class (RID)
Dangerous Goods Class (RID)

Accident Class Class Class Class Class Class
Location 2 3 4 5 6 8
Highways 3 39 3 1 2 9
Country Roads 2 50 1 1
Localities 14
TOTAL 5 103 3 1 3 11
Concerning the number of accidents involving "Danger Class 3", 38% of them
happened on highways, 48% on country roads and only 14% in localities. A release of
the tanker content was involved in 118 of 153 accidents. It was reported that the tank-
wall was damaged in 65 cases. On average the accidents involved 16'000 1 or 3'SOO kg
of dangerous goods. The quantity released to the environment (on average 4'800 liters
of which 1'000 1 could be recuperated) is larger for accidents happening on highways
than in localities or on country roads. For the period 1989-1991, there were six severe
accidents involving more than 30'000 liters lost to the environment. The costs of such
accident have been estimated and reported by "Bundesanstalt fUr Strassenwesen,
Wiesbaden".10
226 CHAPTER 7
accident have been estimated and reported by "Bundesanstalt fUr Strassenwesen,

Wiesbaden" .10
It must be noted that, in West Germany, the part representing accidents involving
the transportation of dangerous goods is extremely small as compared to the much
larger number of other traffic accidents.
7.3.3. OTHER ACCIDENT STATISTICS INVOLVING DANGEROUS GOODS

TRANSPORTS
France
French statistics on trucks accident have been reported by Degranges ll (1989). Data on
risk aSsessment for anhydrous ammonia have been compiled by Castellano et al.,12
(1989), and propane transportation by Hubert, et al.,13 (1990). study for the region of
Lyon has been presented by Hubert, et al.,14 (1988). The French Ministry of Transports
is publishing yearly accident statistics. is Statistics for the French Railways, SNCF, are
also available from the Ministry of Transports in Paris.
Great-Britain
Transport accident frequency data have been reported by Appleton,16 (1988). The
Health and Safety Commission is also a source of interesting publications on
transportation. 17
USA
The following reports are available on: Severity of transportation accidents involving
large packages in the U.S. (Dennis et al.,18 1978), and risk assessment for transporting
propane by truck and rail (Geffen, et al.,19 1980).
Holland
A study regarding LPG transport in the Netherlands has been published by TN020
(1983).
TABLE 7.36. Accident frequencies
Country Accident Frequencies: Occurrence

(Reference, Year) per vehicle aad per kilometer
Rail Accidents Road Accidents
France (Hubert, 1990) 1.0 x 10-7 5.0 x 10-7
U.K. (Appleton, 1988) 0.7 x 10-7 6_7 X 10-7
USA (Geffen, 1980) 9.4 x 10-7 16.0 x 10-7

The Netherlands (1NO, 1.4 x 10-7 4.4 X 10-7
1983)
TABLE 7.37. Release frequencies for road
Country Total Releases That Can Lead to Consequences

Release
Major Release Medium Release Minor ReIease
0.2 m2 or 100% 0.01- 0.001 m 2 < 0.01- 0.001 m2
of content
France 130 x 10-9 33 X 10-9 97 X 10-9

(Hubert, 1990)
U.K.. 0.5 x 10-9 0.048 X 10-9 0.43 X 10-9

(Appleton, 1988)
USA 180 x 10-9 43 X 10-9 137 X 10-9

(Geffen, 1980)
The Netherlands 22 x 10-9 0.66 x 10-9 1.5 x 10-9 20x 10-9

(TNO,1983)
TABLE 7.38. Release frequencies for rail
Country Total Releases That Can Lead to Consequences

Release
Major Release Medium Release Minor Release
0.2 m2 or 100% 0.Ql- 0.001 m2
of content
France not 1.1 x 10-9 not available

(Huiler!, 1990) available
U ..k. 2.5 x 10-9 0.25 X 10-9 2.25 X 10-9 =10 X 10-\0

(Appleton, 1988)
USA (Geffen, not 42 x 10-9 not available

1980) available
The Netherlands 14 x 10-9 0.4 X 10-9 0.98 X 10-9 13 X 10-9

(TNO,1983)
Inumatiolllll Codes and Guidelines

The codes, guidelines and regulations cited below are international codes of general
interest concerning safety aspects, calculation methods etc., which include interesting
information_ In Switzerland there are other codes and regulations in effect, which may
differ from the above mentioned codes_
Most regulatory action in the area of dangerous goods transportation is aimed at

improving the mechanical aspects of spill prevention, i.e., improving the standards for
vehicles, tanks, containers and packaging. Another concern is the design of better road
228 CHAPTER 7
vehicles, tanks, containers and packaging. Another concern is the design of better road
and ancillary equipment, such as oil catchment canalization, barriers, improved
signaling and speed control system, etc.
In Switzerland the reporting of accidents by the local police or the highway police
need to be done according to an accident protocol, which still needs to be defined and
agreed on. Nowadays both statistical data concerning volume and type of dangerous
goods in transit and the number, type and consequences of transport accident are not
easily available. The statistics do not allow conclusions on accident rate per dangerous
goods category to be drawn. It is also not possible to establish a flow chart of the
chemicals and other dangerous goods transported within or through Switzerland. There
is a necessity for implementing a data base concerned with dangerous goods transport
incideQts, as this is done in the U.S., or in other countries. The actual legislation does
not oblige the transport firms and trucking companies to announce dangerous transports
to the authorities nor are they at present obliged to supply information on goods
movements and volume.
. The reporting of accidents should also attempt to break the causes of an accident
down into two main categories, human error or mechanical failure. Among the types of
accidents we find rollover, leak, spills, fire, (explosion), equipment failure. Rollover
and spills are usually quite frequent Liquids transported in bulk require special driver
skills because of the high center of gravity and liquid movements which can easily
cause rollover.
Public Safety Service for the control of dangerous goods transports (Cantonal
Administration / Cantonal Police) should ensure that:
• material safety data sheets for all products are carried in the cabin of the trucks.
• detailed maps of all consignee loading and unloading areas and emergency
equipment location are available at the Public safety Service.
The Public Safety Service should also assist in the development of the trucking
company's policy and procedure manual. Vital elements in preventing motor vehicle
accidents are the driver's skill and knowledge, both of which can be improved through
appropriate training. Driver attitude is affected by the company's philosophy. In view of
increasing concurrence in the transportation business, the pressure on the drivers from
the side of the companies is increasing. As a result violation of speed limits, rest period,
etc., are a frequent cause of accident due to fatigue, speed excess or inattention.
7.4. RaiI/Road Transport Containers
7.4.1. INFORMATION CONCERNING SWITZERLAND
The equipment used for inland road transportation of dangerous goods may vary and
the carriers are relatively free to use a wide range of containers and vehicles, given such
restrictions as physical state of the product, and need for pressurization or temperature
dangerous goods, can only be shipped across the border in compliance with the rules
stated by such regulations and ordinances. Moreover the maximum allowable quantities
and packaging type, which are prescribed, must be adhered to.
The prescriptions concerning construction, testing and examination, permit for use
for dangerous goods' transportation are defined in the following regulations:
• The UNO recommendations are aiming at harmonizing the national regulations in
order to assure the same quality standards for all tanks and containers used for
transportation.
• Both the IMDG code and the new edition of the U.S. Code of Federal Regulations
have taken over the recommendations of the UNO concerning ship transport. The
ADRIRID and SDRIRSD regulations, as well as other European regUlations, however
are based mainly on the 100 years older RID code. Differences between the inter-
national and national codes are to be found regarding the wall thickness of containers
and tanks. The problematic has been explained in an EMPA-report1 (1993).
What concerns the safety of tanks in case of accidents, there has been several test
series conducted by specialized institutions like for instance the EMPA (Swiss Federal
Laboratories for Materials Testing and Research). Three kinds of experiments
concerning the resistance of tanks against piercing have been published:
• Experiments with plates or shells on testing machines
• Experiments with tanks on a full scale
• Experiments with models
The results of the experiments have shown that real accidents are difficult to
reproduce in an experimental set up. Tht: following conclusions can be drawn:
- road tankers with aluminum container, may burst as a result of a chock, when the
tank is full. The resistance to perforation is smaller than for steel containers.
regarding rail tankers, the nonnal containers have a wall thickness of 6 mm, and
the containers for transporting liquefied gases have a wall thickness of 12 to 15
mm. In the latter case a perforation of the container wall is very improbable.
- all the tanks are designed to withstand the internal vapor pressure at 50°C.
Infonnation concerning type and number of rail and road tankers used in Switzer-
land (TABLE 7.39) can be obtained from the following sources: .
• Federal Office for Statistics (for road tankers).
• Data base of EGI ("Eidg. GeJahrgutinspektorat"), that is the Swiss Association
in charge of Controlling Pressure Vessels (SVDB, "Schweizerischer Verein for
Druckbehiilteriiberwachung").
• Data base of ZfW ("Zugforderungs- und Werkstiittendienst SBB"), that is the
repair workshops services of SBB. This data base maintains records concerning
the type and number of rail tankers in service.
230 CHAPTER 7
• Data base of ztw ("ZugfOrderungs- und Werkstattendienst SBB"), that is the

repair workshops services of SBB. This data base maintains records concerning
the type and number of rail tankers in service.
TABLE 7.39. Number and type of rail and road tankers (without gas tankers) in Switzerland
Content Numberol Numberol Numberol Numberol Numberol

in liters Road Tank Platforms Rail Tank
Tankers Trailers Tankers Containers
3'000 127 387 1 3 541
9'000 565 419 11 4 199
12'000 1,545 831 14 3 64
14'000 140 556 32 2 32
15'000 9 16 14 5 4
24'000 435 16 214 297 132
25'000 61 0 30 11 60
40'000 10 0 189 1,966 191
60'000 0 0 0 135 I
85'000 0 0 I 2,422 0
Total: 2,892 2,225 S06 4,948 1,224
The next set of tables indicates the number of rail tankers with a given wall
thickness (TABLE 7.40), and the utilization of rail tankers for liquefied gases for the
year 1988 (TABLE 7.41). The typical dimensions ofraillroad tankers and containers are
given in TABLE 7.42.
TABLE 7.40. Wall thickness of rail tankers used for petroleum produ·::ts
Number of Rail Tankers

in Service
Rail Tankers Type Wall Wall
Thickness Thickness
>5mm S5mm
2 Axles 840 14
4 Axles 0.6 bar test 5S5 SOl
pressure, manhole 650 0
4 Axles 1.5 bar test 142 272
pressure, manhole 500 0
Total 1,567 I,OS7
TABLE 7.41. Utilization of the rail tankers for li-

quefied gases in Switzerland (Year 1988)
Type of Gas Number of

Rail Tankers
Ammonia 4
C-Mixture (Propane) 80
C(h 6
Ethylenoxide 2
Chlorine 32
Sulfur dioxide 6
(h,N2, TK 13
Bromine 3
Phosphorus trichloride 5
Methylene chloride 3
Vinyl-chloride 6
Butadien 2
Dimethyl-sulfate 2
Di-chloride-fluoro- I
methane
TABLE 7.42. Typical dimensions of tankers
Size Tank Rail Road

Container Tanker Tanker
Tank: D:ameter (m) 2.2 2.6 2.2 x 1.8
Width (m) 2.4 3.0 2.3
Height (m) 2.6 4.5 2.8
Length (m) 6.0 12.0 9.0
Weight, empty (ton) 3.4 21.0 15.0
Test Pressure (bar) 4.0 0.6 0.4
Material 1.4571 FE255 Pe460
(DIN 7441) (Aluminum)
Wall thickness (mm) 3.0 6.0 5.0
7.4.2. RELIABILITY OF TANK/COMPONENTS FAILURE
At the present time there are no reliable tanks and components failure probability data
available for Switzerland. Based on data published in the open literature, probability of
material releases were derived for Swiss conditions (Kieselbach 1, 1993). Such data
(TABLE 7.43) must be used with caution since the conditions prevailing in the
countries of origin are quite different from the Swiss conditions:
232 CHAPTER?
• The road conditions are different due to the presence of mountains and hills. The
maintenance is usually better, but the road are subjected to variable climatic
conditions like fog, snow, and ice in winter.
• The driving habits and regulations are not the same: restrictions regarding
driving and rest periods, no truck driving at night, lower average speed, lower
transported weight, etc.
If we assume, for Switzerland, 5'000 road tankers carrying petrol and 100'000 km
driven per tanker and per year, we obtain 5 x 10.8 truck-km per year. This means that
we should reckon with at least one accident (of any type, except rear-end collision).
But, indeed reality looks far better.
TABLE 7.43. Probability of release per road tanker-krn
Type or Side Effect Damage Probability or Release per Road Tanker-kin

Accident Type
Propane Petrol Chlorine
Getting off- Turn-over Impact on

road (side) shell 0 7 x 10.8 7 x 10-8
Skidding 0 0 0
Side collision Side impact 0.5 x 0.8 m2
on cylindrical surface 1.6 x 10-8 1.6 x 10-7 1.6 x 10-7
shell involved
Rear-end Truck HaIfrear-
collision colliding onto end shell 10-9 5 x 10.9 5 x 10-9
other truck
I
Truck being HaIr rear- 10-9 5 x 10-8 5 x 10-8
hit end shell
Frontal Half front- 2.5 x 10-7 2.5 x 10-7 2.5 x 10-7

collision end shell
TABLE 7.44. Probability of material release per road tanker-kilometer
Probability of Material Release per Road-tanker Kilometer
Incident Type Petrol Propane Chlorine
Perforation with - 4.8 x 10- 10 0.8 X 10- 10

leakage> 1'500 kg
Equipment leaking - - 1.4 X 10-8
Leakage over 1'500 kg 2.1 x 10-8 - -

The data of TABLE ?44 are taken from a British study. 22 When the probability of
failure of tank components, like pumps, hose, valves, and relief valve are known, it is
in principle possible to calculate the probability of material release from transport
tanks on the basis of event trees. However it must be stated that there are not many data
from foreign studies that may be transposed to Swiss conditions and particularities.
The principles underlying the design and operation of tanks and vessels for the
containment of hazardous substances are broadly similar for transport by road, rail
and sea. Relevant codes are: the UN Transport Code, the IMCO Code ('Intergovern-
mental Maritime Consultative Organization), RID, ADR and the (British) Home Office
Containers Report.
7.4.3. UNLOADING TERMINALS
Reloading points for road and rail tankers are areas in which a large number of
accidents occur. There are three broad categories of hazard for a tanker terminal:
• Fire/explosion of product in the tanker
• Fire and explosion from leaks and spillage
• Accident arising from transfer of the wrong products.
Terminals are discussed in Truck Loading Rack Safety by the alA (Oil Insurance
Association, USA; Item 8), in API Bull. 1003 (1975, Item 29) and by Hughes?3
Relevant codes are NFC 30 Flammable and Combustible Liquids Code (1973); and
NFC 58 Liquefied Petroleum Gases, Storage and Handling (1974); and the ICI LFG
Code.
Summary
Risk assessment for regional planing comprises information and statistics of the volume
and class of dangerous goods transported or stored in large quantities in a given region.
Specialized databases and information procedures were lately developed in view of
assisting the risk analysis calculation as well as comparative assessment of various
transportation systems or technologies involved in the given technical process. The
present chapter introduces to the reader information, sources of statistical data as well as
the type and quantity of goods transported in Switzerland. Importation / exportation
statistics concerning petroleum products are presented in detail. Traffic accident
statistics involving hazardous materials are presented for Switzerland. Similar data are
given for Germany. Statistical information on rail/road transport containers and the
reliability of tank/components failure are also included in this chapter.
234 CHAP'IER 7
1 Kieselbach. R .• Transport geflihrlicher GUter in der Schweiz. Basisinfonnationen I StandortbestinuDlDlgen

1992. EMPA-Bericht Nr. 227. (Swiss Federal Laboratories for Materials Testing and Research). EMPA,
CH-8600 Diibendorf. (1993).
2 Nicodemus. S.; Eberlein. H .• Schlitzung von Umfang und Struktur des Transportaufkommens gefilhrlicher
Giiter. Statistisches Bundesamt. Wiesbaden. (Januar 1993).
3 Bundesminister fiIr Verkehr: Verkehr in Zahlen 1992. Deutsches Institut fiIr Wirtschaftsforschung (DIW).
Berlin. (1992).
• Elslisser.l.. Strassenverkehrsunflille von Fahrzeugen mit gefilhrlichen GUtem. Auswertung von Unfallakten.
Bundesanstalt fiIr Strassenwesen. (1978). (unverOjfentlicht).
5 Bressin. M .• UnflUle heim Transport gefilhrlicher Giiter auf der Strasse 1982-1984. Forschungsberichte der
Bundesanstalt fiIr Strassenwesen. Bereich Unfallforschung, Nr. 126 (1985).
6 Bamberg, R., Unflille beim Transport gefilhrlicher Giiter in Stiickgutbehiiltem Zwischenbericht fiIr den
Zeitraum 1987-1989. Bundesanstalt fiIr Strassenwesen, (1990). (unveriJjfentlicht).
7 Gail, H.E., Unflille beim Transport gefthrlicher GUter in Stiickgutbehiiltem. Ergebnisse fiIr das Jahr 1990.
Bundesanstalt filr Strassenwesen, (1990), (unverOjfentlicht).
8 Statistisches Bundesamt: Ergebnisbericht zur Statistik der Unflille bei der Lagerung und beim Transport
wassergeflihrdender Stoffe 1988-1990. Beirat beim Bundesminister fiIr Umwelt, NaturschulZ und
Reaktorsicherheit, Ausschuss "Statistik", Statistische Bundesamt, Wiesbaden, (1992).
9 PlIppel, M., Strassenverkehrsunflille von Gefahrguuankfahrzeugen 1989 bis 1991. Berichte der Bundes-
anstalt fiIr Strassenwesen, Unterreihe "Mensch und Sicherheit", Heft MS, (1993).
10 Bundesanstalt fiIr Strassenwesen 1991: Schlitzung der gesamtwirtschaftlichen Unfallkosten fiIr das Jahr
1989, (1991), (unverilffentlicht).
II Degranges, J.P., Les donnees d'acidents routiers disponibles au CEPN, Rapport CEPN No 146, Paris.
France.
12 Castelano, S.; Degrange, J.P.; Hubert, Ph., et al., Le transport de l'ammoniac anhydre, Aanlyse et estimation
des risques, CEPN No 159, Paris, France, (1989).
13 Hubert, Ph.; Pages, P.; Castelano, S.; Degrange, J.P., Le risque associ05 au transport du propane, CEPN No.
168, Paris, France, (1990).
14 Hubert, P.H.; Pages, P.; Degrange, J.P., EsJrnation regionale du risque associ05 au traffic des mati~res
dangereuses: comparaison d'itin05raires roull.,rs II Lyon (1988). CEPN No 129. Paris, France.
IS MfMD, Mission du Transport des mati~s Dangereuses. Accidents et incidents de transports de mati~s
dangereuses par voies routi~ et ferroviaire. Minist~ des Transports, Paris.
16 Appleton, P.R., Transport accident frequency data, their sources and their application in risk assessment
United Kingdom Atomic Energy Authority, SDR R 474, Warrington, UK, (1988).
17 HSC (Health and Safety Commission): Major Hazard Aspects of the Transportation of Dangerous
Substances Advisory group on dangerous substances, Her Majesty Stationery Office, London, (1991).
18 Dennis, A.W.; Foley, J.T.; Hartman. W.F.; Larson, D.W., Severities of Transportation Accidents Involving
Large Packages Sandia Laboratories, Report SAND77-0001. Albuquerque, USA, (1978).
19 Geffen, G.A. et al., An Assessment of the Risk of Transporting Propane by Truck and Rail, PNL-3308,
Richland. USA. (1980).
2OTNO: LPG, A study, TNO report for the Public Ministry of Housing, Physical Planning and the
Environment, TNO, Appeldoom, Netherland, (1983).
21 Hubert, P.; Pag~s, P., International discrepancies in the estimation of the expected accident and tank failure
rates in studies, in F. Saccomanno and K. Cassidy (eds.), Transportation of Dangerous Goods: Assessing
the Risks, Institute for Risk research, Univeristy of Waterloo, Waterloo, Ontario, Canada, (1992).
22 Major hazard aspects of the transport of dangerous substances, London, Health and Safety Commission,
HMSO, (1991).
23 Hughes, J.R., Storage and Handling of Petroleum Liquids - Practice and Law, 2nd ed., Griffm, London,
(1970).
CHAPfERS
DATA BASES AND COMPUTER SUPPORT FOR RISK ASSESSMENT
8.1. Data Bases for Hazardous Chemicals and Materials
The data basesi concerning dangerous goods and health hazards, listed in TABLE S.l,
provide useful information for risk assessment studies.
TABLE 8.1. Data bases for hazardous chemicals and materials
• ASHMIR (Hazardous Materials Incident Reporting System)

• CHEM-BANK
• CHEMDATA and HAZDATA
• CHEMICAL INFORMATION (CCINFO disc series A)
• CHEMSAFE Facts Data Base
• CHEMSURE Safety Information System
• DANGEROUS GOODS CD-ROM
• DECHEMA Information Systems and Data Bases
• ECDIN
• ENERGY LIBRARY
• ENFLEX INFO
• ENVIRONMENT LIBRARY
• EXIS International Dangerous Cargo Data Base
• HAZARD-MASTER
• Information Services of the Royal Society of Chemistry
• Material Safety Data Sheets (MSDS CD-ROM)
• MHIDAS (Major Incident Data Service)
• Occupational Health and Safety Information (CCINFO disc series B)
• OHS MSDS ON DISC
• OSH-ROM
• PC-FACTS
• POISINDEX
• SIGEDA-PC Information Systems
• TOMES
i Note: Ai; listed in CD-ROM in print (1988-1989), An International Guide, Meckler Publishing House.
236 CHAPTERS
The list is not exhaustive and there exist, however, other data bases which belong to
private organizations and are not always easily accessible to the public (e.g., transport
and distribution of petroleum products; accident statistics of trucking companies, etc.).
Paul Scherrer Institute (PSI) and the Swiss Federal Institute of Technology in ZUrich
(ETHZ) have established in 1993 a long-teno project) on "Comprehensive Assessment
of Energy Systems." The ultimate goal of this project is an integrated evaluation,
covering the risk-related, environmental and economical aspects associated with
different energy systems of interest for Swiss conditions. Within the framework of this
project a "PSI-Severe Accident D(ltabase" was created, which includes valuable
infonnation from other data-bases and organization, with emphasis on energy-related
events. TABLE S.2 indicates the origin of the data which were used for, or added to the
PSI data base.
TABLE 8.2. Data bases containing historical accident data
Code Name Description Organization Country

of Origin
OFDA The Office of Foreign Disaster Assistance OFDA USA

Database
MHIDAS The Major Incidence Data Service UKSRD UK
WOAD The WOAD Offshore Data bank DNV Norway
MMS The Minerals Management Service Accident USA
MARS The Major Accident Report System CEC-Ispra EEC
AHE The Acute Hazardous Events Database EPA USA
OPLIS-SIGMA Disaster Data & the OPUS Database SR CH
- The Accidents Handbook UBA BRD
Other valuable infonnation can be found in "PSAM II Conjerence".2 The purpose of

the PSAM Conference is to provide a forum for the presentation of scientific papers
covering both the methodology and applications of system-based approaches to the
design and the effective safe operation of technological systems and processes. These
include nuclear plant, chemical and petrochemical facilities, defense systems, aerospace
systems, and the treatment and disposal of hazardous wastes.
8.1.1. DESCRIPTION OF SELECTED DATA BASES
ASHMIR Data Base

The results of reporting incidents involving hazardous materials (including fires) on a
nationwide basis in Australia have been described? The "Hazardous Materials Incident
Reporting System (ASHMIR)", was developed on the behalf of the Australian
Government, and started in 1989. "Australian Hazardous Materials Incident Reporting
System (ASHMIR) fl.
The system is also capable of integration into international data bases such as the
UK Health and Safety Executive I Safety Reliability Directorate "Major Incident Data
DATA BASES AND COMPUTER SUPPORT FOR RISK ASSESSMENT 237
Service" (MIDDAS). The AFIRS method of coding and the Australian Standard "AS
2577-1983" and certain fields from MHIDAS were applied.
"Austra1ian Hazardous Maurillls Incident Reporting System (ASHMIR)"

The statistics show that:
• From 523 incidents reported in half a year, 213 of the incidents involved petroleum
products, of which the main risk was flammability.
• The most common type of incidents were spills (42%), and leaks (37%). One boiling
expanding liquid vapor explosion (BLEVE) was also reported. Contamination (1.4%)
and explosions were rated the lowest (1.8%).
• Fixed property incidents account for (60%), and mobile property incidents (road
transport)for (40%), water transport (6%), and rail transport (1%).
• A total number of 439 people were injured (67 injuries to fire-fighter, and 372 injuries to
others) and one fatality was observed. The highest number of people injured in one
incident was 92 people.
• The day on which the highest number of incidents occurred was Friday, the lowest on
Sunday.
The reports were entered into a DBase-ill plus and data analysis was carried out using
Lotus 1-2-3. Data were collected on the main hazardous materials by United Nations
(UN) Number. A total of 106 UN Number was reported, and the percentage of incidents
is shown below:
Number of Reported Incidents and Type of Products 'n"olved

• Motor spirit (156 incidents, or 0%)
• Petroleum spirit (107 incidents, or 20.5%)
• Methane (53 incidents, or 10.1%)
• Petroleum gases, liquefied (36 incidents, or 6.9%)
• Hydrochloric acid (17 incidents, or 3.3%)
• Ammonia (13 incidents, or 2.5%).
CHEM-BANK Data Base

Type Database - bibliographic, updated quarterly.
Computer: IBM PC, AT, XTor 100% compatible.
Available: from Silver Platter - World Wide Distribution Network.
Description: Collection of data banks on potentially hazardous chemicals, with three
complete data banks: Registry of TOXIC effects of chemical substances (RTECS) from
National Institute of Occupational Safety and Health, USA, Oil and Hazardous
Materials - Technical Assistance Data System for U.S. EPA, and Chemical Hazard
Response Information System from the US Department of Transportation, (Coast
Guard).
238 CHAPTER 8
CHEMDATA and HAZDATA Data Bases

Description: CHEMDATA4 and HAZDATA from the National Chemical Emergency
Centre at Harwell provide hazanl information and emergency action advice with
sources of specialist advice..It runs on several computer systems, PC, and CD-ROM.
Chemical Information (CCINFO disc series A)

Type : Database - full text, updated quarterly.
Computer: mM PC, XT, AT or compatible.
Available: from Canadian Centre for Occupational Health and Safety (CCOHS)
Hamilton Canada.
Description: CCINFO disc series A contains data bases on chemicals including mate-
rial safety data sheets pure chemicals, natural substances, mixtures and products
registered in Canada. It includes the NlOSH Registry of Toxic Effects of Chemical
Substances Videotext information packages and various OH & S publications.
CHEMSAFE Facts Data Base and DDB Product Data Base

Description: CHEMSAFE provides information and advice on safety, with facts and
relevant figures. DDB is a product data base for chemical technology
CHEMSURE Safety Information System

Description: It is an information and reporting system for dangerous goods which runs
on an IBM AS/400 computer. It has been developed to support safety and environ-
mental protection in plants. It is used for managing plant information on dangerous
cargo/products: safety data sheets, accident sheets, operating instructions are among the
type of information that the system can store, manage and retrieve.
Dangerous Goods CD-ROM

Type CD-ROM Full text, images.
Computer: mM PC AT or compatible.
Available: from Springer Verlag GmbH, Germany.
Description: This dangerous goods data base contains seven sources of information
with over 400,000 names, synonyms and product names for 126,000 environmentally
risky substances. The seven data bases are: HOMMELL Handbook of Dangerous
Goods; CHEMDATA - Harwell Laboratories OPERATION FILES; Swiss fire Brigades
Association; MERCK (catalogue); BAG - Swiss Toxic Substances Lists; SUVA - Swiss
Accident Insurance Institution; VeL-Handbook of forms for use in chemical industry.
DECHEMA Information Systems and Data Bases

Description: DECHEMA5 provides the user with information services and data bases on
chemical technology, biotechnology, materials and corrosion data, product information,
vendors, and research institution in Germany.
ECDIN Data Base

Type CD-ROM Full text - reference.
Computer: IBM PC or compatible, Apple Macintosh.
Available: from Europe CEDIS Edizionic (Italy).
Description: Contains environmental information on 1,700 chemical products. It
permits the evaluation of real and potential risks connected with the use of the chemical
substances, as well as their ecological and economic implications.
Energy Library
Type CD-ROM Database - bibliographic reference, updated quarterly.
Computer: IBM PC, XT, AT, PS/2 or compatible: Apple Macintosh.
Available : from OCLC USA.
Description: This data base provides users with information on oil, coal, nuclear, solar
and other related energy sources. Over 290,000 records on the disc deal with hundreds
of energy related topics.
ENFLEXINFO
Type : CD-ROM Database (full text), updated by-monthly. There is a data base
full text demo disc available.
Available: from ERM Computer Services Inc. - USA
Description: Provides current compilation and the full text of US Federal and state
environmental regulations. The regulations are indexed in six ways; reference applica-
bility, requirement, full text and data.
Environment Library
Type CD-ROM Database - bibliographic reference, updated quarterly.
Computer: IBMPC, XT, AT, PS/2 or compatible: Apple Macintosh.
Available : from silver Platter - Worldwide Distribution Network.
Description: This data base provides over 400,000 records on environment, both
historical and current, on all topics dealing with pollution and environmental issues of
concern to environmental researchers.
EXIS International Dangerous Cargo Data Base

Description: The expert system EXIS uses a data base containing international data and
information regarding transportation, physical and medical properties, classifications/
codes and packaging information, and runs also on IBM-compatible PC'S.6
HA7ARD-MASTER
Type CD-ROM Full text, reference.
Computer: IBM PC, PS/2 or compatible.
Available: from USA-Technical Services Associates Inc.
Description: This disc offers a comprehensive library of safety transportation, and
environmental regulations with a powerful, accurate, and easy to use search and
240 CHAPTERS
retrieval capability. HAZARD-MASTER in perfect for regulatory compliance and

reference.
Information Services of the Rpyal Society of Chemistry

Description: This service is available on-line (DATA-STAR/CHIN, DIALOG
Information Services, Inc. (File No. 317, STN International), or on CD-ROM, and
covers topics such as occupational health and safety, legislation, protective equipment
and storage relating to the chemical and allied industries. Further topics are hazardous
waste management; storage and transportation; leaks, spills and unplanned releases;
fires and explosions, etc.
Material Safety Data Sheets (MSDS CD-ROM)

Type : CD-ROM Database-full text, reference, updated quarterly.
Computer: ffiM PC, XT, AT or 100% compatible.
Available : from Faxon Company Europe/USA.
Description: This data base contains 60 000 material safety data sheets providing
information about chemicals hazards in their use, toxicity, physical properties, ingre-
dients, reactivity, first aid instructions, spill disposal and fire procedures, special protec-
tive equipment, storage and handling procedures and other related data.
OHS MSDS ON DISC

Type : CD-ROM Demo disc - Data base (full text).
Computer: ffiM PC, XT, AT or compatible.
Available : from Occupational Health Services Inc.
Description: Material Safety Data Sheets on hazardous substances are contained on this
disc. All substances have been researched to reflect all appropriate information. Data
base is maintained current by reviewing. all appropriate publications approximately
13,000 per year.
Occupational Health and Safety Information

Type : CD-ROM Database - bibliographic, updated quarterly.
Computer: ffiM, PC, XT, AT or compatible.
Available: CCINFO disc series B from Canadian Centre for Occupational Health and
Safety (CCOHS), Hamilton, Canada.
Description: This CD-ROM includes NIOSHTIC (National Institute for Occupational
Safety) data base of bibliographic references covering international health and safety
documents, Canadian Occupational Health and Safety resources and studies, noise level
measurements, fatality reports Canadian bibliographic references, Videotext inform-
ation packages and various other Canadian publications.
OSH-ROM Data Base

Type CD-ROM Database - bibliographic, updated quarterly.
Computer: IBM PC, XT, AT or 100% compatible.
Available: from Silver Platter - World Wide distribution Network.
Description: This is a data base of Occupational Health and Safety information.

Contains three complete bibliographic data bases; NIOSHTIC, HSELINE and CISDOC.
Collectively they contain over 24,000 citations from over 500 journals and 100,000
monographs and technical reports. Also contains the major incidents data base
MHIDAS with over 5,000 incident records.
PC-FACTS
Description: PC-Facts is a chemical events data base developed by TNO Institute in
Apeldoom (Holland). The system has now been acquired by the Swiss Military Atomic
and Chemical Laboratory ("AC-Laboratorium") and is located in Spiez (Switzerland).
The information stored in PC-FACTS can be used to prepare reports on risk 1IDalysis
based on accident/incidents reported. i
POISINDEX
Type : CD-ROM Database - bibliographic reference, updated quarterly.
Computer: mMPC XT, AT or 100% compatible.
Available: from Micromedex Inc. (USA) Microinfo Ltd. (UK).
Description: This is a detailed toxicology data base designed to identify and provide
ingredient information for over one-half million commercial, industrial pharmaceutical,
zoological and, botanical substances. Also provides detailed symptomatology and
management/treatment protocols in the event of a toxicology problem due to ingestion
dermal absorption, eye exposure or inhalation of any of the substances listed. The data
base may be searched by brand name manufacturers name, generic or chemical name,
sheet or slang terminology or botanical and common name.
SIGEDA-PC In/omUltion Systems

Available: from ECOMED Verlagsgesellschaft mbH,7 Lech, Germany.
Description: This data base on CD-ROM contains over 10,000 chemical products'
descriptions with physical properties and MAK values. Safety data, health and safety
guidelines, as well as literature information are also included. 8
TOMES Data Base

Type CD-ROM Database - bibliographic, reference, updated quarterly.
Computer: mM PC XT, AT or 100% compatible.
Available: from Micromedex Inc. (USA) Microinfo LTD (UK).
Description: TOMES is an expanding industrial chemical data base with in-depth
coverage of clinical effects range of toxicity, work place standards, kinetics and
physicochemical parameters. Detailed, comprehensive and referenced protocols for the
medical evaluation and treatment of individuals exposed to chemicals agents are
provided. Protocols for initial response to incidents (fires, spills, leaks) from hazardous
materials are also included. special attention is given to regulated chemical, including
, AC-Laboratorium. Wimmis/Spiez. Schweiz: Contact person is Dr. P. Roder

(AC-Zentrum. Tel: 033-55 II II).
242 CHAPTERS
those on the, "extremely hazardous substances" list. Agents are indexed by chemical
names, synonyms and commonly associated numbers.
Tools for Hazard Classification

The "Fire & Explosion Index" (1976), is available from the Corporate Safety and Loss
Prevention, Dow Chemical Company. It can be used to categorize installations, pro-
cesses, and products relative to their danger concerning fire and explosion. Another
similar publi::ation is the "MoM Fire Explosion & Toxicity Index" (Lewis,9 1976).
Valuable information and a methodology for hazard classification can also be found in
F &EI, "Fire & Explosion Index Hazard Classification Guide".lo
Although the methods applies to chemical plants, it can be applied to calculation
pertaining to storage area as well. Process units or vessel, contributing most to a fire or
explosion are evaluated on the basis of assessing material factors and hazard factors to
them. The product of the "general and special hazard factors" represents the "unit
hazards factor". This figure measures the degree of hazard exposure of the unit and is
used with the material factor which represents the edge of loss exposure. The product of
the unit hazard factor and the material factor represent the F&EI. This is used to
determine the area of exposure surrounding the unit being evaluated.
8.2. U.S. Transportation Statistics and Information on Commodity Flow
8.2.1. TRANSPORTATION STATISTICS ANNUAL REPORT
~lle Transponation Statistics Annual Repon is a publication that summarizes the state
o~ the Nation's transportation systems and the issues acd consequences of maintaining
such a diverse and complex network, the quality of statistics used to characterize the
transportation system, and planned efforts by the BTS to improve the quality of
statistics. All four transportation modes - airways, highways, railways, and waterways -
are examined through available data and statistical studies. The 1995 edition of the
TSAR will be available in July 1995 and will focus on transportation and the economy.
In addition, the 1995 TSAR will provide analysis and interpretation of the information
contained in the National Transportation Statistics cited above.
8.2.2. NATIONAL TRANSPORTATION STATISTICS
The National Transponation Statistics is a publication containing selected national

transportation and transportation-related data from a wide variety of government and
private sources. The data illustrate transportation activity for the major transportation
modes in five-year increments, 1960-1992193. Transportation trends in performance,
safety, and motor vehicles sales, production and costs are also presented.
Supplementary information includes data on transportation and the economy, energy
consumption, energy intensiveness, energy transport, and energy supply and demand.
Additionally, statistics on commercial space transportation andjoumey-to-work data are
illustrated. The 1995 NTS has been redesigned to complement the Transportation
Statistics Annual Report cited below. The NTS is also available on diskette in Excel and
Lotus formats.
The FAA Statistical Handbook of Aviation is published annually by the Federal
Aviation Administration (FAA). Its prime purpose is to server as a convenient source
for historical data and to assist in evaluating progress. The Handbook should provide a
valuable source of information for organizations interested in aviation.
The 1993 Commodity Flow Survey (CFS) is the first comprehensive effort since
1977 to learn where and how goods are shipped in the U.S. The Census Bureau has
released preliminary national tabulations of value, tons, and ton miles of shipments by
mode of transportation, commodity, distance shipped, and size of shipment. In addition,
the DOT and the Census Bureau have released more detailed information from the CFS
including the flows of individual commodity groups between and within States, as well
as flows among 89 National Transportation Analysis Regions.
8.3. Equipment Reliability Data Bases
• Rasmussen Report WASH-1400 (1975): It is concerned with reactor safety studies.

Availability: published report. Note: One of the early historical sources of nuclear
power plant equipment reliability - Contains some data generic to other industries .
• IEEE Std 500 (1984): A large collection of component reliability data much of which
has originated from nuclear sources. Availability: published
• IPRDS In-plant Reliability Data System: Large detailed data base on nuclear plants
which is updated continuously. Awilability: published and computerized.
• NPRDS Nuclear Plant Reliability Data System: The data base is managed by the
Institute of Nuclear Power Operations - Contains nuclear plant data in a consistent and
comprehensive format. Failure data on more than 45,000 events and it may well be the
most significant nuclear data base in the near future. Availability: on-line access and
data retrieval, floppy disks, summary report.
• Offshore Reliability Data Handbook (OREDA): Contains reliability data from off-
shore drilling and production platforms which concentrates mainly on the North Sea
Availability: published.
• SRD component and reliability data bank: Contains a large amount reliability data
from a wide range of industries which includes nuclear and other forms of power
generation from world wide sources. Also available are incident and accident data bases
on chemical, explosive and environmental incidents. Availability: only to members of
the SRD Association (for information apply to the Business Manager SRD Association,
UKAEA, Wigshaw Lane, Culcheth, Warrington WA3 4NE UK)
• Risk assessment data banks at the Savannah river site: The Savannah River Site
maintains a computer data base system for non-reactor facilities that contains a com-
pilation of the incidents that have occurred since 1953. Five risk assessment data banks
244 CHAPTERS
exists in the areas of reprocessing, fuel fabrication, waste management, tritium, and the
Savannah River Technology Center. They provide data for use such as failure rate
analyses, equipment reliability and breakdown, studies, etc.
8.4. Center for Exposure Assessment Modeling (CEAM)
CEAM serve~ as the focal point for Office of Research and Development's multimedia
exposure assessment modeling and ecological risk assessment. CEAM information and
software products are available through the World Wide Web (WWW).
U.S. Environmental Protection Agency Voice: 706/546-3549

National Exposure Research Laboratory Fax: 706/546-3340
- Ecosystems Research Division BBS: 706/546-3402
960 College Station Road 1,200-14,400 Baud, N/S/l
Athens, GA 30605-2700 USA Internet: ftp@ earthl.epa.gov
Last updated: 06112195
For further information on this and other software products, contact:
Center for Exposure Assessment Modeling (CEAM)

Voice: 706/546-3549
Internet: ceam@athens.ath.epa.gov, dd@hennes.ath.epa.gov
BBS: 706/546-3402 FAX: 706/546-3340
The information and software available through WWW were pr.::pared in support of
U.S. EPA Sdentific Programming Contract, by the Exposure R~search Laboratory,
Athens, GA, USA.
8.5. Transportation Risk Management System
A transportation risk management system (XTRIM) is available for road and rail
transport in Denmark and the Lombardia (Italy). The underlying models for accident
frequencies and consequences are mainly based on empirical analysis of historical
accident descriptions.
A study on the risks of transporting dangerous goods in inland waterways, restricted
waters and coastal waters is currently being undertaken. An accident datallase compris-
ing 3521 accidents happened during transport of dangerous goods has been established
using more than 20 different sources from the open literature. 1869 of the accident
records describe marine and inland waterways accidents. Each accident record contains
42 information fields. The database is operated under Oracle.
An important contact has been made to Ministry of Transport and Public Works in
Holland, where they have a large knowledge on inland waterways transportation. It is
hoped to receive 17,000-18,000 accident descriptions from a ten year period on the
inland waterways of Holland. The future plan for the transportation risk research is to
develop frequency and consequence models suitable for the XTRIM enabling this to
make risk estimates for road, rail, marine and inland waterways transport of dangerous
goods within Europe.
8.6. Spatial Data and Geographic Information Systems Within the U.S. Bureau of
Transportation Statistics (BTS)
An important driving force in BTS (Bureau of Transportation Statistics, U.S.

Department of Transportation) data compilation and analysis activities is the need for
extensive geographic data on transportation facilities and networks, the activities they
serve, and the surroundings they affect. The emphasis on geographic data and analyses
reflects the central purpose of transportation: to connect, separate locations and
accommodate the flow of people and goods. This requires refinement and application of
analytical methods based on geographic information systems (GIS) technology.
GIS provides a practical and realistic method for measuring and understanding the
extent, use, performance and consequences of the transportation system.
The paper by Hancockll (1994) presents a plan to identify and obtain the required
information to support a GIS structure, integrate this information into cohesive tools
necessary to support planning and policy initiatives, and make the information available
for use by the public and other agencies. The highlights of this important paper are
presented below.
Overview
BTS is compiling a geographic data!Jase to support research, analysis, and decision
making across all modes of transportation. The database is designed primarily to meet
the needs of DOT at the national level, but will have major applications at State and
local scales throughout the transportation community. Most of the database is to be
published as a National, Transportation Atlas, including base maps that portray the
extent and connectivity of the nation's, transportation system. thematic maps' that
portray the use, and consequences of transportation, and electronic files that are used to
develop the maps and analyze transportation networks. The remaining part of the
database will be used for internal BTS analyses, but will be left to others to publish.
The National Transportation Atlas database is organized into four layers of inform-
ation:
• The facilities layer represents the location, physical characteristics, and connectivity
of highways, railroads, waterways, fixed guideway transit, airports, pipelines, terminals.
bridges, locks, structures.
• The service layer represents transportation services on and across the transportation
facilities, such as bus lines and railroad trackage rights.
I A thematic map is a map that displays one or more specific types of information usually using symbols to
represent different classes or movements of information.
246 CHAPTER 8
• The flows layer represents interactions between areas, such as commodity

movements and donor-donee financial relationships, which are assigned to specific
transportation facilities and services primarily with analytical models. Much of the flow
data will come from the American Travel Survey and the Commodity Flow Survey,
both of which are being conducted by the Bureau of the Census for BTS, and from the
Census Transportation Planning Package of journey-to-work data that was published on
CD-ROM by BTS.
• Finally, the background layer includes data of use to the transportation community
that are published and maintained by others. Examples include political boundaries,
geographic names, population distributions, economic activity, and environmental
conditions. Boundary and name files that help define locations for the facility, service,
and flow layers are to be published in the National Transportation Atlas. The remainder
of the background layer is used by BTS for creating thematic maps and preparing
analyses, but is left to other agencies for dissemination.
Consecutive to this work is a U.S. national attempt to modernize the Nation's
mapping programs through interagency and intergovernmental cooperations, Coordi-
nation of Geographic Data Acquisition and Access: the National Spatial Data Infra-
structure (NSDI).
NSDI provides basic location (but not necessarily connectivity) data for the facilities
layer of the National Transportation Atlas through the Census Bureau's TIGERfiles and
the Digital Line Graphs (DLGs) of the U.S. Geological Survey; BTS provides updates
on location and assures adequate representations of connectivity for facilities of
potentially national significance that appear on the NSDI transportation layer.
Contents of BTS Geographic Data Bases

BTS spatial data consists of two classes of data, information used in the National
Transportation Atlas and information from sources outside the Department of
Transportation for use within BTS and the Department. The databases used by BTS
have been created by different agencies within the Federal Government. In general, the
networks have been made available at two scales; 1:100,000 and 1:2,000,000. Root-
mean square positional accuracy.
Attributes. Two types of attributes have been defmed for the National Transportation
Atlas databases: locational information, and transportation attributes.
Locational attributes provide information necessary to represent the data spatially,
to uniquely identify every element in the databases, to provide connectivity of trans-
portation facilities and services, and to provide referencing systems for transportation
attributes. In general, these attributes will be located in a single database.; The proposed
attributes include facility network link, facility network node, facility terminal and other
structures, service network, flow layer point, line, boundary, and area databases,
respectively. Facility network nodes and facility terminals are located in separate
databases. Nodes are components of networks (generally, where two links intersect) and
, For the modal networks, each database may consist of one or more files that have link, node and geometry
information.
are not necessarily terminals. Similarly, tenninals are probably not located at a given
node although they will be referenced to the nearest node on the appropriate network.
Note that the locational attributes will probably consist of, or include, the National
Spatial Data Infrastructure (NSDI) core attributes, which have not yet been formalized.
Transportation attributes consist of information used for planning and analysis.
These attributes will probably be maintained in several databases and will be identified
from specific analysis requirements and by the aVailability and quality, of the desired
data.
International Extension. With enactment of the North American Free Trade

Agreement (NAFTA) and growing interrelationships of U.S. domestic-transportation
and international trade, BTS data bases must be expanded beyond the ' U.S. borders'.
U.S. Modal Networks. Of the three principal modal networks, the highway network
data base consists of approximately 400,000 miles of highways, or 10% of the total
roads in the United States. The rail network contains the entire mainline and secondary
rail systems while the waterway network includes all navigable waters.
Information. Information currently available for the NTA is primarily locational in

nature. To increase the usefulness of the NTA databases, additional attributes must be
collected or obtained from other sources such as the Highway Performance Monitoring
System (BPMS). These attributes must then be added to the existing databases or
overlaid onto the networks and facilities. In many cases, attribute information becomes
available on a state by state or regional basis. Generally, BTS will establish an attribute
code in the appropriate database. As the information for that attribute is obtained, the
code will be updated. In this way, the National Transpr,rtation Atlas databases will
contain the most current information but will not lead I:sers into misunderstanding
attributes.
Because the flow of people and goods is usually intermodal, BTS will incorporate
specific connections between networks and terminals. Initial connections include direct
node and link: references within the terminal databases. Currently those databases
include only the major airports, rail/highway tenninals, water ports and transit stations.
As appropriate, standard GIS software will be used to display different attributes
across a network to identify obvious discontinuities within that attribute field. For
example, the different categories of highway functional classifications will be
highlighted in different colors. Sections of highway that have a discontinuous link: (Le.
a link: appears red with yellow links on either side) will be noted as having a
discontinuous attribute. All such variations will be noted and addressed as discussed
above.
Maintenance. The information within each of the databases will be maintained by the
appropriate agency. BTS will maintain selected transportation attributes that are not
maintained by another agency and will support other efforts as requested. For archival
purpose, BTS will maintain archival copies of all information used in the BTS spatial
248 CHAPTER 8
data framework, including the National Transportation Atlas, and for ~y analysis
performed by or for BTS.
Integration and Analysis Tools.

With widespread availability of commercial geographic information systems software,
many tools and applications are commercially available. Whenever the needed capa-
bilities are accessible through vendors, BTS will acquire this software. If necessary,
BTS will develop tools and will make the technologies available to other interested
users.
Examples of these tools include:
• Geographic Data Visualization Tools:
user selected multi-dimensional tabulations
- chloropleth maps
- flow maps
• Network Simplification-Tools:
- link chaining (i.e. end-on link. collapsing)
- network spur removal (i.e. removing dead-ends)
- sub-network selection
- network aggregation (i.e. merging of two or more parallel routes)
• Network Matching Tools.
BTS is also developing a GIS Center within the Department of Transportation
headquarters. The goals of this center are to:
- integrate geographic data products for publication and use by DOT
- prepare maps and other vi!lual di."plays
- analyze temporal and geograpb!c variation in the extent, use, performance, and
consequences of the transportation system, and provide a "test bed" for the
development and demonstration of GIS technology and geographic analysis
methods for DOT and other users.
Combining software and the GIS Center, BTS will have the necessary tools to
provide integration and analysis support to the transportation community. The mission
of BTS is "to compile, analyze, and make accessible information on the Nation's
transportation systems".[7] To meet this mission, BTS is actively developing products
that disseminate this information to the public.
For spatial data, BTS will make available two types of products: database files and
both paper and electronic (displayed on a computer screen) maps. The database flIes can
be used for analyses as well as generating maps and will be available as the National
Transportation Atlas databases. The maps will be available as the National Trans-
portation Atlas and as-thematic maps for use in BTS and Office of the Secretary of
Transportation (OST) reports. Another product will be any user aides and software
tools, or corresponding procedures, developed by BTS for use with the Spatial data.
These products may· be in the form of computer software or technical reports.
Atlas Databases. The BTS will provide annual compact discs of the National
Transportation Atlas databases, initially at two scales: 1:100,000 for local and regional
purposes, and 1:2,000,000 for national evaluations. The discs will contain all associated
databases in ASCII format, the corresponding documentation, and display and
extraction software beginning with Version l.0 in 1995.
Thematic Maps. As required to support the Transportation Statistics Annual Report,

the National Transportation Atlas. and other DOT initiatives, BTS will provide thematic
maps of information from the spatial data framework and from analyses using the BTS
spatial databases. Generally, these maps will be available in both raster and vector
formats as well as on paper at I: I ,000,000 or 1:7,500,000 scale of detail.
Conclusion
Spatial data and GIS technology are as important as traditional statistics to BTS.
Geographic and temporal variations must be understood to understand transportation
issues. In support of this belief, BTS has committed to compiling, developing and
maintaining a collection of databases that represent the current transportation system
along with the necessary user aides to make this data available to the general public.
This collection is generally referred to as the National Transportation Atlas databases
and will be available to the public during the flI'St quarter of Calendar Year 1995. BTS's
goal is to provide data and analysis tools, but not turn-key solutions to transportation
issues.
8.7. Computer Support for Risk Assessment
Different computer support systems for risk assessment and management of dangerous
goods transportation problems are presented below. Software for risk assessment, and
transportation risk in particular, is usually complex, requires large amount of input data
of various kinds (data bases) and a considerable amount of effort in both the problem
formulation and the interpretation of the results. Risk assessment often concentrates on
human fatality risk, and sometimes on property damage. There is, however, an in-
creasing interest for including environmental consequences into a truly multicriteria risk
assessment.
8.7.1. COMPUTER PACKAGES FOR HAZARD ANALYSIS
The software/codes mentioned below were developed and are sold by Technica 12 in
London:
WHAZAN-ll
It is a chemical hazard analysis software package designed to run on an IBM PC. It
helps in assessing the toxic and flammable consequences posed by the release of
hazardous industrial substances using simple analytical methods. Programs predict gas,
liquid or two-phase releases from a leak; thermal radiation from burning pools, jets, or
250 CHAPTER 8
fireballs, explosion blast effects; pool spreading and evaporation; and plume rise and
dispersion. A physical and thermodynamic properties data base is included in the
package for 20 commonly occurring hazardous chemicals. Users may add extra sub-
stances to this library, if required. WHAZAN-ll contains also a sophisticated dense
gas/passive dispersion program, which is particularly useful in simulating the dispersion
behavior of pressurized or refrigerated gases, such as ammonia, chlorine, propane, etc.
The mathematical models included are:
- Liquid Outflow
- Gas Outflow
- Two-Phase Outflow
Spreading & Evaporating Liquid Pool
- Jet Dispersion
- Adiabatic Expansion
- Plume Rise
- Dense Cloud / Gaussian Dispersion
- Gaussian Puff Model
- Passive Plume Dispersion (uses a Gaussian dispersion model with a point source
which can be at an elevated position)
- Burning liquid pool
- Burning fluidjet
- Fireball and BLEVE
- Exploding vapor cloud
- lndoorconcentrationsldoses
SAFEI'I
It is an internationally recognized package for quantitative risk analysis. It automates
the risk assessment of chemical and petrochemical facilities, where toxic and flammable
materials are manufactured, stored and transported. SAFETI combines a complete
library of rigorous mathematical models which, either singly or in combination are used
to derive the individual and societal risk associated with the release of hazardous
chemicals. SAFETI automatically analyses the complex sequences from accident
scenarios, taking into account local population and weather conditions. The output is in
the form of individual risk (contours) and societal risk (F-N) curves.
TECKJEI'
It is a program for modeling continuous jet releases. It assesses the effects of toxic and
flammable clouds produced by pressurized. releases from a vent stack or flare, if the
cloud ignites TECJET shows the extent of the effects. Concerning toxic clouds,
TECJET shows the effect on local population centers.
DATA BASES AND COMPUlER SUPPORT FOR RISK ASSESSMENT 251
CARA
It contains a set of modules for conducting various forms of reliability analysis, i.e., the
module FMECA is used for failure mode effect, and criticality analysis, and the module
CAFTAN for fault tree analysis.
PHAST
It enables the engineer to identify situations which have the potential for causing
unacceptable consequences to the population and the environment PHAST is derived
from SAFETI:
LPG-l model
A model LPG-l has been described by Van Erde et al. 13, which can assess the risks
from transportation accidents related to LPG. The model's first module contains a
representation of the spill mechanism following a transportation accident It determines
the amount of vapor flash-off and the size of any remaining pool based on the
atmospheric conditions at the time of the accident, the amount of pre-heating of the
container, and the thermodynamic properties of the material shipped. Subsequently,
simplified fireball, vapor cloud explosion and BLEVE models are used to compute the
extent to which pre-specified levels of shock-wave and heat radiation damage will be
reached and the radius within which direct flame exposure will be experienced. During
each phase of the analysis, the model considers relevant site-specific mitigating factors,
such as the wind speed, air stability, combustion characteristics and the probability of
encountering ignition sources.
8.7.2. TRANSPORTATION RISK AND COST ANALYSIS MODELS
Risk and Cost Analysis Models

In order to investigate whether dangerous goods transports should be considered as a
problem, it is necessary to investigate the transport and the traffic system. What
distinguishes a dangerous goods transport from an ordinary transport, is the fact that in
case of an accident during transfer, acute injury to people and damage to property may
result
The quantifying of the risk of dangerous goods transports directly from statistics is
not always possible due to the problem of limited availability of statistical data, 14
however it is possible to assess to some extent the effect of selected road parameters on
traffic safety. Methods for estimating true accident rates from historical data using the
''Empirical Bayes Method" accident rates at road intersections using Poisson distri-
butions are emerging. Different aspects of road design and influence of the topography
have been investigated by Polus and Dagan lS (1987), Lamm and Choeiri l6 (1987).
While it is now possible to calculate the influence of curve radius, road width, incli-
nation and intersection on the frequency of traffic accidents, it is not yet possible to
calculate the influence due to the remainder of the parameters.
252 CHAPTER 8
Another problem is encountered when extremely dangerous hazardous materials

must be transported. One is faced with the problem of how to define credible risk
criteria to be used in transport planing, which aims at minimizing the risk, in the event
of an accident on the way. Any model that is considered for extremely dangerous goods
must make use of the fact that the current routing scenario ends at the occurrence of the
first accident. In such a case the best route is the one which minimizes the risk if an
accident occurs on the way. This means that the expected conditional risk, given that an
accident occurs must be minimized. A mathematical formulation of this special problem
has been reported by Sivakumar and Battal1 (1993), Sivakumar et aI.,IS (1991). In their
paper Glickman and Sherali l9 (1991) have introduced a similar model, where the objec-
tive is to determine a route that minimizes the expected number of fatalities given that
the number of fatalities exceed a certain value.
RISKMOD
The computer model RISKMOD was developed to provide assistance in the evaluation
of policies for the movement of dangerous goods, by estimating objective risk. 20 The
model represents individual vehicle shipments of dangerous goods on the truck and rail
networks for which the associated risks are estimated in a series of steps:
• The first step involves an accident rate prediction, followed by a spill rate prediction,
given an accident.
• In subsequent steps, the damages are evaluated of an accident alone, and of an
accident followed by a spill of goods.
• The final tables provide link by link risk estimates and a summary of the total risk for
the entire route.
A novel feature of the RISKMOD model is the valuation of consequences, which
provides a common denominator in terms of expressing risk, so that all risks may be
summed on a route.
Improved RlSKMOD Model

This paper describes recent modifications to the model which have been made, based on
new data. Specifically, the risk associated with the mechanical aspects of an accident are
included separately from the risk due to the release of the dangerous cargo. In addition,
more detailed truck release data have been included to better reflect the range of
consequences which follow a release event. Both modifications assist in providing a more
accurate and representative account of the risks associated with transporting dangerous
goods.
DERACS Simulation Tool

A prototype computer program DERACS has been developed which is capable of
simulating some of the consequences of a train derailment accident. 21 The program is
made up of a number of models which account for train derailment mechanics,
flammable liquid spills, fire effects on remote targets, fire impingement on tank cars
carrying dangerous commodities, explosion blast over-pressure and thermal radiation,

and heavy plume and puff dispersion. Most of these sub-models have previously been
developed and are available in the open literature.
The overall program consists of a pre-processor for inputting the surroundings data
and train consist data, a simulation module containing the various models, and a post-
processor which presents the results of the simulation in graphical form. The present
version of DERACS is limited in scope, and intended for demonstration purposes.
IRIMS Management System

IRIMS is a decision-oriented prototype system for the management of hazardous
substances and industrial risk developed under contract to the Commission of the
European Communities' Joint Research Centre (JRC) in Ispra, Italy, and implemented
by IIASA's Advanced Computer Applications (ACA).22 IRIMS consists of different
modules for information retrieval, simulation, optimization and multicriteria scenario
evaluation. The system is coupled to a number of data bases on hazardous substances,
industrial accident reports, directives and regulations, industries, waste streams. It uses
a geographical information system (GIS), an environmental impact assessment module
(for river water quality, long range atmospheric transport, and ground water quality
management simulation), and an industrial structure optimization module. Also part of
the system is a transportation risk and cost analysis module, which was developed in
collaboration with the Ludwig Boltzmann Institute (Kleindorfer and Vetshera,23 1985).
The major routes of the European transportation network have been included. The
modules allow the user to find several routing alternative (on the basis of the shortest
path and of risk/cost evaluation). The cost evaluation is based on freight rates sampled
from commercial firms and do include insurance and possible liabilities not covered by
the insurance. The risk analysis covers material damages as well ~s injuries and
fatalities. The seh:tion of optimal alternative solution is performed by the user via an
integrated multicriteria evaluation package DISCRET (Majchrzak,24 1985).
DERA Interactive Transport System

Another interactive prototype system has been developed with INRETS (Institut
National de Recherche sur les Transports et leur SecuriM) of the French D6partement
d'Evaluation et de Recherche en Accidentologie (DERA).25 The system integrates a path
generator and risk assessment module with a set of data bases and a GIS. This prototype
is used for finding save routes (rail/road). The digitized maps show land use and
population density, residential areas, industries, agriculture, forests, etc. Paths with the
lowest risk can be found explicitly. Information on accident rates are associated to the
different arcs connecting the traffic nodes. The risk value used is thus a product of this
rate and the population affected (as a function of land use).
XENVIS Environmental Information System

The project XENVIS has been developed under contract to the Dutch Ministry of
Housing, Physical Planning, and the Environment (VROM). This X-based environ-
mental information system has an interactive and graphic-oriented framework and a
post-processor for the risk assessment package SAFETI.26 The XENVIS path generator
254 CHAPTERS
offers the user the choice between railways and roads. Risk analysis for each route
segment is performed using the SAFETY package, which includes as output F-N curves
and risk grids. Risk contours can also be generated
8.8. Special Literature and Risk Assessment Studies
In this section we present some selected papers showing GIS applications related to
mapping and spatial analysis.
In the past, many State agencies have focused information resources efforts on
administrative processes internal to their organizations. In the USA (e.g.,Texas) there is
a strong interest in developing a business plan "to prioritize and schedule the develop-
ment of public funded programs for the acquisition of geographic information of
common interest and to seek partners in the public sector for data development". It is
believed that an open systems environment is necessary for the different U.S. States to
be able to fully utilize and coordinate the various data bases associated with such
project. Each U.S. agency, while continuing to be responsible and accountable for its
primary mission, must also support the open systems concept in such a manner that
allows for the easy accessibility and use of GIS data layers by all other governmental
bodies.
In Switzerland, a Federal Confederation, where the different Cantons have important
rights to decide an govern themselves, there are many problems concerning the
coordination of such enterprise. It should be the role of the Swiss Federal Government
to create an advisory commission to coordinate and promote GIS activities, and to
develop what we could call a "State Strategic Plan for Information Resources
Management". Switzerland is now in the proce~s of preparing digitized maps at the
cantonal level to suit the needs of public uti.lities and environmental protection.
Integration and accessibility of the different data bases is not yet as advanced as in the
USA.
A strong GIS infrastructure will enable the creation of a map that encompasses all of
the elements needed to assist agencies when performing duties that: 1) enhance the
economic well-being of the state, 2) protect the environment, 3) conserve and manage
the vast natural resources, 4) provide new human resource opportunities through
planning, and 5) protect the lives and property of the citizens. The greatest lasting
benefit of GIS will come from assuring the availability of high-quality data to all state
workers.
The Texas Department of Transportation (TxDOT) recently fmished digitizing its
County/City map series and has made it available for distribution through the Texas
Natural Resources Information System (TNRIS). The County map series is the first
digital statewide data layer made available as a result of the coordination efforts of the
Geographic Information System Planning Council (GISPC).
8.8.1. GIS MAPPING AND SPATIAL ANALYSIS
The following papers give information on mapping and spatial analysis using GIS
technology (Figure 8.1).
1) Stockwell, J.R.; Sorensen, 1.W.; Eckert, J.W.; Carreras, E.M., The U.S. EPA
geographic infonnation system for mapping environmental releases of toxic chemical
release inventory (fRl) chemicals, Risk Analysis 13, (1993), pp. 155-164.
2) Von Braun, M., The Use of GIS in Assessing Exposure and Remedial Alternatives at
Superfund Sites, in M.F. Goodchild, B.O. Parks, and L. T. Steyaert (eds.), Environmental
Modelling with GIS, Oxford University Press, Oxford, (1993), pp. 339-347.
3) Wadge, G.; Wislocki, A.; Pearson, E.J.; Whittow, J.B., Mapping natural hazards with
spatial modelling systems, in P.M. Mather (ed.), Geographical Infonnation Handling -
Research and Applications, John Wiley & Sons, Ltd., (1993), pp. 239-250.
4) Wadge, G.; Wislocki, A.; Pearson, E.J., Spatial analysis in GIS for natural hazard
assessment, in Goodchild, B.O. Parks and L.T. Steyaert (eds.), Environmental Modelling
with GIS, Oxford University Press, Oxford, (1993), pp. 332-338.
Figure 8.1. References concerning GIS mapping and spatial analysis.
8.8.2. GIS AND TRANSPORTATION
GIS tools are beeing increasingly used in applications concerning transportation, for
representing ecosystems and other environmental resources, and display the range
affected by the consequences resulting from an incident scenario (e.g., using cercles to
represent the surface area of a possible spill, or plotting corridors along a given road
track). The papers listed below give an idea of such applications:
1) Barkan, C.P.L.; Glickman, T.S.; Treichel, T.T., Tank car spills into aquatic eco-
systems, paper presented at the 1993 TRB Hazmat Transportation Conference,
Albuquerque, New Mexico, USA, (1993).
2) Brainard, 1.S.; Lovett, A.A.; Parfitt, J.P., Using GIS for the risk assessment of aqueous
waste transport in Southern England, paper presented at the Society for Risk Analysis
annual meeting, San Diego, California, (December 1992).
3) Brainard, J.S.; Lovett, A.A.; Parfitt, J.P., GIS as a risk assessment tool for aqueous
waste transport, in EGIS'93: Proceedings of the Fourth European Conference and
Exhibition on Geographical Infonnation Systems, Utrecht, EGIS Foundation, (1993), pp.
391-400.
4) Freckmann, P., Route calculation for dangerous goods transport with a graphical
infonnation system, in EGIS '93: Proceedings of the Fourth European Conference and
Exhibition on Geographical Infonnation Systems, Utrecht, EGIS Foundation, (1993), pp.
1132-1138.
Figure 8.2. Refences concerning GIS applications related to transportation.

256 CHAPTER 8
5) Lepofsky, M.; Abkowitz, M.; Cheng, P., Transportation hazard analysis in integrated
GIS environment, Journal of Transportation Engineering, 119, No.2 (1993), pp. 239-254.
6) List, G.F.; Turnquist, M.A., Routing and emergency response team siting for high-
level radioactive waste shipments, Paper presented at the 1993 TRB Hazmat
Transportation Conference and subsequently for the special issue on Emergency
Management Engineering of the IEEE Transactions on Engineering Management, March
1994. (Available only from authors, as of 10 May, 1995).
7) Padgett, D.A., Assessing the safety of transportation routes for hazardous materials,
Geo Info Systems 2 (1992), pp. 46-48.
8) Parfitt, J.P., et al., The Risk Management of Hazardous Wastes, Their Transport and
Disposal. Report to the British Oxygen Chemical Foundation, University of East Anglia,
(1993).
9) Patel, M.H.; Horowitz, A.J., Optimal routing of hazardous materials considering risk
of spill, Transportation Research, A 28, (1994), pp. 119-132.
Figure 8.2. Refences concerning GIS applications related to transportation (continued).
8.8.3. SPECIAL METHODS FOR MODELING AND EYALUATING TRANSPOR-

TATIONRISKS
Somer papers presented in this section demonstrate interesting methods for modelling
transportation risk, optimizing routing of dangerous goods, or evaluating the economic
of alternative routing strategies:
1) Abkowitz, M.; Eiger A.; Srinivasan, S., Estimating the release rates and costs of trans-
porting hazardous waste, Transportation Research Redord 977, Transportation Research
Board, NRC, Washington, D.C., (1984), pp. 22-30.
2) Current, J.R.; ReVelle, C.S.; Cohon, J.L., An interactive approach to identify the best
compromise solution for two objective shortest path problems, Computers and Operations
Research 26 (1990), pp. 187-198.
3) Glickman, Th.S.; Sontag, M.A., The tradeoffs associated with rerouting highway
shipments of hazardous materials to minimize risk", Risk Analysis 15(1) (1995).
4) Gopalan, R.; Kolluri, K.S.; Batta, R.; Karwan, M.H., Modeling equity of risk in the
transportation of hazardous materials, Operations Research 38(6) (1990), pp. 961-973.
5) List, G.; Mirchandani, P., An integrated network/planar multiobjective model for
routing and siting for hazardous materials and wastes, Transportation Science 25(2)
(1991), pp. 146-156.
6) ReVelle, C.; Cohon, 1.; Shorbrys, D., Simultaneous siting and routing in the disposal
of hazardous wastes, Transportation Science 25 (1991), pp. 138-145.
7) Saccomanno, F.F.; Chan, A.Y.-W., Economic Evaluation of Routing Strategies for
Hazardous Road Shipments, Transportation Research Record 1020, Transportation
Research Board, NRC, Washington, D.C. (1985), pp. 12-18.
Figure 8.3. Refences concerning modeling and optimization methods for transportation
8) Turnquist, M.A., List, G.F., Multiobjective policy analysis of hazardous materials

routing, in L.N. Moses and D. Lindstrom (eds.), Transportation of Hazardous Materials:
Issues in Law, Social Science and Engineering, papers presented at Hazmat Transport '91.
Kluwer Academic Publishers, Norwell Massachusetts, (1993), pp. 103-116.
9) Vincent, P.J., Computerised transport routes, Civil Protection 11 (1989).
10) Wijeratne, A.B.; Turnquist, M.A.; Mirchandani, P.B., Multiobjective routing of
hazardous materials in stochastic networks, European Journal of Operational Research 65,
(1993), pp. 33-43.
11) Zografos, K.G.; Samara, S., Combined location-routing model for hazardous waste
transportation and disposal, Transportation Research Record 1245, Transportation
Research Board, NRC, Washington, D.C., (1989), pp. 52-59.
12) Kalelkar, A.S.; Arthur D. Little, Inc., Proceedings of National Conference on Control
of Hazardous Material Spills, AlChE (1974) "Use of Multidimensional Utility Functions
in Decision Analysis in Hazardous Material Transportation: Applicability and Limit-
ations" (with Proceedings of the 80th National Meeting of AlChE, 1975).
13) A Review of Decision Analytic Methods Applicable to Natural Hazards Risk:
Reduction Policy Analysis, Graduate School of Business Administration, University of
Southern California. Working Paper No.3-3-1976.
14) Use of Multidimensional Utility Functions in Hazardous Shipment Decisions,
Graduate School of Business Administration, University of Southern California, Working
Paper No. 4-2-1976.
Figure 8.3. Refences concerning modeling and optimization methods for transportation (continued).
Summary
This chapter is dedicated to the presentation of data bases and computer support
systems for risk assessment. A number of relevant and specialized databases are
identified and summarized. Details are given on the US transportation statistics and
information on commodity flow. Spatial data and Geographic Information Systems
(GIS) are integrated within the computational framework of transportation risk of
dangerous materials. Computer software is intensively used for the calculation of risk
resulting from a given transportation system. Relevant software was identified and
summarized. Risk and cost analysis models are discussed in the context of risk
assessment tools for the decision making process.
258 CHAPTER 8
1 Hirschberg, S. et al., Assessment of Energy Systems (Ganzheitliche Betrachtung von Energiesystemen -

GaBE), Detailed outline of the project, 2nd version (1993), Paul Scherrer Institute, Wtirenlingen and
Villi gen.
2 PSAM II Conference, Proceedings of, Vol. 3 (Session 73-108), March 20-25, (1994), G.E. Apostolakis, and
J.S. Wu (Eds), San Diego, California, U.S.A.
3 Winder, C.; Tottszer, A.; Navratil, J.; Tandon, R., Hazardous materials incidents reporting: Results of a
nationwide trial, J. Hazard. Mater., 31 (1992), pp. 119-134.
4CHEMDATA, Vendor: National Emergency Centre, B7.22 Harwell Laboratory, Oxfordshire OX11 ORA,
UK
5DECHEMA, Deutsche Gesellschaft fUr Chemische Apparatewesen, Chemische Technik und Biotechno-
logie e.V., Frankfurt/Main, Germany.
6 EXIS Expert Information Systems, K.O. Storck Verlag, Stahltwiete 7, D-2ooo Hamburg 50.
8 ECOMED Verlagsgesellschaft mbH, Justus-von Liebig-Strasse I, Postfach 1752, D-891 0 Landsberg/Lech.
8 SIGEDA-PC, Das Gefahrstoff Informationssytem (Sicherheitstechnische Kenndaten). Vendor: ECOMED
Verlagsgesellschaft mbH, Justus-von-Liebig-strasse I, Postfach 1752, D-891O Landsberg I Lech.
9 lewis, D.1., The Mond Fire Explosion & Toxicity Index - A Development of the Dow Index, AIChE Loss
Prevention Symposium, Houston, (April 1979).
10 "Fire & Explosion Index Hazard Classification Guide", 5th Edition, Report No. LC 80-29237, Published by
American Institute of Chemical Engineers, 345 East 47th St. New York, N.Y. 10017, (1981).
11 Hancock, K.L., Spatial data and geographic information systems within the bureau of transportation
statistics (BTS), Vanderbilt University, prepared for the Center for Transportation Analysis, Oak Ridge
National Laboratory, and the Bureau of Transportation Statistics, U.S. Department of Transportation,
Washington, DC., (October 3,1994).
12 Vendor of TECHNICA products: DNV TECHNICA Ltd., Lynton Housde, 7-12 Tavistock Square, London
WCIH9LT, U.K.
13 Van Aerde, M.; Stewart, A.; Saccomanno, F., Estimating the impacts ofL.P.G. spills during transportation
accidents, J. Hazard. Mater. 20 (1988), pp. 375-392.
14 Abkowitz, M.; Cheng, P.D.M., Developing a risk/cost framework for routing truck movements of hazardous
materials, Accident Analysis and Prevention, 20 (1), (1988), pp. 39-51.
15 Polus, A.; Dagan, D., Models for Evaluating the Consistency of Highway Alignment. Transportation
Research Record 1122, TRB, Washington DC., (1987)
16 Lamm, R; Choeiri, E.M., Recommendations for Evaluating Horizontal Design Consistency Based on
Investigations in the State of New York. Transportation Research Record 1122, TRB, Washington DC.,
(1987).
11 Sivakumar, RA.; Batta, R, Establishing credible risks criteria for transporting extremely dangerous
hazardous materials, in Transportation of Danger{)us Goods: Assessing the risk, F. Saccomanno and K.
Cassidy (eds), Institute for Risk Research (lRR), (1993).
18 Sivakumar, RA.; Batta, R.; Karwan, M.H., A net-work based model for transporting extremely hazardous
materials, Working Paper, Department of Industrial engineering, State University of New York, Buffalo,
(1991).
19 Glickman, T.S.; Sherali, H.S., Catastrophic transportation accidents and hazardous materials routing
decisions, Probabilistic Safety Assessment and Management Conference, Beverly Hills, California,
(1991).
20 Stewart, A.M.; Van Aerde, M.; Shortreed, 1.H., Enhancements and updates to the RISKMOD risk analysis
model, J. Hazard. Mater. 25 (1990), pp. 107-119.
21 Birk, A.M.; Anderson, R J.; Coppens, A. J., A computer simulation of a derailment accident: Part I - Model
Basis. J. Hazard. Mater. 25 (1990), pp. 121-147.
22 Fedra, K., Advanced Decision-oriented Software for the Management of Hazardous Substances. Part II: A
Demonstration Prototype System. (Final Report to EURATOM CEClJCR, Ispra. Italy). Reprinted as CP-
86-10, (1986). International Institute for Applied Analysis, A-2361 Laxenburg, Austria.
23 Kleindorfer, P.R.; Vetshera, R., Risk-Cost Analysis Model for the Transportation of Hazardous Substances,
Final Report VoL I. October 1985, Ludwig Boltzmann Institute, Vienna.
24 Majchrzak. 1., DISCRET: An Interactive Package for Multicriteria Decision Making Problems, Report,
June 1985, System Research Institute, Warsaw, Poland.
2S Fedra, K.; Lassarre, S.; Weigkricht, E., Computer-Assisted Risk Asssessment of Dangerous Goods Trans-
portation Case Study for Haute Normandie. Final report submitted to the Institut National de Recherche
sur les Transports et leur Securite. International Institute for Applied System analysis. A-2361
Laxenburg, Austria.
26 TECHNICA: The SAFETI Package, Computer-based System for Risk Analysis of Process Plant. Vol. I-IV
and Appendices I-IV, (1984), Technica Ltd., Tavistock Sq., London.
CHAPTER 9
TOWARDS INTEGRA TED RISK ASSESSMENT AND SAFETY

MANAGEMENT AT REGIONAL LEVEL: - TRANSPORTATION OF
DANGEROUS GOODS
9.1. Hazard Assessment Considerations
9 .1.1. INVESTIGATING AND ASSESSING THE RISK
One objective of hazard assessment is to quantify the hazard and the risk of its
occurrence. Another is to reduce the risk.
Hazard Identification and Safety Audits

The tools which can be used to identify hazardous situations are summarized below:
• Checklists, applicable to management systems in general and to projects
throughout all its stages.
• Use of reactivity rating method for estimating physical and chemical properties
(including also reactivity, instability, and explosivity).
• Use of computer pmgrams to calculate equilibrium products and heat release of
decomposition and/or combustion reactions: for instance the code CHETA
(CHEmical Thermodynamics And energy Hazard evaluation), developed by
ASTM (Davis and Ake l , 1973), or the code from the NASA (Gordon and
McBride2 , 1971).
• Use of hazard indices: DOW index, MPPD (Maximum Probable Property
Damage index), Fire and Explosion index, TI (Toxicity Index), etc.
• HAZOP (Hazard and Operability Studies).
• Plant Safety Audits.
There are four situations with respect of the probability of an event:
• The probability of the event is low.
• The probability of the event can be made low by the application of a standard or
code of practice.
262 CHAPTER 9
• The probability of the event can be made low by the application of measures
which can be shown by a simple risk analysis to be of equivalent safety to the
normal standard or code of practice (equivalent risk).
• The probability of the event must be assessed quantitatively by use of (a) a Risk
Analysis, or (b) a Fault Tree (and eventually an Event Tree), and (c) Cause-
consequence diagram.
In some cases events with rather low probability may be neglected, while in others
they still make a significant contribution to the risk and must be taken into account.
A risk analysis may also be carried out to determine safety relative to an absolute
risk criterion such as FAR (Fatal Accident Rate). Whereas afault-tree is constructed by
defining a top event and then defining the cause events and the logical relations
between these cause events (top down approach). An event-tree begins with a particular
initial event and works from the bottom up.
• On the basis of HAZOP studies prepare an overall survey considering major
hazards such as large explosions or toxic release, and the events which can give
rise to them.
• Represent the situation revealed by this overall survey and its possible causes in
the form of a fault tree. The individual branches of this overall fault tree are then
considered and their probabilities assessed.
• On the basis of the hazard assessment and in relation to the risk criteria take
measures for the control of the hazard and for the reduction of the hazard risk.
One problem of risk assessment is that there is no method of ensuring that all
eve.1tualities have been taken into consideration. AnotLer problem is that the data
required for the assessment are often inadequate or non-exi!:tent.
9.1.2. ASSESSING THE CONSEQUENCES
Another important method of hazard assessment is the defiuition and modeling of

scenarios for loss of containment with potential for fire, explosion or toxic release. So-
called vulnerability models are used for that purpose. Some models deal with the
emission of material from containment and with dispersion of vapor into the
atmosphere, while other models are concerned with the evaporation and dilution of
spillage of material of different volatility under different conditions, or relate to the
flame on a pool of burning liquid. Some other models are capable of assessing damages
and/or resultant doses resulting from an accident. We indicate here two simple
equations relating the intensity as a function of the distance (a) for an explosion, and (b)
for toxic gas release.
a) For explosion
(9.1)
INTEGRATED RISK / SAFETY MANAGEMENT AT REGIONAL LEVEL 263
where pO is the peak overpressure; r the distance; and W the mass of explosive; n =
1.6 (for overpressures of 1 - 10 psi), respectively, 2.3 (for overpressures of 10 - 100
psi).
b) For toxic release
(9.2)
where Q is the rate of release; r the distance; and C the toxic gas concentration. The
value of n is about 1.76.
In many situations it is necessary to relate the intensity of some phenomenon such
as heat radiation from a fire, overpressure from an explosion or toxic gas concentration
from a toxic vapor release to the degree of injury or damage which can result from it. A
method currently used is Probit Analysis. The probit (or probability unit), Y, is related
to the probability, P, by the equation:
1
P=---J Y-5 (u )
exp - - du
2
(9.3)
(21t )1/
2 - 2
The probit is a random variable with a mean 5 and a variance 1. The probability
range (0 - 1) is usually replaced by a percentage (range 0 - 100). A general form of the
probit is: Y = kJ + k2 In V, where the probit, Y, is a measure of the percentage of the
vulnerable resource which sustains injury/damage, and the variable, V, is a measure of
the intensity (overpressure, etc.). Probit equations have been developed for all kinds of
hazards and for different toxic releases.
9.2. The Potential Use of Com.,arative Assessment of Various Transportation

Systems
Methodological key issues (Gheorghe3 , 1994) to ensure a credible comparison to be

made among different technologies or technological systems can be summarized as
follows:
a) The delineation of consistent and comparable boundaries for different transportation
technologies or chains of technologies (e.g., modal transportation systems). The
definition of principles and delineation of "reference technological chain" for risk
comparison purposes is a priority issue in this regard.
b) Methods and dose-effect relationship for the risk assessment of the health effects of
different technological transportation systems (e.g., rail, road, pipelines, ships,
barges) including, in particular, the deviation of consistent indicators of health and
environmental risks for comparative assessment.
c) Methods and relationships for the assessment of impacts of different transportation
technologies on the natural environment and associated ecosystems, and particularly
264 CHAPTER 9
the derivation and application of indicators of environmental risks for comparative

assessment.
d) How to deal, in the comparative risk assessment process, with the time and techno-
logical dependencies of risk estimation, given variations in technologies and in the
state of technological development (over time). Can the principles of "risk discount-
ing" be used to compare existing risks by postulating and accounting for possible
future technological development. How can future technological development be
accounted for in the comparative assessment process.
e) How to deal in the comparative risk assessment process with the issue of "uncertain-
ties". Such uncertainties exist in particular in the type, nature and extent of various
impacts and with the variable nature of some impacts over time. Uncertainties also
exist in the estimation (both qualitatively and quantitatively) of the various risks,
both in terms of consequences and probabilities.
t) Comparative risk assessment for severe accidents in different transportation techno-
logical systems. The main issues here relate to the tools and criteria used for
comparison; the methods for presenting the results and the comprehensiveness of
the comparative results, particularly in relation to environmental impacts from
severe accidents.
g) The integration of the different elements of risks, including the question as to
whether one overall integrated indicator of risk is possible or feasible; the applica-
bility of comparative risk assessment to regional safety management studies; the
weighting of the various dimensions of different risks in the comparative risk
assessment process.
h) Methu:!s and tools for the presentation of the results of comparative risk assessment
9.2.1. SETIING BOUNDARIES FOR COMPARATIVE RISK ASSESSMENT OF

DIFFERENT TRANSPORTATION TECHNOLOGIES FOR REGIONAL SAFETY
MANAGEMENT
The definition of consistent boundaries for the different transportation of dangerous

goods technologies which might operate in a given area is important for credible
comparative risk assessment results. When dealing with comparative risk assessment of
various technologies which might coexist in a given region, a consistent boundary
definition has to be chosen.4 When comparing technological or activity risks in a given
area there is obviously a need to agree on some boundaries.
The principles for setting system boundaries are as follows:
• The setting of boundaries has to be adapted to the purpose of the assessment
• The boundaries should be set taking into account:
time (horizon)
space (workers and public exposure, environment)
phases of the systems under investigation
impacts to be considered
material balance consideration in terms of risk
per net production (e.g., energy) output
acceptable threshold criteria.
• At a regional level, the setting of boundaries will greatly benefit from structured
public scoping/participation process, so that all concemed::ould be involved in
boundary setting. This process would greatly enhance the credibility and accep-
tance of the study process.
9.2.2. ASSESSMENT OF ENVIRONMENTAL RISKS (IMPACTS)
Usually in a region where complex risk assessment studies are made, one has to take in
consideration that a large number of compounds are released into the atmosphere and
water and disposed on land with associated environmental impacts (risks). These
impacts may be immediate or long term. Implicit "environmental impacts" are often a
direct or indirect "inference" of "health impacts". TABLE 9.1 and TABLE 9.2 show a
generalized environmental transfer model outlining the various essential components of
health and environmental risk estimations in large industrial complexes/regions.
Transportation of dangerous goods systems should be included and/or integrated into
such a comprehensive analysis.
TABLE 9.1. Categories of risks usually adopted to assess and compare the health impacts of different
pollutant sources
Health Risk
Source People at Risk Exposure Effects
Routine or accidents Workers and the Short, or medium and Fatal and non-fatal
public long term Immediate/delayed
respectively
Long-term/delayed
TABLE 9.2. Categories of risks usually adopted to assess and compare the
environmental impacts of different pollutant sources
Environmental Risk
Source Effects
Duration Extent
Routine or Short or medium and Local, regional and
Accidents long term global
A linear pollutant pathway modelS indicating the amount reaching the receptor
(target at risk) as a function of the amount emitted, altered by dilution and removal and
enhanced by environmental accumulation factors is shown in Figure 9.1.
266 CHAPTER 9
As shown, distance (space) and rate of movement (time) are critical parameters to an
environmental impact assessment study. The complex interdependencies between time,
space and feedback mechanisms (the degree of resilience of a given environment to
external factors) are not fully known; very often it is difficult to normalize them on a
common scale for comparison.
In comparative risk assessment of different technologies one has to "compare risks",
which may be different in a subjective way from impacts, effects, emissions, etc. Two
major limitations have to be considered when dealing with the assessment of environ-
mental impacts of different technologies or technological chains, namely:
- the effects are not always susceptible to quantification
- there is no general agreement on what should be quantified.
Methods of relevance for making comparisons of environmental impacts in
integrated regional risk assessment and safety management studies are:
ranked matrix environmental assessments
emission values and ambient quality indices
critical loads and critical levels.
- A matrix approach is a ranking method to quantification of environmental impacts. It

attempts to relate dose to response.
- The emission values method is a dose oriented approach.
- Critical load model whilst being dose based. is in fact essentially of a response type. the
response being the outset of damage systems.
- The Critical Loads concept provides an adequate scientific basis to set limits to
environmental impacts above which damage would occur.
Critical Levels. It expresses the concentration of pollutants in the atmosphere above

which direct adverse effects on receptors such as the ecosystem, may occur, according
to present knowledge.
Critical Loads. It is a quantitative estimate of an exposure to one or more pollutants

below which significant harmful effects on specified sensitive elements of the environ-
ment which do not occur, according to present knowledge.
The critical loads approach is a procedure for developing optimized abatement
strategies by which leveled emission reductions are obtained on the basis of scientifi-
cally derived critical values.
Additional concepts associated with the critical loads and levels are:
target levels and loads
- actual loads and levels.
INTEGRATED RISK AND SAFETY MANAGEMENT AT REGIONAL LEVEL
Rate of Effects
Raleof Rate of removal or
altha
emission transport accumula-
receptor
tion
Amount
Amount
reaching
emitted
receptor
over
over
unHtime
unHtime
Figure 9.1. Pollutant pathway model used for assessing environmental impacts from emissions.
N
~
268 CHAPTER 9
The critical loads approach for regional environmental impact assessment requires
the following types of information:
inventories of current emissions and projections of future emissions rates
estimates of the potential emissions reductions
long range transport models
maps of critical loads and target loads
integrated assessment modeling.
Indicators of environmental risks: Qualitative and quantitative indicators could be
used for environmental impact indicators of different technologies which might coexist
in a region (including the transportation of dangerous goods ). Such indicators are: site
specific, society specific, value judgment dependent. Synergistic effects could be taken
into consideration when dealing with comparative environmental impact assessment.
Environmental performance measurements already accepted by international orga-
nizations and practice and which can be used in comparative analysis are:
environmental performance indicators (e.g., river quality, air quality, soil quality
etc.)
- environmental goals (critical loads, sustainability index)
environmental emissions (SOx, NOx, CO2, etc.).
9.2.3. COMPARATIVE HEALTH RISK ASSESSMENT
The assessment of health risk is usually based on:

i) data from epidemiological studies for particular situations
ii) estimations based on dose-effect relationships for different levels of exposure.
Mathematical models enable the determination of ambient concentrations of
pollutants in different media, and the prediction of health effects through dose-effect
relationships. The state of knowledge concerning exposure paths, dose-effect relation-
ships and the health risks for different technologies varies amongst different systems.
Both approaches to estimate health effects from various technological systems
which can coexist in a given region, only rarely deliver quantitative risk coefficients
with high confidence intervals. Dose-effect relationships for many chemical pollutants
are not available.
There is a particular lack of data concerning the delayed (chronic) health effects of
many chemical pollutants. The health effects at the low spectrum of continuous
exposure need particular attention in this regard. Another issue of research need relates
to the indirect health effects of pollutants where relatively little information is available.
A short list of parameters, being used as indicators of health risks for comparative
assessment of various technological systems, includes:
number of fatalities (acute)
number of fatalities (delayed)
incidence of injury
INTEGRATED RISK I SAFETY MANAGEMENT AT REGIONAL LEVEL 269
- projected incidence of genetic effects

years oflost life (YOLL)
relative utility loss (RUL).
The following comments are offered in relation to the use of the above indicators:
• Although compound indices such as YOLL and RUL offer the prospect of improved
reliability in assessing and comparing health detriments, simple indicators such as
numbers of fatalities and serious injuries should continue to be used in parallel, at least
for the time being.
• Some classes of detriment should not be aggregated (e.g., prompt and delayed
fatalities ).
• Health risks to different groups and of different types should be distinguished,
although it may in addition be permissible to aggregate in some cases.
Time Dependence in Comparative Risk Analysis

In comparative risk assessment (CRA) , appropriate time frames must be defined for
different stages of technological systems. In this respect one can consider micro-time
and macro-time as possible concepts to be used in comparative risk assessment.
9.2.4. UNCERTAINTIES IN COMPARATIVE RISK ASSESSMENT
When dealing with comparative risk assessment for activities which take place in large
(regional) areas, including the transportation of dangerous goods, uncertainties occur in
the prediction of scenarios, models and data.
Uncertainties in scenarios
- erroneous probabilities
- factors not considered
- factors screened out.
Uncertainties in models
- imperfect conceptual model
- imperfect mathematical model
- imperfect computer model.
Uncertainty in data
- general vs. site specific data
- measurement errors
- data reduction.
Expert judgment is inherent in the evaluation of uncertainties. Uncertainties must be
delineated and exposed whenever appropriate and attempts made to deal with them.
Uncertainties take many fonns and it is essential that a coherent and clearly visible
approach is adopted, both in the computational process and in the interpretation of
results.
270 CHAPTER 9
Within the complex study process of integrated regional risk assessment and safety
management (including the transportation of dangerous goods), the uncertainties arise
from a number of sources:
Data. All data are subject to sampling errors. Statistical uncertainty requires the
reporting of range or confidence levels.
Models. Estimates of risks through scientifically based models (lack of direct observa-
tions) brings additional uncertainty in a comparative risk assessment. When models are
the only tool to be used for risk assessment, uncertainty analysis of the results should be
exposed and taken into consideration.
Knowledge. Lack of knowledge or uncertainty in the foundation of basic physical

phenomena can lead to erroneous results and faulty predictions. At the present time, this
uncertainty in results due to knowledge is solved by using simulation procedures or
inter-comparative computer exercises.
There are some techniques already in use which can help to evaluate uncertainties;
among them are sensitivity analysis, inferential statistics (Bayesian methods), complex
statistical and probabilistic analyses, Monte Carlo simulation. In case of lack of know-
ledge, bounding calculation and conservative assumptions can be used. As a rule of
thumb, one can consider that risk assessment should be realistic and, if this is not
possible, then the results should be considered as conservative.
There are different mathematical approaches to dealing with uncertainty: probabi-
lity theory, Markow models, possibility theory (fuzzy sets), event trees, influence
diagrams, heuristic models, !leural networks, etc. The fact that uncertainties exist does
not invalidate the use of comparative risk assessment as a useful tool for o,'crall
assessment of the impacts of various technologies in a given area, as long as such
uncertainties are clearly understood, exposed and acknowledged in the presentation of
the results of the study.
9.2.5. UNCERTAINTY ESTIMATION IN ENVIRONMENTAL REGIONAL RISK

ASSESSMENT AND THE TRANSPORTATION OF DANGEROUS GOODS
In risk estimate, uncertainty pervades each element of environmental risk assessments, as an

examination of estimations of the risks to humans from the release of a harmful substance
into the environment due to normal operation or accidents.
The mathematics in a quantitative risk estimate may enter the following elements:
Risk = Probability • Quantity • (llDispersion) • Population· Uptake * (Dose vs. Response)
It is clear that an increase in any of these six elements leads to higher estimates of
risk levels. There is always some uncertainty about each of these elements in the risk
estimation process. In the current discussion, the terms "extreme certainty" and
"extreme uncertainty" denote opposite ends of the spectrum of how well a particular
phenomenon is understood. Qualitative modifiers "high", "fair", and "moderate" depict
gradations of this understanding. These gradations, though arbitrary, are appropriate to
the range of uncertainties found in environmental risk assessment.
1) Uncertainties on the probability of a release of a harmful substance:
- May be unnecessary for many environmental risk assessment.

- Estimating the probability or frequency of such accidental releases is the first
step in many environmental risk assessment studies.
There is sizable uncertainty about estimates of the probability of accidental
releases of harmful substances (N.B., historical data offer some guide to this
probability for frequently occurring accidental releases, but are less information
on some kinds of accidents involving certain substances).
Mathematical models produce fairly uncertain estimates of the probability of
accidental releases of radioactive substances, it would appear that most estimates
of the probability of chemical releases would be no better than fairly uncertain.
2) Uncertainty on the quantity of a harmful substance released:
- Such estimates improves with the quantity of the information gathered on the
harmful substance released into the environment.
3) Dispersion of a harmful substance and uncertainties on the resulting concentrations
of that substance in the environment:
In practical situations estimates of concentrations at particular points in the
dispersion pathway range from fairly to highly uncertain.
- When measurements of concentrations are available, risk analyst can omit
estimation of the probability. of release of a given substance.
- Actual measurements of concentrations of particular substances in the environ-
ment reveal that concentrations are highly variable ambient concentration of
some pollutants are moderately invariable to moderately variable from day to
day in a particular season. These concentrations may vary by a factor of 5 to 10
between warm and cold seasons and among different years. Human exposure to
some chemicals in the workplace may be extremely variable among individual
work situations.
4) Uncertainty on the population exposed to release of a harmful substance:
In the absence of site-specific data, any estimate of the population potentially
exposed to the release of a harmful substance from such sites would be extre-
mely uncertain.
5) Uncertainties on the uptake of harmful substances by humans and other organisms:
- Pharmaco-kinetic models reflect how the physiology of humans differs from that
of test animals with respect to uptake, metabolism, and excretion of particular
chemicals. Pharmaco-kinetic models may modify risk estimates in significant
272 CHAPTER 9
ways, but their structure and the data they contain make these modifications
moderately to extremely uncertain.
6) Uncertainty on the relationship between dose of a harmful substance and adverse
toxicological response:
The toxicity of a particular dose of a substance not only varies across species,
but among individuals of the same species.
Most acute toxicity estimates are moderately uncertain.
Cancer potency estimates are fairly to extremely uncertain.
7) Measurement error:
Measurement error increases the uncertainty associated with each of the above
elements in an environmental and health risk assessment.
Among all the elements of risk assessment, measurement error generally intro-
duces the least uncertainty.
8) Overall uncertainty:
Uncertainty in each of the elements of a quantitative risk assessment combines to
produce an overall uncertainty in the estimate of a particular risk.
Only a few studies have carefully integrated the uncertainty associated with
individual elements of a risk estimation into an overall estimate of a risk and its
overall uncertainty.
Estimates of human cancer risks due to exposure to chemicals have been shown
to be moderately to extremely uncertain.
9) Credibility of Risk Analysis; Uncertainty on Policy Implications:
- Uncertainties in environmental and health risk assessment are present in all
elements of the assessment process. .
There are several approaches to deal with uncertainty:
i) Risk analysts should present the particular level of uncertainty in the risks they
are estimating.
ii) Risk managers should reduce uncertainty when the reductions would be cost-
effective and would affect the choice of risk management option.
iii) When conservative assumptions must be applied, this must be done in the
context of the particular uncertainties of the risk being addressed.
Reducing the uncertainty in anyone or more of the elements of a risk assessment
will reduce the overall uncertainty of a risk estimate; not all such reductions are equally
cost-effective. Risk managers can also use the economic concept of the efficient frontier to
determine which reductions in uncertainty are likely to reduce overall uncertainty the most
and at the lowest cost:
with regards to making conservative assumptions in the context of uncertainty,
risk mangers should consider when and how to apply these assumptions
some critics assert that conservative assumptions have resulted in grossly

inflated risk estimates and, thus, in regulations that are too stringent; others
assert that these assumptions are needed in order to compensate for various
uncertainties.
Some Guidelines on How to Deal with Uncertainty

• Action 1. Explicitly acknowledge and approximate the uncertainties in each element of
the risk estimation process regarding the transportation of dangerous goods.
• Action 2. Carry the uncertainty through the calculation process and calculate not only the
median or mean risk, but its overall uncertainty as well.
• Action 3. Develop and apply decision criteria that explicitly address acceptable risk levels
for individuals and the overall population.
• Action 4. Offer to the decision makers a description of the nature and the magnitude of
risks, including the uncertainty of these risks.
TABLE 9.3 presents some verbal description pertaining to the range of uncertainty.
TABLE 9.3. Range of uncertainty and variability in environmental risk assessments
Range ofUac:ertaintyl GSD Descriptive Interpretative Tenn Phrases

Variability
More than 10.00 Extremely uncertain Extremely variable
10.00 Highly uncertain Highly variable
5.00 Fairly uncertain Fairly variable
3.00 Moderately uncertain Model ,.t~ly variable
2.00 Moderately certain Mode[~.ely invariable
1.50 Fairly certain Fairly invariable

1.25 Highly certain Highly invariable
1.05 Extremely certain Extremely invariable
1.00 There i~ absolute certainty
9.2.6. INTEGRATED ENVIRONMENTAL AND HEALTH RISK ASSESSMENT IN

LARGE INDUSTRIAL AREAS
Uncertainty in Risk Analysis and Results Representation in Data Bases

There is a need to represent the uncertainty bound~ of such data and knowledge in
comprehensive studies and in specialized data bases. The purpose of this section is:
i) to help in identifying (direct) uncertainty measures of data and knowledge
specific to health and environmental risk assessment studies in large industrial
regions including activities related to the transportation of dangerous goods, and
ii) their representation in documented and specialized data bases.
274 CHAYfER9
No measured or calculated value related to environmental and health impact (risk)

assessment has any meaning unless it is accompanied by a declaration of quality, in
form of an uncertainty statement The uncertainty declaration(s) can be regarded as the
quality assurance tool for that process.
Uncertainties Associated with Model Predictions in Risk Assessment

Uncertainties in models may result from transformation of the given model into a
mathematical .1lodel. One has to distinguish between parameter, model and conceptual
uncertainties, and treat them separately. Often results from modeling activities are stored in
a data base (e.g., PC-FACfS as a specialized data base related to accidents, including those
of transportation of dangerous goods). Models for assessment of environmental, health and
safety impacts are important elements in integrated and comparative risk. assessment for
different technologies which might coexist in a given region.
A formal procedure of parameter uncertainty analysis is given next:
1) List all the parameters that are potentially important contributors to uncertainty in
the final model prediction.
2) For each parameter listed, specify the maximum conceivable range of possibly
applicable alternative values:
- Specify the degree of belief (in percentage) that the appropriate parameter value
is not larger than specific values selected from the range established above and
select a probability distribution that best fit the quoted degrees of belief.
- Account for dependencies among model parameters by introducing suitable
restrictions, by quoting appropriate degree of belief, or by specifying suitable
measure of the degree of association.
3) Set up a subjective probability density junction (pdf) for the combined range of
parameter values. This will subsequently be referred to as a joint pdf. Propagate this
joint pdf through the model to generate a subjective probability distribution of
predicted values:
Derive quantitative statements about the effect of parameter uncertainty on the
model prediction.
- Rank the parameters with respect to their contribution to the uncertainty in
model prediction.
- Present and interpret the results of the analysis.
Expert judgment elicitation is preferred when one of the following sit~tions exists:
- no other means are available for quantifying an important issue.
- the information available is characterized by high variability.
- some expert question the applicability of the available data.
- the existing results from code calculations need top be supplemented.
- the analysts need to determine the current state of knowledge.
Uncertainty Declarations in Databases

There are reasons for including uncertainty indicators in the databases which store
information relevant to integrated risk assessment or comparative risk assessment of various
technological systems. Uncertainty declarations are quality declarations and a good
description of uncertainties is decisive for good practice and scientific acceptance.
Uncertainty criteria may be used as acceptance criteria in quality assurance for
choosing data to be accepted and included in the data base. A good scientific practice is a
strategy for accepting material in the data base.
Validation. The advanced assumptions (theories) should be subject to a validation

procedure.
Criticizable. The assumption in risk assessment should be formulated in a well defined

manner so that it can be criticized.
Acceptability. The assumptions for risk estimation should be discussed when necessary
in a forum (e.g., steering committee).
9.2.7. INTEGRATION OF THE DIFFERENT ELEMENTS OF RISK AND THE

ROLE OF COMPARATIVE RISK ASSESSMENT FOR SAFETY MANAGEMENT
Lack of a single integrated risk indicator; the elements of risks from various technolo-
gical transportation systems can be broadly categorized as indicated in TABLE 9.3. In
reviewing risk assessment and management results for their use in area safety planning,
one should take in consideration the followings:
• The various elements and dimensions of environmental and health risks cannot be
integrated into one overall indicator of total risk. No comparison on the basis of a single
indicator is possible. The comparative risk assessment process must specify on which
basis (indicator) the comparison is being made.
• It is necessary to expose all the dimensions and elements of health and environmental
risks in thl! comparative process. Differences between transportation systems, regions
and societies make the development of one overall indicator meaningless. There is no
"global" risk value. The results should not be transferred from one study to another
without appropriate investigations of differences between technologies, regions and
countries.
• One approach to develop an overall indicator is the estimation of "extemal costs" of
impacts in terms of monetary value.
• The integrated approach to health and environmental risk means all risks should be
identified, assessed and considered in the comparative risk and management process.
276 CHAPTER 9
More Comments on Continuous Emissions

The most serious risks (impacts/effects) from continuous eDllSSlOns from different
industrial activities (e.g., power generation, chemical plants, heating, etc.) may be
expected to be those related to human health and of environmental effects.
A damage function approach can be used to assess and value the impacts resulting
from various activities - including the transportation of dangerous goods - which are
taking place in a large industrial area. This type of approach allows calculation of
marginal damages; many impacts are very often side and technological dependent. A
pathway is chosen to show the progression of impact from the industrial activity and
corresponding emission through effects on plants in natural ecosystems to economic
valuation.
In order to quantify different risks from a specific industrial site it is necessary to
identify the distribution of receptors that may be affected by operation of the plant in
some detail. The type of data necessary to make this calculations are of the following
form:
- Distribution of people in different risk categories
- Distribution of various species in different risk categories.
It is important that health and environmental damages are assessed over as much of
the area as possible. One has to be aware that the artificial truncation of the reference
environment for risk assessment will lead to underestimation of impacts.
Impact estimation can be performed on a marginal basis using at least three types of
procedures
• Statistical relationships
• Dose-response functions
• Mathematical models containing a series of cause-effect relationship.
International team of experts convened to discuss, agree and recommend the most
applicable dose response functions and models for risk assessment Valuation can be
performed most easily when the commodity is directly traded. When this is not
possible, it is necessary to assess values indirectly in terms of Willingness to Pay (WI'P)
for improved environmental quality, or Willingness to Accept (WI'A) increased environ-
mental damage.
External Costs. They are defined to be the costs which fallon one group of people
due to the social, or economic activities of another group, and where the latter group
does not take these costs into account. There are often difficulties in defining to what
extent damage to health or the environment is an external cost.
In order to be comprehensive, consistent and transparent it is necessary to consider
the environmental impacts and their monetarized values over all space and time (the life
cycle analysis). All part of the analysis are subject to uncertainty, some of which is very
considerable. In many cases it is not even possible to describe the uncertainty in terms
of distribution of values. The valuation of uncertain events (risk evaluation) poses
considerable challenges. Discounting of the monetary values of environmental and

health impacts is controversial.
TABLE 9.4. Indicators representing damage categories
Public health
- PM IO mortality - PM IO respiratory hospital admission
- S~ mortality - S~ cough episodes for children
- PM IO bronchitis in children - S~ adult chest discomfort days
- PM IO symptoms days - SO, eye irritation
- PM IO asthma attack - NO, eye irritation
- PM 10 restricted activity day - Transport accidents
- PM IO emergency room visits - Construction accidents
Occupational health
- Mining (respiratory diseases; mortality - Transport (death, major/minor injuries)
and morbidity)
- Mining (death, major/minor injuries) - On road accidents (death. major/minor
injuries).
Agriculture Aquatic impacts

- on wheat, barley, rye,
S~ -direct effects - Mining (groundwater and surface water)
oats, peas and beans, or other crops
- Acidic deposition -soil liming - Acid deposition -fisheries
- Effects of ozone Forests
Biodiversity
Materials
• Stone refacing - Galvanized steel replacemOent
- Mortar repointing - Aluminum replacement
- Concrete replacement - Cultural impacts
- Repainting - Building damage from subsidence
- Steel replacement
Noise Global warming impacts
• Plant operation • Aquatic life
- Road traffic - Air quality (ozone)
- Rail traffic - Glacier regression
- AgriculturallForest growth
- Desertification
Impact Categories
A list with guiding priority impact categories is given next:
atmospheric poUutionlhuman healthlmaterialslforests/cropslfreshwater ftsheries
accidents (continuous emissions) occupational and public
land use/natural ecosystems
278 CHAPTER 9
global warming
emissions to water/drinking water/ecosystem
noiselhuman health/amenity.
Damage Categories Indicators

Appropriate indicators for damage categories related to the above impact categories list
are given in TABLE 9.4.. They have to be considered as indicative, which can guide the
analyst in his effort to quantify consequences in the risk assessment stage.
9.3. Some Methodological Issues in CRA for Severe Accidents
The assessment and hence the comparative risk assessment (CRA) of different transpor-
tation technologies or appropriate technological systems as regard major accidents,
focused almost entirely on estimating acute fatalities to people from historical records
or using techniques of Probabilistic Safety Assessment (PSA) as a predictive tool.
Methodologies for the estimations of the late health effects to people and of the
environmental impacts from major accidents are limited, or in need of significant
development and applications. A defmition for severe accidents should encompass all
the elements of health, environmental risks and damage to plant, equipment, buildings
and should be expressed in terms of both the potential as well as the actual damage and
risk.
9.3.1. METHODS AND TOOLS OF COMPARATIVE RISK ASSESSMENT FOR

SEVERE ACCIDENTS
The comparison cannot be made on the basis of the consequence of such accidents in
isolation. The likelihood (or probability) of occurrence should also be taken into
account. Hence, estimation of the frequency of such accidents is relevant. Such
estimations necessitate reliable information on the past records of such accidents and
their effects and/or the application of probabilistic methods that predict the likelihood of
their future occurrence.
It is difficult to assess and compare the frequency and the health and environmental
damages caused by severe accidents because such data are not systematically collected
by a single national agency.
There are no data concerning in particular the delayed effects on health from severe
accidents for non-nuclear energy systems. All health effects in such cases are reported
in terms of immediate fatalities, with immediate injuries reported in a few cases. This
makes complete comparison difficult, since the total impact may be underestimated.
The ultimate long term environmental effects, particularly from severe accidents, are
difficult to establish; it may be difficult to establish whether the effect is irreversible or
whether a recoverable effect is possible.
Presentation of Results of Comparative Risk Assessment for Severe Accidents

In regional risk assessment and safety management studies, in presenting the compara-
tive risk assessment results for severe accidents, the following principles should be
adopted:
a) Risks from severe accidents should be presented and compared separately from
the risks resulting from routine operations.
b) Data based on historical (actual) occurrences should not be compared Ilirectly
with data on probabilistic predictions of likely future events.
c) Comparative results should not be presented in terms of aggregate risk in
isolation.
d) All dimensions and elements of environmental and health risks should be ex-
posed and presented.
The presentation of results in the form of F-N curves (frequency distribution) is a
useful presentation technique. However, the values usually presented without additional
information may not, by themselves, be sufficient since other dimensions of risks may .
not be exposed. Another method of presenting the results in tabular form should ensure
the completeness of all relevant dimensions of risk rather than relying on isolated
information (e.g., on fatalities).
9.4. Basic Aspects in Decision Aiding Techniques
It is already acknowledged that the process of initiating, promoting risk analysis and
implementing safety management studies for regional areas involves complex decisions
as well as the participation of a large number of actors (e.g., public, experts, systems
analysts, environmentalists, safety engineers, administrators, politicians etc.). The
process of integrating various aspects of risk such as environment, health, hazardous
installations, transportation systems, safety culture, management, involves decision
aiding techniques which are close to the management science field known as decision
analysis. 6
There are, in general, positive and negative aspects associated with decisions.
Indeed, many decisions are made intuitively by experts and do not use structured
processes or techniques. For many decision problems related to risk assessment and
safety management in large and complex industrial areas, the solutions and then
advantages and disadvantages may not be immediately apparent because of the
complexity of the issues involved. There is a need for systematic processes to be
followed that help structure thinking and analysis, and allow different viewpoints to be
taken into consideration. Structuring helps avoid inappropriate ad hoc decisions and
allows the process of reaching a decision to be more open and the decision itself to be
more readily defensible (decisions which are taken today very often have long term
effects). In the end, the use of various decision aiding techniques and the overall
process and technology of decision analysis allows ·the integration of various risk at
regional and area level. The integration of various risks into the decision making
280 CHAPTER 9
process is the appropriate mechanism which allows displaying various risks and
choosing the most appropriate safety improvement strategy.
There are many inputs, influences and constraints that a decision maker will
consider when deciding whicb actions to initiate regarding risk reduction or safety
management to a particular plant or for the region under analysis. Decision aiding
techniques (DAT) are tools for decision makers; they are decision aiding techniques,
but not decision making techniques. In the content of regional safety management such
techniques could i) assist structuring the complexity of issues derived from hazard
identification and prioritization, ii) assist in improving safety of individual installations,
iii) siting of new installations to comply with numerous criteria and performance risk
indicators, etc. There are complex decisions that need to be made in the field of regional
safety management which may involve conflicting positions. A large number of tools
are available to assist in solving and structuring decisions of such complexity.
The main purpose of this section is:
- to emphasize and strengthen the use of DAT for integrated regional risk
assessment and safety management, including the transportation of dangerous
goods.
to provide guidance as to where such processes mayor may not be appropriate
(to avoid the application of sophisticated processes to trivial problems).
- to recognize that DATs are tools, not panaceas.
The main stages of DAT in relation to the integrative process of various risks which
could coexist in a given region, are:
.~: Definition and description of the problem.
.~. Consideration and defuntion of appropriate quality ~surance requirements.
• ~: Formalizing the descriptive model of the problem.
.~: Obtaining the necessary information for modeling.
• Step 5: Analyze in order to determine the set of alternatives and criteria.
• Step 6: Ensure the selection of the proper method to make the decision regarding the
proper integration of various risks in the region and their minimization
• Step 7: Establish a clear record of the process and any decisions taken as a result of
the integration process of various types of risk and appropriate safety
measurements.
To make a decision means to select a way of action out of a set of possible alterna-
tives. The decision process is complex and sometimes iterative; the set of alternatives or
criteria may vary from an interaction to iteration.
Decision making involves three major elements:
a) alternatives among which the ''best'' one will be chosen
b) criteria for judgment
c) methods for selecting one alternative from the whole set.

Decision aiding process (DAP) is needed because it helps to generate a degree of
shared understanding among interest parties who are concerned with issues on risk
assessment and safety management in regions with complex and large industrial
installations and transportation systems. There are a few issues to be highlighted: in this
domain one has to involve complexity, uncertainty or even fuzziness, multiple criteria
and group interests with the objectives in conflict. DAP can provide models that
integrate these features into an adequate methodology. DAP provide a framework
within which informed discussions of key issues (e.g., environmental or health risks)
can be conducted, and facilitate generation of consensus and integration of such issues.
The relative costs and benefits to the user of applying a decision aiding process
should be compared to the costs and benefits of not using one of these decision aiding
processes. The cost of not using a decision aiding process can be high, especially in
terms of inefficient use of human resources and in the end the incapacity of integrating
various types of risks and selective safety improvement measures in the overall decision
process.
In the risk management phase for studies concerning complex industrial areas, all
possible options (e.g., technological, organizational, managerial etc.), have to be listed
exhaustively as well as all relevant choice factor from which the appropriate
comparison of the set of options will be made. After having identified the sets of
options and factors, each option has to be assessed and weighted according to each
factor. These assessments involve generally uncertainties, especially when probabilistic
events are concerned for which occurrence probabilities and possible associated
consequences cannot be easily evaluated.
When the best option has been selected, one has to check the robustness of this
choice in relation with the variety of ~isk impact indicators, safety improvement
measures and a sound cost effectiveness strategy. Various uncertainties or value
judgments have to be tested in a sensitivity analysis phase. At the end of this selecting
process, results have to be presented as simple and as clear as possible.
The following steps have to be considering when involving DAP for integrated
regional risk assessment and safety management:
Choice of options and relevant technical, environmental, health and economical
factors.
Assessment of options according to each relevant factor and the relative
importance of each factor in the political or technical decision environment.
Choice of the most appropriate and relevant decision aiding techniques.
Estimation of weighting factors.
Use of sensitivity analysis.
Presentation and interpretation of results.
The wide practice of the use of decision analysis shows that no decision aiding
technique is appropriate for all circumstances.
282 CHAPTER 9
Decisions related to risk management in large (regional) areas involve negotiations

between people with different viewpoints or value judgments, such as representatives of
the cantons, health officials, the environment departments, representatives of the
workers and the management of a given facility, and the public. One objective of the
negotiations is to determine the "choice factors" which are representative of the
different viewpoints.
The range of possible decisions in any investigated area is generally limited by
certain constraints which may be imposed, for example, by a regulatory authority at the
federal or cantonal level, by operating policies, or by generally accepted good practice.
Other constraints may be less tangible, such as acceptance by the public or by the
industry practice, limited perceived threats to environmental quality and health risks,
and political conditions. New techniques are now available (decision conferencing)
which enable the integration of various aspects into a comprehensive decision policy.
9.4.1. OVERVIEW OF DECISION AIDING TECHNIQUES
Before carrying out any decision aiding process for integrated risk assessment and
safety management at regional level, it is essential to identify all of the available options
and relevant factors that would influence the outcome. "Brainstorming" can present a
constructive approach. The use of multi-disciplinary scientific and engineering
expertise groups should be considered for dealing with the more complex assessment
problems, particularly those involving the need to choose between aVailability of
alternative or competing action pathways.
Generally, one option is devoted to a given objective function (e.g., reduction of the
probability of occurrence of an undesired event, reduction of various releases, etc.). One
can thus structure a brainstorming session as follows:
What are the involved functions?
- Is each function necessary?
- It is possible to ensure a given function by different ways?
- For each way, is it possible to ensure the function differently?
The selection of the relevant factors can be aided by the application of analytical
hierarchy process or value trees which enable the complexity of a given problem to be
broken down into smaller constituent parts. As an example, TABLE 9.1 and TABLE
9.2 illustrate the criteria hierarchy for the overall health risks. As shown, distance
(space) and rate of movement (time) are critical parameters to an environmental impact
assessment study. The complex interdependencies between time, space and feedback
mechanisms (the degree of resilience of a given environment to external factors) are not
fully known; very often it is difficult to normalize them on a common scale for
comparison.
In comparative risk assessment of different technologies one has to "compare risks",
which may be different in a subjective way from impacts, effects, emissions, etc. Two
major limitations have to be considered when dealing with the assessment of
environmental impacts of different technologies or technological chains, namely:
- the effects are not always susceptible to quantification

- there is no general agreement on what should be quantified.
Methods of relevance for making comparisons of environmental impacts in inte-
grated regional risk assessment and safety management studies are:
- ranked matrix environmental assessments
- emission values and ambient quality indices
- critical loads and critical levels
- objective (e.g., routine, accidental, public, occupational etc.).
When different groups are directly involved in the decision making process, it is
important to take into account the different objectives of each group. In this context it is
necessary to recognize the fundamental objectives and the boundary conditions,
together with any particular preferences with respect to these criteria. For instance, for
the objective "minimize hazardous potential" the associated criterion is "hazardous
potential" and the preference is for minimization.
When a particular option cannot be uniquely prescribed by a single figure within
acceptable confidence limits, it may be possible to use alternative values or distribution
of values, or distribution functions, or even possibility distributions or fuzzy sets
evaluations.
9.4.2. DECISION AIDING TECHNIQUES IN USE FOR INTEGRATED RISK

ASSESSMENT AND SAFETY MANAGEMENT
A number of decision-aiding techniques are currently available. For the purpose of this
book, the application of Cost-Effectiveness Analysis (CEA), Cost-Benefit Analysis
i.CBA), Multi-Attribute Utility Technique (MAUT), and Multi-Criteria Outranking
Technique (MCOT) are considered in further details.
Cost Effectiveness Analysis (CEA)

CEA technique can be applied when two related factors are considered, one of which
being a cost. The relative interest of each option is assessed through a "cost effective-
ness ratio" illustrating the amount of effort (cost) devoted to a decreasing of one unit of
risk. This technique is relatively easy to apply and can be recommended when only two
relevant factors are envisaged because in this case the sensitivity analysis is very easy to
make.
Cost Benefit Analysis (CBA)

CBA originates from the economic theory of welfare; it compares the benefits and harm
associated with different options. The principal characteristic of this technique is that all
relevant factors have to be expressed in monetary terms. When all these factors are
expressed in the same unit it is very easy to aggregate all these different costs in a total
cost. The best option is then the option presenting the minimum total cost.
284 CHAPTER 9
Multi-attribute Utility Technique (MAUT)

The essence of this technique is to derme a scoring scheme (or a multi-attribute utility
function), measured on a scale between 0 and I, with the property that if the score (or
utility) is the same for two options there is no preference for one or the other. If
however the utility for option i exceeds that for option j, then option i is preferred to
optionj. Such utility functions are established for each relevant factor (or attribute) and
then aggregate in a total utility function representing the global interest for each option.
The MAUT is an interesting decision analysis model for high level strategy
problems involving different factors (quantitatively or qualitatively). Recent experi-
ments show that this technique can be implemented in practice when dealing with
different actors involved in a complex decision making environment. Fuzzy sets theory
can also be adopted to this technique in order to accommodate for imprecision in
evaluating of the factors.
Multi-Criteria Outranking Technique (MCOT)

The aggregation techniques described in the previous sections combine aU the evalua-
tions according to the different relevant factors into a single figure of merit expressing
the global interest of each option. Instead of expressing the performances of each option
in terms of a single overall figure of merit, outranking techniques compare each option i
to every option j in order to evaluate whether option i outranks (or is preferable to
optionj.
9.4.3. SENSITIVITY ANALYSIS
Any decision aiding procedure involves the use of data, models, techniques, value
judgments which contain uncertainties of various types such as:
Uncertainties associated with an imperfect knowledge of the performances of the
options under different circumstances (scenarios) or of the parameters and data
used in the assessments.
- Uncertainties associated with the use of models.
- Uncertainties due to imperfect knowledge (e.g., about the future of each option).
Intrinsic uncertainties resulting from the statistical treatment of the variable (this
is very important when assessing the expected outcomes from low probability
events).
- Value judgments.
The sensitivity analysis highlights the aspects which have the greatest influence on
the results, where it may be desirable to attempt to reduce the uncertainties, if possible.
Types of sensitivity analysis involve the following process:
- modification of a given parameter.
assignment of probability distributions to the important variables (probability
encoding).
9.4.4. PRESENTATION OF THE RESULTS FROM APPLYING A DECISION

AIDING TECHNIQUE
Aspects to consider when presenting the results of integrated decision for risk
assessment and safety management at regional level are as follows:
List of the assumptions, hypothesis and initial conditions considered within the
case study
Comment on the weighting factors
Comment and reference on the various models and techniques used as well as
their relevance and integration within the overall study
- Indication of the ~ uncertainties, quality of data
Presentation of the sensitivity analysis results.
From these different information the final decision will be made by the decision
maker or the group involved in preparing recommendations and policy strategy at area
level.
9.4.5. ADVANCED TECHNIQUES FOR DAP
Artificial intelligence models have been lately implemented for computer aided design
for risk assessment and management, as well as for expert systems development and
implementation. A few comments are worth considering:
a) Expert systems have been developed for assisting the probabilistic safety
analysis for complex nuclear and other industrial facilities and processes; they
are relevant in integrating various aspects concerning risk in a given region.
b) Decision support systems by means of expert systems, operation research, multi-
criteria decision models with objectives in conflict are already in use in making
decisions concerning risk reduction and cost optimization for marginal safety
measures implementation.
c) Connectionist expert systems (neural networks)7 which exhibit characteristics
and capabilities not provided by any other technology; this a complementary
technique designed to assist solving ill-designed risk assessment and manage-
ment problems.
d) Integrated knowledge based decision support systems by using systems analysis
techniques and procedures and by integrating in an efficient way expert systems
and neural network technology into an advanced methodology of systems
analysis and information technology.
e) Influence diagrams which enables capturing numerous elements which could
lead to an unsafe operational environment
286 CHAPTER 9
Decision Conferencing
Decision Conferencing is an efficient tool for integrated regional health and environ-
mental risk assessment and safety management. Decision Conferencing is a relatively
new decision aiding tool, which has been developed from the practice and need of
integrating various aspects of risk, which has grown up in a number of organizations or
projects, when complex situations and difficult issues arise. In a recent report
("International Chernobyl Project - input from the Commission of the European
Communities to the Evaluation of the Relocation Policy adopted by the former Soviet
Unions) it is highlighted the followings:
"Rather than discuss the matter amid the hurly-burly of their day to day
activities, the decision makers take time to go away from their regular working
place in order to concentrate fully and solely on the issues before them.
Sometimes straightforward discussion can lead to a clear decision and a view of
a way forward, but the complexity of the issues and the uncertainties involved
may be too great for simple discussion to resolve. In such cases, it has become
the practice in some organizations for the team responsible for the decision to
meet together for two days or more away from their nonnal working
environment to discuss and explorer the issues".
Decision conferencing uses the services of a moderator, who is skilled in the process
of group decision. The moderator is seldom a person with experience in the context of
the issues at hand or have expertise in the discussion. The moderator has a very definite
role, namely to smooth the team's work:, to help the process and make the team
productive and more creative. Within a decision conference event, the content of the
discussion comes from the team themselves. A decision conference is a two-day event
in which all the "owners" of the problem (e.g., integrating health and environmental
risks in a given region and making appropriate decisions for safety improvements)
gather together to agree upon a strategy. The decision makers, or advisors (the problem
owners) are supported by a moderator and a decision analyst. Two important comments
are made:
i) the moderator has the role to lead the conference (he acts as a disinterested
chairperson guiding the discussion forward constructively, building and interpreting
decision models): this would help the decision makers or their advisors to appreciate
various facets of the problem before them.
ii) the analyst deals with the details of the model building, by using information
technology (models and computers).
The technology of decision conferencing brings together knowledge, skills and
techniques arising from the following fields:
• Decision analysis which provides a structure and language (e.g., multi-attribute value
theory) in which decision makers can think and talk about their problem.
• Information technology assists in "real time" the decision making process; the results
of decision analysis can be presented to the decision makers in simple, easily
comprehended ways.
INTEGRAlED RISK I SAFETY MANAGEMENT AT REGIONAL LEVEL 287
• Group dynamics an awareness of which helps the moderator and analyst ensure that
the group of decision makers or advisors interact constructively.
• Within a decision conference the team can discover the importance of differences in
opinion between the team members and the sensitivity of their conclusions to these
differences and to those judgments of which they are most unsure. Controversial
discussions are unimportant when discussion is focusing on the critical issues without
being deflected. The moderator is expert in group dynamics and rational decision
theory, as well as a skilled communicator.
In practice, each decision conference is different. It evolves according to the needs
of the team and not according to some fixed agenda. The topic of integrated regional
risk assessment and safety management and its complexity is an ideal topic for decision
conferencing. No model should be taken as definitive: revised models are build and
much sensitivity analysis undertaken until no further insights are obtained and agreed
conclusions are reached. Decision conferences are creative events; they create an
environment where participants create, evaluate, modify and re-evaluate options,
building a strategy which they all support.
In accortiance with the practice in this field: "Surprisingly, the analysis in
decision conferences needs much less hard data than one would, at first, think.
Strategies have to be priced: that is clear. But the cost estimation need only be rough. It
is a broad brush picture that the event seeks to create. Detail can be added at a later
date".
The lack of a single risk indicator to deal with risk assessment, health and environ-
mental impacts at the regional level, should be compensated by the use of this
specialized model - decision conferencing - able to integrate objective and subjective
decision aspects when dealing with complex issues and costly options and decisions.
9.5. Loss Prevention and its Economic Aspects
9.5.1. CAUSEOFLOSSES
When considering the different aspects of regional risk assessment, one must keep in
mind that part of the risk is represented by fixed installations, like factories, chemical
processing industry, warehouses, marshaling yards, loading/unloading stations (for
hazardous products in liquid form), etc., and by the means of transportation related to
hazardous materials (trucks, railroad, ship, barge, pipeline).
A major hazard in modem process plant and in transportation usually materializes
due to loss of containment, resulting in leakage of hazardous liquids or escape of toxic
or flammable gases.
The big three major hazards are fire, explosion and toxic release. Some factors
which determine the scale of the hazard are:
• inventory
288 CHAPTER 9
• energy factor
• time factor
• intensity - distance relations
• exposure factors (reduction of exposure; mitigating effect).
TIle control of hazards is possible only through effective management Some
features which are characteristic of loss prevention are:
• development of techniques for the identification of hazards
• principle of risk criteria and of the quantification of hazards
• development of techniques for the quantification of hazards
• use of the techniques of reliability engineering
• principle of independence in critical assessment and inspections
• critical review and critique of codes and standards or regulations, where these
appear outdated by technological changes.
Planing for emergencies is also an aspect of importance in loss prevention.
9.5.2. ECONOMICS OF LOSS PREVENTION
Loss prevention is concerned with the avoidance both of personal injury and of eco-
nomic loss. Some costs related to loss prevention arise through failure to take proper
loss prevention measures; others are incurred through uninformed and unnecessarily
expensive measures. The financial viability of new project is often determined by loss
prevention factors.
Cost of Losses
The economic factors affected by accidents include loss of human life, property, insur-
ance, cost of being absent from work, plant shutdown and downtime, loss of marlrets,
etc.
The Concept of "Total Loss Control". The relation between accident and propertl
damage costs is the central concept of "total loss control" (Fletcher and Douglas ,
1971). Here the basic idea is that loss due to personal accidents is only the tip of the
iceberg of the full loss arising from accidents. Indeed accidents can be classified as
disabling injury, minor injury, property damage and no damage accidents. A disabling
injury being one which involves one or more days off work other than the day of
accident. The ratio between the different types of accident is a key concept in loss
control. A ratio for these accidents can be given in the form of an "accident pyramid":
1 (disabling injury) : 100 (minor injury) : 500 (property damage) : 10,000 (no damage).
Another ratio is the ratio of uninsured to insured costs of accidents.

It is suggested that the competent authorities responsible of a damage control center
maintain a data base concerning not only injury accidents but also damage accidents
and the cost which were involved.
The modeling of accident process is important and has been studied by Houston
(1971,1977), who proposes the following tools:
• Principal cause model ("proximate cause" used by lawyers and insurers): not
well suited.
• Fault tree model
The following statistics are needed for risk assessment and estimation of the cost of
losses:
- Loss statistics (fire and explosion). Explosion losses are usually included in the
fire statistics.
- Personal accident statistics classified by the type of manufacturing industries
(e.g., injury or deathsl1()8 exposed hours), and the corresponding rates/ratio
should be established as a mean of comparing different types of activities.
Cost of Losses at the National Level. The necessary information can be extracted from:
Statistics for fire loss in factories.
Surveys regarding the national costs of machine repairs and maintenance, both in
terms of direct maintenance costs and of lost production.
The cost of downtime in the process industries.
Information regarding other costs, such as those of design and commissioning,
delays, loss of markets are more difficult to obtain.
Cost of Losses at Factory Level. These can be estimated along the following lines:
• Damage cost may be estimated at the factory level by the method of "total loss
control" 9. Plant unavailabliity or downtime, mayor may not incur serious cost
penalties. There exist methods of estimating both the mean and spread of plant un&vail-
abilities.
• The effect of downtime on plant profitability is another standard topic in process
economics.
• Data on equipment failure are available as generic failure data given in the literature
or as data obtained from the works. Similar considerations apply to equipment repair
times and costs.
• The estimation of the cost due to public reaction is very much a matter of judgment.
This involves also the costs due to legislation changes or changes in the practice of a
particular branch of the industry under public pressure.
• Insurance costs need to be considered both in terms of the rates which are currently
quoted and of those which might be obtained by adoption of further loss prevention
measures.
Loss prevention is in large part the application of probabilistic methods to the
problems of failure in the process industries. Reliability engineering in the process
industries, and in transportation and storage, is especially concerned with mechanical
equipment reliability. Reliability may be defined as "the probability that an item will
290 CHAPTER 9
perform a required function under stated conditions for a stated period of time". We
indicate below some definitions concerning reliability relations:
Probability of unions
Joint and marginal probability
Conditional probability
Independence and conditional independence
Bayes theorem
- The reliability function and hazard rate
- The failure density and failure distribution functions
Relation between basic functions
- . Exponential distribution
Probability and event rate
- Unreliability and failure rate
- Mean life
There is a close link between reliability and qUality control of equipment, welds,
piping and instrumentation. Maintenance and operability controls taking place at
regular interval is an asset for good performance of equipment and is an integral part of
loss prevention.
9.6. Risk Communication
There is a growing awareness that rational risk management (based on estimates of the
probabilities and consequences of an accident) is only a partial solution to the problem
associated with the risk from the transportation of dangerous goods. Engineering
technology may effectively deal with the problem of reducing real risks; it very often
neglects the perceived or subjective risks. In terms of public response, perceptions are
more instrumental than objective risks. The relationship between ,,real" and perceived
risk is an imperfect one. When saying that risk perception is irrational vis-a-vis
objective risks, one does not imply that there is a lack of logic or pattern to the process.
A number of concerns relating to public perceptions of risk, risk tolerance and risk
acceptance could be considered from the perspective of communication strategies.
When dealing with risk communication programs within a decision making process,
one has to consider the followings:
- Public concerns and priorities
Knowledge and images
- Risk acceptance
Individual and community context.
Public Concerns and Priorities. The first issue is to understand what is bothering
people most about the transport of dangerous goods. One can make the assumption that
a whole range of factors, quite apart from "true" risk, will determine perceptual
response, and ultimately public acceptance. For the purpose of strategic planning, the
perceived risk is every bit as real and significant as the objective risk.
Public concerns and priorities could be considered from the following prospective:
The relative concern paid to the issue of the transportation of dangerous goods
vs. other threats to health and safety in the investigated area.
Perceptions of the relative risks posed by different modes of transport or types of
goods.
Levels of concern over various consequences of the transport of dangerous
goods.
A number of questions should be considered when dealing with the heading of
public concerns and priorities:
How prevalent and profound is the concern about the transport of dangerous
goods ?
How concerned are people about this issue relative to other threats to health and
safety ?
Is transport by rail more dangerous than by other modes ?
- Are nuclear wastes perceived as more risky than chemical goods ?
- Are people more concerned about risks to health and safety, or the impact of the
transport of dangerous goods on property values and the quality of their living
environment ?
Do these concerns vary systematically according to the background character-
istics of the respondent?
Knowledge and Images. A number of tasks should be addressed:

- Document any systematic or random distortions that occur in the public
perception process
Explain why these occur
- Indicate what implications these discrepancies between public perceptions of
risk and "objective" risks suggest for a rational risk management strategy.
Public awareness of hazards is a dynamic process, rather than static. Definitions of
risk tend to be arbitrary, reflecting the way in which people orient themselves to the
world. A number of questions related to the above issue could be addressed, namely:
- How aware are people about the transport of dangerous goods ?
How accurate are public perceptions of the risks involved ?
How much experience have respondents had with dangerous goods or their
transport ?
Do levels of knowledge vary with other respondent characteristics?
Public Acceptance. There is a number of dimensions related to risk acceptance:

- evaluations and perceptions of acceptable levels of risk
292 CHAPfER9
importance of public perceptions regarding the parameters of control and

responsibility
- the credibility of government and industry
public preferences among options.
When dealing with public risk acceptance attitudes and preferences, a number of
questions should be addressed:
- Does the public accept the notion that some minimal level of risk is inevitable?
How satisfied is the public with existing safety procedures and emergency
response measures ?
What is the minimum level of intervention necessary to make the public feel
. secure and thus accept the transport of dangerous goods ?
What is the trade-off between level of acceptable risk and the cost of improving
the safety of the system ?
- What faith does the public have in the competence and integrity of those
responsible for transport safety ?
What factors or combinations of factors, reduce or enhance public acceptance ?
Issues related to risk communication and public acceptance within the decision
making process are presented next by various categories summarized above:
i) Public Concerns and Priorities: level of concern (perceived level of risk from
living near dangerous goods transportation corridors), relative concern (perceived
level of risk from other situations which threaten health and safety), variation by
mode of transport (likelihood of accidents occurring by mode of transport, overall
level of concern about accidents involving dangerous goods by mode of transport),
variation by type of good (perceived likelihood of transport accidents and levels of
concern by type of product), relative importance of impacts (level of concern by
type of impact: accident, noise, property values, amenities), variation by respondent
(background characteristics and locallproximity of respondent).
ii) Public Knowledge: awareness of transportation of dangerous goods (accuracy of
perceived frequency of transport accidents involving various types of dangerous
goods, accuracy of perceptions of proportion of volume of freight involving
dangerous goods), fidelity of knowledge related to mode of transport, source of
knowledge (direct experience with accidents, information sources, work experience
with dangerous goods).
iii) Risk Acceptance: acceptance of inevitable risks, risk tolerance, minimum
intervention, trade-off costs for improvements to system, integrity of government
and industry, role of government, explanatory factors (background characteristics of
respondent, location I proximity of respondent, perceived level of personal control).
iv) Background Characteristics: location I proximity, life cycle, socio - economic
status (level of educational attainment, household income, homeowner I renter).
Summary
Integrating various types of risks, and their sources, existing in a region is not an easy
or trivial task. Various types of techniques, methods and tools could be employed in
helping the integration process. One potential I relevant instrument for the integration
process of various types of risks in a region is that of Comparative Risk Assessment.
The chapter investigates the process of setting boundaries for comparative risk assess-
ment of different technologies, the assessment of environmental risks/impacts, com-
parative health risk assessment, uncertainties in comparative risk assessment. Methods
and tools of comparative risk assessment for severe accidents are also discussed.
Recently, Decision Aiding Techniques playa significant role in the integration
process of various types of risks and their management at the regional level. This
chapter reviews such techniques and their relevance to the risk integration process.
Special attention is given to decision analysis and decision conferencing.
294 CHAPfER9
I Davis, E. I.; Ake, 1. A., Equilibrium thermochemistry computer programs as predictors of energy hazard
potential. In Loss Prevention, 7 (1973), pp. 28.
2 Gordon, S.; McBride, B. I., Computer program for calculation of complex chemical equilibrium
compositions, Report NASA SP-273, (1971).
3 Gheorghe, A.V., Comparative risk assessment of the health and environmental impacts of various energy
systems, In!. I. of Environment and Pollution, 4(3-4), (1994), pp. 329-349.
4 Haddad, S.; Gheorghe, A.V., Issues in comparative risk assessemnt of different energy sources, In!. I. of
Global Energy Issues, 4(3), (1992).
S United Nations Environment Programme, The Environmental Impacts of Production and Use of Energy.
Part IV. The Comparative Assessment of the Environmental Impacts of Energy Sources, Phase L
Comparative data on the emissions, residuals and health hazards of energy sources, Report ERS-I4-85,
UNEP, Nairobi (1985).
6 Gheorghe, A. V., Decision Processes in Dynamic Probabilistic Systems, K1uwer Academic Publishers,
Dordrechl, (1990).
7 Gheorghe, A.V., Connectionist expert systems for analysis of health problems associated with industrial
activity and electricity generation systems, In!. I. Environment and Pollution, 4 (1994), pp. 107-124.
• International Chernobyl Project - Input from the Commission of the European Communities to the
Evaluation of the Relocation Policy Adopted by the Former Soviet Union", Report EUR 14543 EN,
Commission of the European Communities, (1992).
• Fletcher, I.A.; Douglas, H.M., Total loss control, Assoc. Bus. Programmes, London, (1971).
CHAPTER 10
CASE STUDIES IN REVIEW AND MISCELLANEOUS INFORMATION
The information presented up to now gives a general overview of the situation on the
transportation of dangerous goods. This includes methodology, statistical data for
probabilistic risk assessment, and other information of interest. There has been different
approaches concerning risk analysis, which were followed by the authorities or
engineering companies in different countries. In this chapter we intend to present a
short summary of selected case studies and miscellaneous information of interest to the
reader, which may show different aspects of the methodologies in use and how a
potential user might implement its own applications. The topics presented below are
intended as examples, and in view of the restricted amount of space, this selection had
to remain very limited, and is in this sense not exclusive.i
10.1. The Vancouver Case Study
10.1.1. THE MODEL
In the Vancouver transport risk analysis model, ideas about the evaluation of risk are
organized around two underlying themes: probability and consequences. Comparing
transport routes in order to select the way presenting the least risk, or to focu~; remedial
actions on road segments which are known to be dangerous, is an important task. The
general philosophy followed by the Canadians is highlighted below.
The Government of Canada has sponsored a "transport of dangerous goods study" for the
Vancouver area, and the development of a risk analysis model. The principal issues of this
interesting report are summarized below.
Probability of Transport Accident

In the study of traffic accident the probability can be expressed as the number of
accidents events taking place over some measured time interval. In the study of traffic
accidents, this rate is conventionally made very general by scaling it according to the
length of the roadway (or railway) over which the accidents are recorded, and by the
number of vehicles present on the roadway over the course of the time interval. This
, We wish to apologize in advance to the reader and to any companies, which may feel that a particular
survey should have been included at this place. We also want to thank the Cantonal Authorities and the
private companies, who supplied us with the requested information and allowed us to disclose it in the
present book.
296 CHAPTER 10
results in a rate which is independent of the links road length or the actual number of
vehicles using it. The expected accident rate, ~,is expressed as:
Number of Accidents j
P = . length. (l0.1)
I vehicle j .length j I
The subscript, i, is an index number used to denote a unique value for a specific
section of a roadway.
To estimate the actual number of accidents over an interval of time, it is necessary to
specify the traffic volume.
Defining the Consequences from Transport Accident

Consequences are introduced in a way of trying to describe the actual effect of an
accident involving a dangerous commodity. It is advisable to restrict the analysis to the
examination of stereotype accidents involving the complete release of a "standard
volume" of each commodity (worst case approach). It is then possible to make
comparison between the relative risk among a group of routes.
This internal consistency means that if route A had a risk of 1000 (the units are not
important) and route B had a risk of 200, then the risk associated with route B would be
only 20% of route A. This scaling relationship (20%) will hold regardless of the
probability density function (PDF) underlying accident severity if the PDF's of selected
routes are identical.
It was found that such distributions are not sensitive to road-type per se, but depend
more upon the packaging technology associated with the commodities transported, and
upon driver training. It is felt that the risk of route A relative to route B will serve as
good measure of their comparative properties. Thf'!se assumptions provide the most
realistic way of making comparisons in a system in which accidents effects are non-
linear (the impact of simultaneously releasing two standard volumes may be greater
than twice one standard volume) and in which very large releases are seldom observed
and poorly characterized.
Weather and terrain-related effects are ignored and a "standard release zone" is
being adopted. This is defined as the maximum distance from an accident each
commodity will affect resources. This distance is determined on the basis of the physical
characteristics of a standard release of each commodity. The zone can be thought of as
extending an equal distance in all directions around a release. Within the radius of this
circle, the value of the affected resources should not be considered as bearing any
defmite relationship to injuries or fatalities within the zone, or to the destruction of
property. By doing so, a way of making risk comparisons among links (selected road
segment) is established: a link in which more of a resource is affected has a higher risk
than a link with less of the same resource affected (i.e. within the reach circle).
The final effect chosen to bear on the consequence of a standard accident, is a
commodity- and resource-specific modifier, a value between zero and one, which
facilitates rating the relative hazards of the commodities on each resource. Given two
commodities of equal reach, this "damage effects" modifier satisfies the condition that
one commodity may have a more severe impact on a resource, and thus should be given
CASE STUDIES IN REVIEW AND MISCELLANEOUS INFORMATION 297
a higher modifying value. By making these values resource-specific, variations in the

ranking of the effects of the commodities on each resources can be incorporated.
Estimation ofthe Consequences from a Transport Accident: - Definitions

Resolll'Ces: Term A represent the amount of resource k reachable by commodity j on link i.
These include people, property, bird habitat, fish habitat.
The value of each resource on each link is found by calculating the weighted average of
census zones within the reach of each of each commodity, and is determined using GIS.
This is highlighted as term A of the equation.
Reach Circle: Term B is the reach circle for each commodity.
The reach of each commodity is assigned a unique value. For each link a reach circle is
definedfor each commodity.
The area of this circle divided by the area of the corridor around the whole link
represents the fraction of the corridor affected in one standard release. This relationship
forms a part of the equation and is highlighted as Term B.
Most HfJl.IJI'dous Product Rating: Term C is the most hazardous product rating.
Each commodity is assigned a modifier value between 0 and I.
The most hazardous product has a rating of 1.0, all other products are compared to
this. This "damage effects" term is highlighted as the third term C of the equation.
The consequence resulting from the actual effect of an accident involving a
r .
dangerous commodity may be defined as:
7t • (reach j
'k =value··k · ·modijier·k (10.2)
'J IJ 2 . length .. reach. J
1 J
The three indexes. i, j, and k refer to the road links, hazardous commodities, and
resources, respectively. The value of each resource (valueijk ) along any given link i,
may be derived from the GIS system (Geographical Information System), which can
perform an interpolation based on the information in its database. This interpolation
algorithm accounts for possible mixtures of high and low resource densities along each
link and computes a weighted average of the resource within the corridor defmed by the
reach of each commodity. For any given accident event on a particular link it is
assumed that only the resources within the "circle of effects" are at risk. Therefore in
the equation above, the value of each resource within the commodity specific corridor is
multiplied by the ratio of the area of the circle to the area of the corridor (i.e., the
fraction of the corridor actually at risk from a release of the "standard volume"). Finally
this value is further scaled by the damage effects modifier. The formula for estimating
consequences from a transport accident is given below:
It R2
C iit = V'it ·ITT·Hjt
I J
(10.3)
Term.4. Term B Ter". C
where, the nomenclature and definitions shown in the box were used.
298 CHAP1ER 10
Calculation of Risk
When Eq. 10.1 and 10.2 are multiplied, they estimtJte the amount of a given resource at
risk (in the restrictive sense) on a given link, for a given commodity, when one standard
unit is released. The final evaluation of risk to any particular route is made by
specifying the volume of traffic carrying each commodity along the road link that forms
the route. This volume is defined as the number of "standard volume" units of each
commodity transported along each link per year.
The total risk can then be calculated by adding up the link specific risk for each
commodity of interest, and arriving at a score for each resource, expressed as the total
value of each resource at risk per year.
It is advisable not to express the resources of interest, i.e., people, property, and
environment in a common unit such as the dollar or any other currency value.
Combining the first two equations with traffic volume, the final equation for total
risk to each resource along a route containing 'i' links and l' commodities can be
expressed as:
(10.4)
The reSUlts, which this model produces, are expressed in units which depend upon
the k resources:
In the case of: Results are expressed as:

People People/year
Property Million SFr. I year
Habitat Number of hectares I year
Euch of these numbers is typically very small, and so to make them more convenient
to discuss they are given a factor of 1 x 10 as a standard multiplier.
10.1.2. FEATURES OF THE COMPUTERIZED MODEL
. The risk analysis model describe previously for the Vancouver case study has been
programmed and implemented on an IBM PC/AT micro-computer. The computer
environment enables the user to create and evaluate routes made up of arbitrarily
complex link combinations, carrying arbitrary combinations of commodities along the
great many possible links in the study area. Using an interactive window-oriented
screen, routes and commodities are easily specified.
Data for the Geographic InjormtJtion System (GIS)

The following specific information should be stored in the GIS, and should be used in
conjunction with the risk analysis model:
• Rail movements of dangerous goods: a map of rail links, and the quantities of
each commodity transported along each link.
• . Road movements: map of roads and transport corridors, with associated accident
rates for each section of road, and the volumes of each commodity transported
along each link.
• Population and property: map of concerned areas together wKh population
density and average residential property values for each case.
• Environment sensitivity: broad areas based on watersheW; and land/water cover
with ranked values on a ten point scale accounting foJ' game, bird and fish
habitats at risk.
The GIS is used to provide information about the resources at risk along each link in
the road and rail transport networks. For each item, a map and its associated textual
attribute data must be available. An example of a typical GIS application is shown
below:
A Typical GIS Application

• Within a study team, determine for each transpon mode the "reach" distance of greatest
likely impact for each commodity
• The moximum reach for a given commodity is estimated to be 4 km, thus the GIS will
generate corridors of 1,2, and 4 /em to each side of every road and rail link (i.e., corridor
widths of 2, 4, and 8 km).
• These corridors can be overlaid on the population, property, and environment maps. The
GIS can compute the number of people, property value, and environment surface area
within each sensitivity zone along the corridor.
• Use a risk model to compute the specific resource values for each commodity, through
interpolation between these predetermined corridors.
10.2. Transport Study for the Lyon Area
A study sponsored by the "Groupe de Prospective" of the French Ministry of

Environment, and cosponsored by the DG XII of the CEC, has been made on the
comparison of two routes for hazardous materials transported by road in the city of
Lyon (France). No major accident has yet occurred in Lyon where about 4 hazardous
material accidents are observed every year. However some catastrophes were avoided
only by chance (20 ton of propane released on the parking of a supermarket in 1985).
A comparative probabilistic risk assessment was prepared for two itineraries:
(a) through the center of the city of Lyon (11 km; travel time: 30 min.), and
(b) suburban itinerary (17 km; travel time: 35 min.).
The number of vehicles re-routed was estimated:
• about 350 vehicles could be rerouted
• 20% of which are empty vehicles
• 60- 70% of the traffic is for gas oil or gasoline delivery in town.
300 CHAPTER 10
For the computation of risks,four products were selected on the basis of their traffic
or their potential hazards: motor spirit (200 trucks/day), LPG (30 truckS/day), ammonia
(2 trucks/day), chlorine (0.2 trucks/day). The route was subdivided into a finite number
of potential accident points (about 20 for each route) to which an accident probability
was attached. A city map was prepared with indication of the population density
(resident and working population) corresponding to selected potentially impacted areas.
The grid was 1 x 1 kIn. In addition meteorological data were gathered and aggregated
according to the wind rose, and three weather conditions were selected.
Key figures for the study concerning the Lyon's area are the release probability per
accident and lethal area knowing the size and nature of a release (6% of total release for
LPG accidents). Lethal area estimation implies hypotheses on release rate, fire ball size,
turbulence, atmospheric dispersion, and on toxicity, or on heat exposure dose response
relationship.
Lethal areas have been computed at the 100% and 10% level according to the 12
weather categories for the four product categories. A few thousand of basic events (e.g..
given physical phenomena following an accident at a given point on the route and under
a given weather category) were taken into consideration. In every case the probability of
the event is known as well as the consequences within each cells, since the number of
deaths is the product of the population density by the fraction of the lethal area within
the cell. Aggregating along the cell provides the total number of deaths due to the basic
event, pooling all the events for a given route provides the risk profile for the route
which can be displayed as a probability - consequences curve or aggregated in an
expected number of deaths. The number of victims among the drivers was also
estimated.
The Results of the Lyon Transporl Study

They indicate that by rerouting, the risk is divided by a factor of about four when the
annual expected number of deaths is considered. On the "city center" route (0.35
accidentslyear, of which 1.5 10lyear concern a release followed by ignition) , the hazardous
nature of the cargo multiplies the rodd traffic risk by about ten. The individual risk was
2.10, it does not reach 10 on the suburban route (0.30 accidentslyear, and 1.3 10 for major
hazard).
When looking a the four products, the risk reduction is about the same. However chlorine
accounts for one eight of the total risk while it represents less than one thousandth of the
traffic. In the case of Lyon, rerouting not only reduces the risk but also its spatial
concentration. As a matter offact the risk analysis shows that accidents with about 10,000
victims can be expected, although at a very low probability level (l to 10 per year).
Splitting for both itinerary the expected number of deaths according to categories of
accidents depending on their severalties gives a better picture than the probability -
consequences diagram The drastic reduction of the most catastrophic events (>50 deaths)
appears then clearly.
The introduction of weighting factors has been recommended for the US nuclear
industry by the US Nuclear Regulatory Commission (NRC). The weight is the number
of deaths. In Germany, a risk aversion factor for railroad accidents in tunnels was
considered and a weight of ten was assigned to accidents with more then 20 deaths.
Cost Benefit Analysis

The usual practices in road safety imply the use of cost benefit analysis so that formal
decision aiding methods can relay on accepted values. The problem of quantifying the
catastrophic aspect of a risk requires an important work in developing risk indices that are
appropriate for the situation.
10.3. Assessment of Transportation Risk for the Canton "Baste-Landscape"
On behalf of the authorities of the canton Basle-Landscape ("Basel-Landschaft"), a

study concerning the assessment of transportation risks for dangerous goods was
prepared by an engineering consortium in collabomtion with the relevant Safety
Inspectorate Office ("Sicherheitsinspektorat Kanton Basel-Landschaft"), and a prelimi-
nary report was published in October 1992. The main goal was to collect traffic data as
well as data characterizing the different road segments and the environment in order to
prepare a traffic model describing the relevant parts of the road network concerned by
the transportation of dangerous goods and to quantify the risk of pollution along the
segments regarding air (population/animals), water (surface/ground water) and soil
(respectively vegetation). The traffic model was developed by the company RAPP, and
the environmental risks were analyzed by Swisselectra. In the approach presented here,
which differs from the method to be found in the guideline of BUWAL, the different
factors which contribute to the total risk were in a first step considered as independent
parameters, before they were combined for estimating the total risk. The same "Pilot
road network N2", (Region Basle _. Augst), was used as a basis for preparing the
succinct report according to the BUWAL guideline III and for calculating the risks
according to the method proposed by Swisselectra. The results of both studies were
finally compared together before the new method was applied to a larger part of the
road-network (main report). The main report was concerned with an integral analysis of
the risks involving rail, road and the river Rhine in the Canton Basle-Landscape.
Description of the Swisselectra TDG-RiskAssessment Methodology

The risk arising from the transportation of dangerous goods is not only determined by
the physical, chemical, and toxicological properties of the substance, but also by the
risk related to the transport process itself and the chamcteristics of the transport roads.
A selected transport road is usually traversing a region with different population
densities, a varying sensitivity towards the environment (ground and surface waters,
forests), changing building/construction elements (like road type, presence of tunnels
and viaducts). The factors influencing the traffic play also a preponderant role: traveling
speed, traffic composition, accident frequency, number of dangerous goods transports
on a given road segment, traffic management, etc. This means that the contribution of
the different factors is very complex (see Figure 10.1).
302 CHAPfERIO
In this context, the Swisselectra company has developed a method for connecting
the different parameters together, and has on purpose avoided to define and use accident
scenarios, as this is the case, when accessing the risk of a single event (e.g., spill, gas
release, fife, or explosion).
Product Related Environmental Frequency of Accident

Danger Potential Parameters D.G Transport Frequency
I I I I
1 I
Magnttudeof Probabiltty of
Potential Losses I Occurence
Damages
L I
I
Risk of a Dangerous
Good's Transport
Figure 10.1. Elements describing the risk of a dangerous goods transport.
Among the different factors defining the total transportation risk, there are four main
elements to be considered:
• parameters related to the dangerous goods transported (Physical-chemical
properties, toxicity, energy content, etc.) and the quantities transported.
• parameters describing the environment (soil, water, air). Such aspects like
populationlhabitation density, distance to the next buildings, distance to the next
ground or surface water and size of the concerned protected zone, biotopes,
tunnels, viaducts.
• accident frequency. This describes the number of accidents on a selected road
segment resulting from the road trace, traffic density, type of road, topography,
and other factors influencing the fluidity of the traffic.
• frequency of dangerous goods transports on selected road/rail/waterway
segments.
The first two elements, i.e., product related danger potential and environmental
parameters, describe the magnitude of potential losses as a result of some given
consequences. The last two elements, accident frequency and dangerous goods transport
frequency are a measure for the probability of occurrence of a hazardous event. By
combining these factors we obtain the total risk arising from a dangerous goods
transport (Figure 10.1).
From relevant cantonal statistical data it is possible to estimate the quantity

transported per year for a selected number of dangerous products, and the transport
frequency for the product in question.
Scrutinize the statistical data for Establish a concept for assessing

each given dangerous goods the danger potential according 10
transport: the respective dangerous goods'
• using the respective transport classes
quantity
• using the related transport
frequency
I I
il
Determine the respective danger
potential for each hazardous product
based on the corresponding transport
quantity:
• repeat this for each environmental

compartment
Figure 10.2. Procedure for assessing the potential danger of a dangerous product.
For each selected product the danger potential is assessed, using RID, or TERA lists
(Figure 10.2 and Figure 10.3). The aim is to assess the risk to the different
environmental compartments, PH, PA , and Ps. The potential danger does also include
the dispersion range of a toxic substance.
The resulting danger potential, P, for a given product is obtained as follows:
(10.5)
where
PH : Danger potential concerning air, respectively human beings I animals
PA : Danger potential concerning water
Ps : Danger potential concerning soil, respectively vegetation.
304 CHAPTER 10
Compile the parameters describing the endangering of environmental

compartments
I
Classify the substance properties according to danger classes
I
Prepare a finer classification and distribute the substances into further
subclasses
I
Define and adjoin the evaluation criteria
I
Combine these evaluation criteria (ranking factors) with the potential
danger of the substance, in such a way that the resulting figure
represents a measure of action radius and effect range of the hazard
with respect to the different environmental compartments.
Figure 10.3. Procedure for defining the potential danger associated with a substance.
The danger potential for human beings I animals does include the following factors:
(10.6)
where
PI : Inhalation toxicity
PD : Dermatologic toxicity
PE : Potential danger due to the energy content of the product (explosion, fire,
thermal irradiation).
Similar defmitions and algorithms for estimating the danger potential of a substance
relative to water and soil have been defmed in the original Swisselectra report, to which
the reader is kindly referred.
The sensitivity of the environment can only be assessed after the construction
aspects of a road!locality are known, since these may affect the severity of an accident
and influence the retention of spills. A distinction must be made concerning the type of
environmental compartment (air, soil, water, forests) concerned. A matrix procedure is
used to classify and evaluate the respective sensitivities (Figure 10.4).
By combining the potential danger of a substance with environmental data, it is possible
to estimate magnitude of the potential damagesllosses caused to the environment by the
release of toxic products (Figure 10.5). The combination of these elements is defined as
the magnitude of potential losses, or damages. This gives an overview of the critical
points to focus on, along a given road segment concerned with the transport of
dangerous goods.
Compile parameters Apply matrix procedure Compile road/surrounding Apply matrix procedure
describing environmental for classifying lestimating construction aspects for classifying! estimating
characteristics I sens~iv~y environmental sens~ivtty construction aspects
I I I
I
Determine sens~ivtty Determine correcting
of environment factors for Increasing I
decreasing sensitivtty
I
Determine sensitivity of
environment related to:
• Environmental compartments
• Products
• Local conditions
Figure lOA. Procedure for·assessing the environmental sensitivity.
Potential Danger of a Product Environmental Parameters
I
Magnitude of Potential Losses I Damages
corresponding to:
• Environmental compartments
• Product transport
• Local conditions
Figure 10.5. Procedure for assessing damages extent

306 CHAPTER 10
The transport frequency can be directly derived from statistical surveys.and traffic
counting for the given region, respectively road segment The data must be evaluated to
obtain the number of dangerous goods transport per year for selected hazardous
materials.
The Swiss VSS-Norm SN 641 3tOa defines the relative accident rate as:
U·t0 6
Accident rate = _...:'-- (to.7)
D1V; Lj t
where
Uj = Number of accidents on the road segment i
D1Vj = Average daily traffic on the road segment i per kilometer
Lj = Length of the road segment in km
t = Observation period in days.
The average accident rates defined in above mentioned Swiss VSS-Norm are shown
in TABLE 10.1.
TABLE 10.1. Average accidentrales (according to VSS-Norm SN 641 310a)
Road Category valid Average Accident Dispersion

for Switzerland Rate
National Highways 0.48 ±0.20

(Max. Speed 130 km/h)
National Roads 0.5 ±0.1O
(semi-highway with a
max. speed of 100 kim)
Main Roads, outside 1.2 ±0.40
localities
(max. speed: 801100 km/h)
Main Roads, inside 2.1 ±0.40
localities
(max. speed: 50160 km/h)
The accident frequency of road segments, for which no concrete data are available,
can be estimated by taking average accident rates for similar road categories. The
accident rates shall be corrected by applying some correction factors which takes in
account the road particularities (Figure 10.6).
By combining the accident rate (related to a given road segment) with the transport
frequency for dangerous product of concern, it is then possible to estimate the
probability of occurrence of that event (Figure 10.7).
Finally, by combining the magnitude of potential losses/damages with the event
probability, one obtains the transportation risk for the dangerous goods under consid-
eration. Estimation of the risk according to the numerical procedure described here
makes it possible to apply normalized (dimensionless) key values for defining some
characteristic risks and to combine these values (or severity scales) to finally obtain the
total risk.
Classify the road segments Apply correction factors for

and define the corresponding local particularities of the road
accident rates segment
I
Calculate the corresponding accident
rate for each road segments
Figure 10.6. Procedure for classifying the accident rates.
Frequency of dangerous Accident rate for a given

goods' transport road segment
• related to the corresponding • related to the road segment
dangerous goods' transport
I I
I
Probability of occurence
for a given scenario
Figure 10.7. Procedure for assessing the probability of occurrence of a given scenario.
10.4. Transport Study.for the Road Tunnel Baregg
MORIKA is a risk analysis for the Baregg-tunnel l (National highway Nt, Baden,
Switzerland), which has been prepared by "Gruner AG, lngenieuruntemehmung"
(Basle) for the Canton of Argovia The actual status has been investigated and a risk
analysis prepared on the behalf of the Health Department ("Gesundheitsdepartment des
Kantons Aargau, Kantonales Laboratorium").
308 CHAPfER10
Some statistics on road transport of dangerous goods through the Baregg tunnel
(Status June, 18, 1991) are reproduced in TABLE to.2. There were 50 accidents in the
Baregg tunnel for the year 1989, but since this road tunnel exists there has been no
accident involving the transportation of dangerous goods through the Baregg tunnel.
The accidents can be classified as follows:
• Drive up accidents due to heavy traffic
• Accidents resulting from a queue in the tunnel
• Accidents at night due to repairs and maintenance work in the tunnel
• Single accidents during light traffic periods
• No accident involving dangerous goods transports
TABLE 10.2. Statistics concerning the road traffic in the Baregg twmel, Baden. Switzerland
(Status Iune. 18. 1991)
Road Traffic Data for the "Daregg Tunnel" Direction Direction

East-West West-East
Number of Road-Tankers per Day: 165 169
Number of Trucks with Parcel Goods: 104 116
Danger class No Description

1 explosives 0 0
2 gas (H2) 16 19
3 flammable, liquids 134 120
4 flammable, solids 0 0
5 flammable, oxidants 0 0
6 poisonous (class 60) 1 3
7 radioactive materials 0 0
8 corrosive 6 7
~: alo.• Petrol (113). Diesel (213).
For risk assessment pwpose, incident scenarios were developed to suit the condi-
tions prevailing in this tunnel (see box below):
Road Transport Incidents ScelUlrios for the Baregg Tunnel

The following incidents scenarios were taken into consideration for risk assessment:
a) Small fire corresponding to the burning of 100 I gasoline (occurrence: < 1 every 1-2
years; 10 minutes without intervention, 5 MW; 2 m. 450°C).
b) Mediumfire with 1.000 I gasoline (occurrence: < 1 every 25 years; 20 minutes without
intervention. 30 MW. 8 m, 1200°C).
c) Large fire with 10.000 I gasoline (occurrence: < 1 every 350 years; 60 minutes without
intervention, 100 MW, 30 m, 1400°C).
d) Dispersion of toxic vapors and liquids (occurrence: < 1 every 650 years).
e) Explosion with deflagration of 100 I gasoline vapor-air mixtures (10% of 1.000 I
gasoline); (occurrence: 1 every 500 years).
The scenario include a characterization of possible accident/incident. together with

an estimation of the:
Probability of occurrence.
Consequences for people inside and outside the tunnel, such as: smoke, irritating
effects, poisoning, gas/vapor release, heat radiation, pressure effects as a result of
deflagration, number of injured or killed people.
Consequences to the infrastructure and surrounding: barrage of the N I-highway,
damages to the tunnel installations (i.e., concrete, reinforcing steel, electrical
cables, etc.).
Consequences to the ecosphere (in the present case the "Limat" river) as a result
of: extinguishing water/foam; release of transport goods dangerous to environ-
ment;etc.
TABLE 10.3. Risk evaluation of incident occurring in the "Baregg tunnel"
Effects on People
Degree of Light Medium Heavy Severe
DIIIIlIIpS/
Consequences
Poisoning Nuisance Irritant effects PoIsoninc Permanent

(Medical ) damages/
deaths
- Spatial immediate neighborhood Community Region
consequences vicinity
- Effect duration dday <1 day 1 week
- No. of medical single >10 >50 >50

cases people/case
- Medical ambulant stationary stationary stationary

treatment
Physical Injuries Disagreement Light injuries Heavy Injuries Permanent

(surpry) Influence on damages! deaths
surrounding
- Consequences immediate neighborhood Community Region

vicinity
- Effect duration d day < 1 day 1 week
- No. of medical single >10 >50 >50

cases people/case
- Medical ambulant stationary stationary stationary

treatment
The consequences of the corresponding scenarios were finally appreciated regarding

their respective effects on human beings, infrastructure , and ecosphere. The results of
the study were presented as "probability of occurrence" versus "consequences", and four
310 CHAPTER 10
groups of consequences, Le., light, medium, heavy, and severe were retained, according
to TABLE 10.3.
TABLE 10.3 (Continued)
Effects on Ecosphere
Surface waters, Individual On year's Many year's Long lasting

incl. cana1ization damages damages damages damages
- Spatial brook; imparted river; locally river; regionally river; inter-

consequences length < 500 m imparted length imparted length regional imparted
<Skm <50km length
>SOkm
- Effect duration hours days weeks months
- Fish dying some dead fishes < 100 kg < 1'000 kg > 1'000 kg
- Regeneration weeks months within I year several years

period
Ground water Individual On year's Many year's Long lasting

(incl. soil) damages . damages damages damages
- Contamination <SOm <500m <Skm >Skm
extent
- Duration of none <1 week < 1 year > I year
contamination single well several wells regional extension
(drinking water)
- Remedial action simple work simple project coordinated coordinated
(extentldurat<on) I year
Effects on Infrastructure
Supply Traffic Individual Regional Inter-regional
Interruption damages damages damages
- Spatial neighborhood Community Region Neighboring

consequences region
- Effect duration 1 week
- Financial 100 Million SFr.
damage
Traffic Traff"IC Trame Traffic Barrages

slackening deviations interruptions
- Spatial immediate neighborhood Community Region

consequences vicinity
- Effect duration <1 hour < 1 day <1 week > 1 week
A survey of the preventive measures and intervention plans was also prepared, and
the benefit of implementing some new proposed measures for preventing incidents
and/or improving the technical standard and safety of the tunnel were assessed.
New additional security measures will be implemented in the near future, and an
improved coordination of the different instances (police, fire-brigade, hospitals, etc.) is
foreseen.
10.5. Transport Study for the Canton Uri
The Authorities of the Canton Uri have asked the engineering company EB&P (Ernst
Basler&Partner Ltd., Basle, Switzerland) to prepare a risk analysis concerning the
environmental protection status of rivers, ground waters, lakes, and tunnels, what the
transport of dangerous goods by road is concerned. Risk assessment for this area is of
prime importance since the region includes the national highways N2 (from
"Seelisberg" to the "Gotthard tunnel") and N4 ("Axenstrasse"), and tunnels like the
Gotthard tunnel and the Seelisberg tunnel through which an important international
traffic is mowing daily.
We reproduce here the content of EB&P's report, which was summarized and
translated into English to serve as an example of a "Swiss Dangerous Goods Transport
Study".
The statistics concerning the traffic, especially the transport of dangerous goods,
was complemented by direct surveys at some strategic points. The dangerous goods
traffic through the Gotthard tunnel amounts to 7% of the total goods transports through
the tunnel. On a regional basis it represents orJy 5%. The traffic surveys of the strategic
points, for the year 1990 (fABLE 10.4), have shown that the proportion of trucks
carrying dangerous goods can vary between 1 and 3% depending on the road link
concerned and date of the survey. A value of 4% was considered for the risk analysis.
The national highway N2 is an important North-South connection. There are approxi-
mately 600,000 trucks per year transiting on this axis.
TABLE 10.4. Total nwnber of bUcks respectively number of trucks carrying dangerous goods 01\ the
National highways N21 N4 (for the year 1990)
Localities
GGschenen Erstleid Bolzbach Fliielen
(Getthard (Seelisberg
tunnel) tunnel)
VehicleIYear VehicleIYear VehiclelYear VehiclelYear
Road-trucks (total) 620,000 750,000 490,000 270,000

Road-trucks (with 25,000 30,000 20,000 11,000
dangerous goods)
Source: EB&P (Risk analysis for the national roads of the Canton Uri, 1991)
312 CHAPTER 10
The following statistical infonnation and documentation were used:
Statistical Information and Documentation Used for the Transporl Study of the Canton
Uri
• Statistics 1984 on Road Transportation of Goods (Erhebung Uher die Giitertransporte
auf der Strasse 1984. Bundesamt for Statistik).
• Statistics 1989 on Road Transportation of Goods in the transalpine region (Erhebung des
transalpinen Giiterverkehrs 1989. Dienst GVF, Sigmaplan, August 1990).
• Statistics concerning traffic accidents in the canton Uri (Verkehrsunfallstatistik des
Kantons Uri 1986 bis 1989).
• Road traffic accidents in Switzerland 1987 (Strassenverkehrsunfiille in der Schweiz
1?87).
• List of classified liquids dangerous to waters (Liste der klassierten wassergefiihrdenden
Fliissikeiten. BUWAL, 1981).
• Products and Goods Classification. (Klassierung von Stoffen und Waren, Brandverhii-
• tungsdienst BVD).
• Catalog of substances potentially dangerous to water ("Katalog der wassergefiihrdender
Stoffe" des Bundesministerium des Innern der BRD un die schwarze Liste der EG,
1.3.85).
• Guide for road transportation of dangerous goods (Leitfaden for den Transport gefdh-
rlicher Giiter auf der Strasse SDR. Bundesamt /iir Polizeiwesen, 1988).
Among the critical objects to be protected, there are the river "Reuss", the protected
natural area between the "Reuss delta" and "Polenschachen", the lake "Vierwaldstiitter-
see", and the ground water basin from "Amsteg" to the lake. On the road network of the
Canton Uri, we reckon with one accident every ten years leading to a light pollution of
water surfaces. A more severe contamination of say 1 km of surface water, or ground
water would happen once every 100 years. Note that a weighted surface area equivalent
was used here for comparing the consequences of water contamination by toxic cargo
releases.
The approach followed for this study was as follows. A list of critical substances
was prepared according to the following criteria (TABLE 10.5):
Criteria Regarding Effects and Impact of Pollutants in Water and Aquifers

• Organic or inorganic product
• Solubility in water
• Potential danger to water and aquifer
• Behavior of the substance in water
• Impact on Health
• Impact regarding water pollution
• Expected reaction of the substances in contact with air, water, etc.
TABLE 10.5. List of toxic products dangerous to water
Product Water solubility Toxicity Class Danger Type ..

(for water) D= directly
I = indirectly
dangerous to water
Phenol good I D
Sulfuric Acid very good 1 D
Caustic Soda very good 1 D
Gasoline not soluble 2 D
Acryl nitril not soluble 3 D
Benzene not soluble 3 D
Trichlorethylene bad 3 D
Ethylmercaptan bad 3 D
Potassium cyanide very good 3 I
Ammonium very good 1 I
nitrate
Ethyl-Parathion not soluble 3 I
'" According to HOMMEL "Handhuch tier gefiihrlichen GUIer", Springer Verlag.
Regarding this study it was recognized that 90% of all dangerous goods transported
had to be considered as toxic to water. The distribution of the products into danger
classes was done on the basis of three documents:
• The German catalog of substances potentially dangerous to water ("Katalog der
wassergefiihrdender Stoffe des Bundesministerium des lnnem der BRD ") considers four
toxicity classes (for water):
Class 1: in general not dangerous to water
Class 2: slightly dangerous to water
Class 3: dangerous to water
Class 4: very dangerous to water
The classification is primarily done on the basis of the toxicity to mammalian
animals, toxicity to bacteria, and the toxicity to fishes. This list contains the following
main representatives:
• The BUWAL list of liquids dangerous to water.
• The Swiss classification list for commercial products and substances prepared by the
Fire Prevention Office for Industry and Trade ("Brandverhiitungsdienst Jilr lndustrie
und Gewerbe").
Applying such classification lists to the dangerous goods transported in the Gotthard
region indicates that 85% of all dangerous goods are directly toxic to water (i.e., all
toxic liquids). This corresponds to about 6% of the total transported goods, respectively
to 3.5% of the total number of trucks.
The goods which are indirectly toxic to water include gaseous and solid products
which cannot reach the water directly. About 5% of the dangerous goods and 5% of the
314 CHAPI'ER 10
goods that are not dangerous can however contaminate the water. Altogether this
corresponds to 5% of the total transported goods. respectively 3% of the total number of
trucks.
Transport of dangerous goods through the Gotthard tunnel is subjected to special
regulations and restrictions (TABLE 10.6) regarding the type of dangerous goods which
are allowed to be transported. as well as the quantities. toxicity class. and combined
truck loads (parcel. bulk).
The philosophy followed in this study for estimating accident frequencies that may
lead to contamination of a water bodies was based on two premises:
a) The accident rates of trucks transporting dangerous goods were deduced from the
general trucks accident statistics of the Canton Uri. the Swiss accident statistics. and
some other international statistics. Based on the percentage of trucks transporting
dangerous goods. the accident rates of trucks transporting dangerous cargo was then
estimated.
b) On the other side. the expected number of accidents was deduced directly from the
accident statistics concerning dangerous goods.
Both premises are not totally independent of each other but allow to check up the
results.
TABLE 10.6. Quantity restrictions for special products transiting through the Gotthard
tunnel
Product Free Maximum Quantity
Phenol
1 Quantity
(up to)
-
Quantity, with
special permit
-
forbidden
(above)
any quantity
Sulfuric Acid Skg Skg Skg
Caustic Soda 200 kg 200 kg 200 kg
Gasoline SOOkg 500-S.000 kg S.OOOkg
Acryl nitril 10 kg 10 kg 10 kg
Benzene SOOkg 500-S,OOO kg S,OOOkg
Trichlorethylene Skg Skg Skg
EthyJmercaptan - - any quantity
Potassium cyanide - - any quantity
Ammonium 200 kg 200-2,000 kg 2,000 kg
nitrate
Ethyl-Parathion Skg - Skg
What concerns the potential hazards to water bodies, two types of accident scenarios
were considered:
• Accidents leading to a leak involving toxic cargo release (with or without fIre
event) and presenting a direct hazard to water bodies
• Trucks fire incidents, which can cause an indirect hazard to water bodies, if
extinguishing foam/water is used to combat the truck's fires.
According to the Swiss statistics2 concerning the general traffic for the year 1986 to
1989, there were approximately 5,000 accidents per year, of which 40 to 50 accidents
involving trucks. Around half of them, that is 20 to 25 accidents took place on national
roads. Taking into consideration the number of traveled kilometers per truck this
corresponds to an accident rate of 7 x 10 per truck x kIn. In order to calculate the
accident rates for the different road segments it was also necessary to consider the
accidents statistics for personal cars, and on this basis to deduce the segment specific
accident rates for the trucks. These amount to:
20 x 10·' accidents per truck x kID on the "Axenstrasse",
4 x 10·' accidents per truck x kID in the "Seelisberg" tunnel, and
7 x 10-' accidents per truck x kID for the other road segments.
Frequency ofAccidents Involving Cargo Leakage

Based on the above assumptions, the number of road trucks per year, and the segment
lengths, an expected number of 21 road trucks accidents was estimated for the national
roads of the Canton Uri. Under the assumption that 3.5% of all the trucks are
transporting goods which are directly hazardous to water bodies, and that accident rates
of such vehicles is quite similar to the accident rates of the other trucks. A number of
0.7 accidents per year can thus be deduced. Following the same approach a number of
165 accidents per year can be prognosticated for the whole of Switzerland. For the year
1987 there were 115 road-tanker accidents in Switzerland In the Canton Uri there are
only two road-tanker accidents which are known for the year 1988. Assuming a period
of ten years this gives an accident rate of 0.2 accidents per year. This means that this
value (0.2) is three times smaller than the value derived using the previous method
(0.7).
For assessing the direct hazards to water bodies only accidents presenting a leak of
toxic liquids are relevant. The number of leaks was deduced from the accident statistics
for the Canton Uri, 3 assuming that this was the case in 5-10% of the trucks accidents.
The last percentage figure is based on an evaluation of about 100 larger accidents
concerning the transport of dangerous goods that occurred in Switzerland and in other
countries, and on a report prepared by TUv Rheinland (D).
For the national roads of the Canton Uri a figure of 0.05 road-tanker leakage-
accident per year was derived (i.e., one every 20 years). The global value for the whole
of Switzerland amounts to 13 leak-accidents per year. The latter value cannot be
directly compared to the figures taken from the different statistics.
Frequency of Accidents Involving Cargo Fire

The result of these estimations and the frequencies used for assessing the transportation
risks for the Canton Uri are reported in TABLE 10.7.
316 CHAPTER 10
TABlE 10.7. Estimated accident frequency (Canton Uri, 1991)
Tanker I Trucks Accidents Switzerland Canton Uri

Total Accidents I Year Accidents I Year
Leak Fire Leak Fire
- All Trucks Accidents 5,000 600 21 2.6
- Tankers Carrying a Cargo 165 18 0.7 0.08

Dangerous to Water Bodies
- Tankers Accidents Involving 13 9 0.05 0.04
Cargo Leakage
Source: EB&P (Risk analysis for the national roads of the Canton Uri, 1991)
TABlE 10.8. Number of transport accidents involving leakage of dangerous cargo
Quantity or Number or Accidents per Year

Cargo Released 1985 1986 1987 1988 1989 1990
500 - 5,000 liters 1 1 4 3 1 0
> 5,000 liters 1 0 1 2 1 1
TABlE 10.9. Frequency of occurrence of the accident scenarios for the Canton Uri
Product Properties Key Products Quantity or Cargo Released
up to 500 500- 5,000 >5,000 Canton Uri

Dangerous to Water kg kg kg (Total)
- Directly Danger Phenol A B C

Classl,2 Sulfuric acid 0.0250 0.0125 0.0060 0.0435
(Water) Caustic soda (50%) (25%) (12%) 87%)
Gasoline
-Directly Danger Trichlorethylene D E F

Class 3 Ethylmercaptan 0.0036 0.0018 0.0009 0.063
(Water) (8%) (4%) (1%) (13%)
Dangerous to Water Potassium cyanide G H I

-Indirectly Ammonium nitrate 0.0280 0.0080 0.0040 0.04
Phosphoric acid (70) (20%) (10%) (100%)
ester (Extinguishing
foam)
The frequency of accidents involving a fire was detennined on the basis of reported
trucks fires (TABLE 10.7). There are no figures available for Switzerland. The figures
were derived by using the data reported by the fire brigade during interventions on the
national roads of the Canton Uri. For the last 10 years period there were only 2-3 trucks
fires reported. About 3% of these trucks were carrying cargoes that were indirectly
dangerous to water bodies. A frequency of 0.08 accidents per year involving trucks fires
can thus be derived (TABLE 10.8). Usually half of these fires are localized to some
parts of the truck and can be extinguished by the truck driver itself. For this study it was
assumed that in 50% of the cases the cargo took fire. According to this premise we find
an frequency of 0.04 accidents per year.
TABLE 10.8 shows the number of transport accidents involving leakage of
dangerous cargo. The compilation was established on the basis of answers to an inquiry
sent to some 18 Swiss Cantons.
Accident Scenarios and Definitions

For both types of accidents, leakage and fire, the frequency of occurrence the scenarios
(A-I) was estimated, as shown in TABLE 10.9. The distribution of the frequencies
shown in this table is based on the results of comparative studies (TOY Rheinland, D).
The natural objects potentially in danger (such as natural reserve, lakes, rivers,
ground water, etc.) which need to be protected, were listed, and their surface area
estimated. Depending on the type of object, a weight of 1 (small importance) to 10
(very relevant) was attributed to each object. The national roads in the Canton Uri were
subdivided into 10 segments and their lengths recorded. In order to estimate the
consequences of a toxic release possibly contaminating some water bodies, two indica-
tors were used:
• the surface area (km) of a potentially endangered ecosystem
• the surface area (km) of a potentially contaminated ground water area
It was also assumed that leakage would take place in a relatively short time.
When estimating the degree of destruction, it must be kept in mind that the effects
and damages are increasing with the toxicity of the cargo and the quantity released, but
are decreasing with the distance to the point of release. The degree of destruction/conta-
mination was dermed as follows:
• no or very little destruction
• 10% destruction
• 40% destruction
• 80% destruction.
For each of the scenarios (A-I) and for each object the degree of destruction/conta-
mination of the ecosystem was finally estimated and presented in matrix form as shown
in TABLE 10.11.
On the horizontal axis are represented the different road segments. On the vertical
axis we represent the endangered objects. The degrees of destruction are indicated for
each object and road segment with a small matrix showing the 9 scenarios defined
previously. The corresponding scenario box is then shaded to indicate visually the
degree of destruction/contamination (or alternatively marked with the corresponding
percentage figures).
318 CHAYl'ER 10
TABLE 10.10. Risk assessment methodology for trucks accidents
Task 1: Collect the following road accidents stotimcs (global traffic accidents, respectively
truck accidents only): Swiss statistics, foreign statistics, and regional stotistics, showing the
number of tonker-trucks accidents per year.
~: Ust all dJJngerous goods transported in the region (type, toxicity class, quantity)
and classify, if possible, the list according to the transit paths.
Task 3: Extract or estimate the Number of Trucks-Accidents (NTA) involving dJJngerous
goods transport, and select all reported accidents involving tonker-truck fire or cargo
leakage, and express the Accident Frequency (AF) on the basis of Traveled Kilometers per
truck (Tr./em), (i.e., AF = NTA ITotol Tr.Km).
Task 4: Using the above accident frequency figures, and /awwing the number per year of
trucks traveling in the region, and the length of the road segments, it is possible to estimate
the number of trucks accidents on the national roads of the region under consideration.
ThllJ.: Select key products according to the BUWAL's list of classified dJJngerous goods.
Classify again and group the dJJngerous goods according to their physical and
toxicological properties in order to obtoin a reduced list corresponding to 10 to 20 key
products toxic to water bodies.
Task 6: Select and list in tobular form all the road segments belonging to a given passage.
Task 7: Indicate also the type of object (i.e., ecosystem, river, lake, ground water, etc.).
Task 8: Estimate the surface area of the object which may reasonably be damaged or
contaminated.
Task 9: Define a weight factor for each object representing the importQnce of the object
with regards to its use (recreation area,jishing, drinking water, game,forest, etc.).
Task 10: Define different scenarios and estimate the accident frequency for each one.
I.-,§!.1l: Estimate for each scenario the potential degree of c:.estruction of the object as a
function of its distance from the road.
Task 12: For each segment multiply the surface area (/em) of the endJJngered object
(belonging to this segment) by the weighting factor (-) and by the degree of destruction (lib)
which has been estimated for the object. The resulting figure (expressed as modified sUrface
area, in /em) represents the extent of the damages for the corresponding segment.
Task 13: Finally add all the modified surface areas belonging to the same type of object to
obtoin the total sUrface damagedlcontominated corresponding to a given class of objects.
Task 14: Establish three or more different scales all of the same length, and place them
above each other on a diagram:
- - one representing the damagedlcontominated surface area (/em) for an ecosystem,
ranging from 0.1 to 10,000 km (on a logarithmic scale).
- - another one for ground water, ranging from 0 to say 4,000 /em (on a logarithmic
scale), etc., and the last onea special linear descriptive scale is used to compare, and
describe the extent of the damages/contamination. This scale is used to obtain the
resulting event severity value, ranging from 0 to 1 (on a linear scale), by projecting the
values (/em) of the previous scales onto it (that is, it relates them to this descriptive
scale). A value of 0.2 corresponds,for example, to an accident; a value of 004 to a big
accident; and a value of 0.5 to 1.0 to a catastrophe.
TABLE 10.11. Matrix representing endangered objects and the degree of destruction/contamination to
be expected for each one
Endangered Surface area Object Road Segment Road Segment

Objects of Objects Type I R-t R-2
Weight
Object t Ground Water 10
3km
Object 2 Ecosystem 3
6km
In order to calculate the extent of the potential damages, one has to solve the tasks
described in TABLE 10.10. The methodology for assessing the transport risks of road
trucks and road tankers was proposed by EB&P for the Canton Uri (Figure 10.8).
Normal AccIdent BIg AccIdent Catastrophe
I I I I I I Ewnt aevertIy
0 0.2 0.4 0.6 0.8 1.0
I I I I I I Ecoaystem
0.1 1 10 100 1000 10000 km2
I I I I I I Ground water
0.1 1 10 100 1000 km2
Figure 10.8. Example of scales used for connecting the different damaged I contaminated
surface areas with the event severity scale.
The resulting event severity for combined damage/contamination of both the

ecosystem and the ground water may be obtained by using a mathematical relation
based on fuzzy set theory (Wyler and Bonenblust, 1991).
10.6. The Project STRADA·DB
STRADA-DB ("Datenbank for STRAssen-DAten") is a project dealing with the

development and implementation of a data base (SDB, "Strassen-Datenbank"), and an
extensive infonnation system for road maintenance management (MSE, "Management
der Strassenerhaltung") dedicated to the Swiss roads network under the hospice of the
320 CHAPTER 10
Federal Department for Roads Construction. The data base shall contain all the
necessary information related to the national roads and their associated maintenance. It
is an integrated management information system using knowledge expert and
ORACLE. The following information shall become available:
• Description of the road network and its particularities
• Description of traffic, environmental factors, and traffic management systems
• Description of the road conditions, state of repair works, etc.
• Air quality measurements data results
• Noise measurements data results
• Accident statistics I Description of
• Land Information System, LIS (INTERGRAPH)
• Heavy transport module
• Expert system with management information and decision support for planing
road repairs and maintenance activities (including GIS and ORACLE).
The system will include decentralized information systems on the basis of
ORACLE, which shall be interconnected over a network.
10.7. Companies Profiles and Miscellaneous Information
Several engineering and/or software engineering companies have specialized in the field
of dangerous goods transportation. In the course of the PPR&S(i) we have contacted
and interviewed a few of them, as well as some Cantonal Authorities in charge of
assessing the risks of transportation, in order to gain more insight in the practical
aspects of dangerous goods transportation and risk assessment. It was, however, not
possible to visit all of them since our resources in man power and time'were very much
limited. The companies and authorities mentioned below are just given as example to
show what is being done, in Switzerland, concerning the risk assessment of dangerous
goods transportation.
10.7.1. SOFIW ARE ENGINEERING COMPANIES
Miscellaneous information dealing with transportation or risk assessment are presented

below:
ASIT Company
Company: AS IT, Aktiengesellschaft flir Sicherheits- und Informationssysteme im
Transportwesen.
Address: ASIT, Thunstrasse 97, CH-3000 Bern, Tel: 031-352 94 77.
Director. Mr. R. Hanni.
(il PPR&S: Poly Project Risk & Safety of technical systems, Swiss Federal Institute of Technology
ASIT is dealing with safety and infonnation systems in the field of transportation.
They are, at present, involved in the preparation of short risk assessment reports (as
required by BUW AL), environmental impact assessment, and proposals for remedial
actions concerning dangerous wastes. It is worth mentioning, that ASIT participated to
the elaboration of BUWAL's Handbook III.
They have developed together with "Springer-Verlag", Heidelberg, the first
European "Safety Infonnation System" (on CD-ROM), dealing with dangerous goods.
This data base gives information on transport classes, physical and toxicological
properties, remedial actions in case of accident, first help measures, etc. The inform-
ation sources were taken from "Hommel Handbook", "Chemdata", "Action Plans of the
Swiss Fire Brigade", "Merck Catalog", "Swiss Poison List" (Schweizer Giftlisten,
BAG), "Goods Catalog of the Swiss Customs Offices", "Safety Data from SUVA"
(Swiss Accident Insurance Institution), "Handbook of the Chemical Industry Compa-
nies" (VCI, Germany).
Another product developed by AS IT, is a "Dangerous Goods Module for Radio-
active Substances", which was developed jointly with the "School for Radiological
Protection" in Villigen, Canton of Aarau. The company ASIT offers also an expert
system to be used in conjunction with their product "Classification of Dangerous Goods
Transports", which integrates the following regulations:
- ADR: European Agreement Concerning the International Transport of Danger-
ous Goods by Road, complemented by SDR and GGVS;
- RID: International Convention Concerning the Transport of Dangerous Goods
by Rail, and RSD and GGVE;
- lATA (DGR): Dangerous Goods Code of lATA (International Air Transportation
Association);
- IMDG: International Maritime Organization;
- P'IT: Al traffic regulations of the Swiss Post-Telephone-Telegraph Authorities.
ASIT has developed different software programs for the Communal and Cantonal
Environmental Protection Services ("Wehrdienste"). These include action plans for
combating oil or chemicals spills, and fires, material management, compatibility lists
for chemicals and materials being stored or transported, alarm plans, cadastral surveys,
etc.
They are suppliers of Swiss national maps (1:25'000) on CD-ROM with an
intelligent program for panning the digitized maps (stored as pixel/vector maps).
ASIT has prepared several reports for assessing the transportation risks:
• Kurzbericht Storfallverordnung (StFV), Teilstrecke N2: Hergiswil-Seelisberg-
tunnel (Kt. Nidwalden), ASIT Nov. 1993.
• Kurzbericht Storfallverordnung (StFV), Teilstrecke Bern-Neuenburg Bahn (BN),
Bemer-Alpenbahn Gesellschaft Bem-Lotschberg-Simplon (BLS), ASIT 15 Juni,
1993.
322 CHAPTER 10
FlNAJOUR Company
Company: FINAJOUR Company in Therwil. BL
Address: Therwil. BL. Tel: 061-721.72.92
Director. Mr. Desboeuss
This software company has developed a computerized railway timetable and is also
offering a computerized global positioning system (GPS) for personal cars, Including
optimal selection of roads. The system is called FINASAT. and makes use of a
Rockwell GPS in connection with an Olivetti Codemo computer (sold for about Fr.
2.500.-). FINAJOUR is presently developing information system for trucking com-
panies dealing with dangerous goods transports.
Other global positioning systems are. for instance. the systems manufactured by
GARMIN (GPS 45. etc.) and the GPS system Pixie from SONY.
10.7.2. MISCELLANEOUS INFORMATION
Program ALPHA1
Supplier : Rudolf van Loh. Entwicklung und Vertrieb von Software. Cologne,
Germany .
Computer : IBMlDOS-compatible.
Description: The software package ALPHAI ("Ausbildungs und Lem-Programm als
Hilfe fUr Ausbreitungsprognosen") is written in Turbo Pascal and runs on an IBM/AT.
It was developed for use as learning and teaching tool for understanding gas dispersion
and explosion behavior with respect to emergency and rescue actions (i.e., exercising
rescue in case of catastrophes).
Since the mathematical models are very simple. this computer tool cannot be used
for real events but only for training purpose. By using this simple Gauss dispersion
model and other correlation it is possible to consider and make "worst case estimates"
of release and dispersion events on flat terrain (The model does not consider terrain
effects). The approach presented here follows the German directives of "TA-Luft",4 and
VOl directives. s The calculation of BLEVE and associated heat effects is based on
reference6 •
The program allows the calculation of the following events:
• Hot and cold continuous gas emissions
• Cold release (as a result of container failure)
• BLEVE (of known mass)
• BLEVE (of known volume)
Note: BLEVE stands for Boiling Liquid Expanding Vapor Explosion. It is the case
when a container. containing liquefied gas under pressure, is bursting due to violent
boiling as a result of heat effect. In such a case the content of the vessel is released and
vaporizes at once. In case of ignition it bums suddenly. In such a case it is difficult to
predict the time of explosion.
VCE stands for Vapor Cloud Explosion. It concern the escape of a combustible gas
which may ignite within its flammability limits (i.e., 2.2 - 10 vol.-% for propane). If the
gas mixture is too rich, it will only bum and dissipate much heat radiation.
In the case of a BLEVE, heat effects affecting persons or buildings can be considered,
as well as pressure effects resulting from an explosion.
Program PCMLUS-92
Description: The software package PCMLUS-92 was developed by van Loh and
Romberg. 7 It simulates emissions from automotive vehicles along roads with scarce
constructions, with or without vegetation alongside the road. 8 The program can
calculate the following emissions: CO, HC (hydrocarbons), NO, N0 2, Pb, S02' PM
(particulate matter/carbon black), and Benzene. The emissions concentrations can be
calculated on a yearly average or a 98-percentile. The programs assumes different road
conditions like noise protection walls or vegetation and their distance relative to the
road side. Traffic volume, speed and composition are also required.
EXIS Expert information system

Supplier : K.O. Storck Verlag, Stahlwiete 7, D-2000 Hamburg 50.
Access : PC with modem.
Description : E~IS is a data base with international information concerning dangerous
goods transports (raillroadlair/ship), including physical properties, packaging,
classification, and international legislation.
Decision Making Tools

Luckritz and Schneider,9 (1980), have presented two tools for decision making in
hazardou5 materials transportation. A description of these tools is given in the box
below.
The Equivalent Safety Concept and the Population Vulnerabilily Model

Two tools to determine the hazard presented by a large quantity of hazardous materials are
the Equivalent Safety Concept and the Population Vulnerability Model
• The Equivalent Safety Concept is a non computer technique that develops indexes for
cargo hazards, for vessel design, and port safety. These are used to assist in a judgmental
decision of authorization of vessel transit.
• The Population Vulnerabilily Model is a computer simulation of a cargo spill integrated
with census data. The damage to life and property are calculated using the census data and
the cargo properties to determine the number of deaths and injuries to personnel and dollar
losses from the cargo release. The results and relative hazards are calculated using these
techniques.
324 CHAPTER 10
10.8. Swiss Cantonal Authorities Involved with Dangerous Goods Transportation
10.8.1. ADDRESSES OF THE EXECUTING AUTHORITIES:

SWISS CONFEDERATION
Swiss Cantonal Authorities Involved with TDG

The addresses of the Federal and Cantonal Authorities responsible for implementing
and controlling the application of the BUWAL directives concerning the risk assess-
ment of traffic ways (according to BUWAL's Handbook III) are listed in TABLE 10.12
(Stand 1992).
TABLE 10.12. Addresses of selected Executive Authorities (Swiss Confederation and Cantons)
Canton Addresses of Executive Authorities (Swiss Confederation and Cantons)

AG Kantonales Laboratorium, Sektion Chemiesicherheit
5000 Aarau, Kunsthausweg 24
AI Baudepartement FachsteUe fUr Umweltschutz

9050 AppenzeU, Blattenheimatstrasse 2B
AR Amt fliI" Umweltschutz Abteilung Gewiisserschutz und Stoffe

9102 Herisau, Kasernenstrasse 17
BE Allg. Koordinationsaufgaben: Kantonales Laboratorium, Abteilung

Umweltschutz und Gifte
Postfach 3000 Bern 9
BL Sichemeitsinspeictorat
4410 Liestal, Rheinstrasse 29
BS Kantonales Laborratorium, Kontrollstelle fUr Chemiesicherheit, Gift und

Umwelt
4012 Basle Burgfelderstrasse 25
FR Office de la Protection de rEnvironnement

1700 Fribourg, Route de la Fonderie 2
GE Office Cantonal de l'Environnement

1205 Geneve, 5, Bd des Philosophes
GL Amt fUr Umweltschutz

8750 Glarus, Postgasse 29
GR Amt fliI" Umweltschutz

7ooiChur, Giirtelstrasse 89
JU Laboratoire Cantonal des Eaux

2882 Saint-Ursanne, Champs-Fallat
TABLE 10.12 (continued)
Kanton Addresses of Executive Authorities (Swiss Confederation and Cantons)
LU Amt flir Umweltschutz

6002 Luzern, Klosterstrasse 31
NE Service Cantonal de la Protection de l'Environnement GTPA

2034 Peseux, Case postale
NW Amt fUr Umweltschutz

6370 Stans, Oorfplatz 6
OW Amt flir Umweltschutz

6060 Sarnen, Postfach 311
SG Amt fur Umweltschutt, Abteilung Stoffe und Abfalle

9001 St. Gallen, LinsebUhlstrasse 91
SH Kantonales Laboratorium
8204 Schaffhausen, Postfach 37
SO Amt flir Umweltschutz Abteilung, Chemiesicherheit

4500 Solothurn, Baselstrasse 77
SZ Amt flir Feuer- und Zivilschutz

6430 Schwyz, Schlagstrasse 87
TG Amt fUr Umweltschutz und Wasserwirtschaft Verwaltungsgebaude

Promenade
8500 Frauenfeld
Tl Dipartimento del territorio Sezione energia e protezione dell'aria

6501 Bellinzona, Viale Stefano Franscini
UR Amt flir Umweltschutz, Abteilung Gewasserschutz

6460 Altdorf, Gurtenmundstrasse 33
VO Service de Lutte contre les Nuisances Les Croisettes

1066 Epalinges, Case postale 33
VS Service Cantonal de la Protection de I'Environnement

1950 Sion, Place des Cedres
ZG Amt flir Umweltschutz

6301 Zug, Postfach 897, Aabachstrasse 5
ZH Oirektion des Innern des Kantons ZUrich, Koordinationsstelle flir

Storfal1vorsorge
8090 ZUrich, Selnaustrasse 32
326 CHAPTER 10
10.8.2. ADDRESSES OF THE CEN1RALIZED CANTONAL INCIDENT

REPORTING OFFICES FOR PROTECTION AGAINST CATASTROPHES
The addresses and telephone ,numbers of the official incident reporting offices of the
Cantons (According to article 10, paragraph 2 of the Law for Environmental Protection
("Umweltschutzgesetz"), which have to announce and transmit impressively any
incident to the Federal National Alarm Office (NAZ, "Nationale Alarrnzentrale des
Bundes") are listed in TABLE 10.13 (Status 1992).
TABLE 10.13. Addresses of the centralized Cantonallncident Reporting Offices (for catastrophe pr0-
tection)
Kanton Addresses of the Centralized Cantonal Incident Reporting Omces

for Protection Against Catastrophes
AG Polizeikommando des Kantons Aargau
5000Aarau, Tellistrasse 85
AI Kantonale Polizei
9050 Appenzell, Polizeiposten Gaiserstrasse 8
AR Polizeikommando des Kantons Appenzell AR
9043 Trogen, Rathaus
BE Einsatzzentraie Kapo
Postfach 300IBem
BL Kantonspolizei
4410 Liestal, Polizeikommando, Mtihlegasse 12114
BS Kantonspolizei Basle-Stadt
4001 Basle, Spiegelgasse 6
FR Centre d1nformation et Transmission de la Police Cantonaie (CIT)

1700 Fribourg
GE Commissariat de Police
1211 Geneve 8,17-19, Bd. Carl-Yogt, Case Postale 236
GL Polizeikommando
8750 Glarus
GR Polizeikommando GraubUnden
7000 Chur, Ringstrasse 2
IV Gendarmerie Cantonaie
2800 Delemont, Route de Bille 23
LU Polizeikommando des Kantons Luzem

6000 Luzem, Kasimir-Pfyffer-Strasse 26
NE Gendarmerie
2000 Neuchiitel, Poudrieres 14
Kanton Addresses of the Centralized Cantonal Incident Reporting Ot1ices

for Protection Against Catutropbes
NW Kantonale Polizei
6370 Stans, Kreuzstrasse
OW Kantonale Polizei
6060 Sarnen, Foribachstrasse
SG Kantonale Mellle- und Alarmstelle Kantonspolizei

9001 St. Gallen, Klosterhof 12
SH Kantonspolizei
8200 Schaffhausen, Klosterstrasse 9
SO Einsatzzentrale Kantonspolizei
4502 Solothurn, Ambassadorenhof
SZ Polizeikommando
6430 Schwyz, Bahnhofstrasse 7
TO Polizeikommando des Kantons Tburgau

8500 Frauenfeld, Zi1rcherstrasse 325
TI Commando della Polizia Cantonale

6500 Bellinzona
UR Kantonales Polizeikommando
6460 Altdorf, Tellsgasse 5
VD Centre d'Engagement et de Transmission de la Police Cantonale,

1000 Lausanne, Rue Cite-Derriere 28
VS Police Cantona1e Va1aisanne, Avenue de France 69

1950 Sion
ZG Polizeikommando des Kantons Zug, Aabachstrasse 1

6300Zug
ZH Polizeikommando des Kantons Zurich, Einsatzzentrale

Postfach 370
8021Ziirich
Summary
This chapter includes a short summary of selected case studies and miscellaneous
information of interest to the risk analyst, when dealing with risk assessment of various
transportation technologies and systems. The topics included here are considered as
examples to complement the information presented in the other chapters of this book.
The selection was large in order to include the experience from various countries (e.g.,
328 CHAYIER 10
Canada, Switzerland), and different methodologies currently in use in different

environments and schools of thoughts are described. Finally, a list of Swiss Federal and
Cantonal Authorities responsible for implementing and controlling the correct
application of the BUWAL directives concerning the risk assessment of traffic ways
have been collected and listed at the end of this chapter.
1 MORIKA: Mobile Risikcn Kt. Aargau, NI Baregg-Tunnel, Sicherbeitsanalyse Transport getlhrlich«

OUter (30. September 1991), Gruner AG, Ingenicuruntemehmung, Basel. Auftraggeber: Kantonales
Laboratorium, Gesundheitsdepartement des Kt. Aargau, Schweiz.
2 Sttassenverkehrsunflille in der Schweiz 1987, Bundesamt fUr Statistik, (1988).
3 VerkehrslDlfallstatistik des Kantons Uri, 1986 bis 1989.
4 Schiegel und SchorIing, "Technische Anlage zur Reinhaltung der Luft", TA-Luft, Ecomed, (1986), pp. 10-
18, and pp. 54-59.
S VDl Richtlinie 3783 Blatt I, VDI.
6 Major Hazard Control. A practical Handbook, n.O, Geneva, (1988).
7PCMl.US'92, Benutzer Handbuch (1992). Vanl.oh, R.; Romberg, E.; Keune, H., IndiTecGmbH, Wartburg
Str. 5, D-4630 Bochum 7, Deutschland.
8 Merkblalt Uber Luftverunreinigungen an Sttassen. Teil: Sttassen ohne oder mit lockerer RandbebaulDlg,
Ml.uS-92, Ausgabe 1992, Arbeitskreis ''Luftverunreinigungen an Strassen" bei der ForschungsgeseU-
schaft filr Strassen und Verkehrswesen, Bochum, Deutschland.
9 Luckritz, R. T.; Schneider, A.L., Decision making in hazardous malerials transportation, J. Hazard. Maler.
4 (1980), pp. 129-143.
SUBJECT INDEX
AC-Laboratory • 241
Actual levels • 266
A Actual loads • 266
Accident Addresses
assessing true accident rates· 58 Swiss Federal and Cantonal Authorities •
failure rate data· 57 324
frequency· 227 Air-mixtures transport • 211
frequency during transportation· 65 Alarm organization
prevention· 222 ARMA, Swiss Meteorological Institute •
propagation· 43 86
reporting· 58; 221 Ammonia· 10
reporting, of· 31 Ammonia pipelines • 190
scenarios· 13 Gulf Central Pipeline· 191
severe, major· 278 Anhydrous ammonia· 197
study on accident type • 31 Annual fatality rates· 160
Accident analysis ARCINFO·4
of rail and road transport risks • 73 Assessment of energy systems· 236
Accident frequency· 306 Average accident rates· 306
estimation, of· 314
involving cargo fire· 316 B
Accident rate
computation, of· 5; 296 Bayes method • 58
estimation for ship transports· 160 BAZL office· 78
railroad • 6; 71 BLEVE·322
road traffic • 5 BLEVE rue ball • 73
Accident scenarios Bulk transport
for ship· 163 by train • 125
for the Canton Uri • 316 liquefied gases • 126
Accident speed liquid chlorine • 126
offreight trains • 130 motor gasoline· 125
Accident statistics of hazardous materials· 158
by accident circumstances· 224 Bulk transportation
by location and dangerous goods class modes, of· 71
(RID). 225 Bulletin
damage to people· 225 accident! incident and inventory· 71
for France· 226 Bunds
for Great-Britain .. 226 for ammonia storage· 196
for Switzerland· 221 purpose, of· 196
for the Netherlands.• 226 types, of· 196
for USA. ·226 Bumingrate
for Western Germany· 223; 224 for liquids· 196
Accidental release BUWAL·121; 136
probability, of· 30 BUWAL Handbook· 10
Accumulation factors· 265 BUWAL Handbook m • 301
Acetylene transport· 211 BUWAL office· 82; 86; 102
332 SUBJECT INDEX
c for comparing alternative routes • 60

for selecting toxic water pollutants· 312
CAER program· 32 Critical levels • 266
Canvey survey Critical loads • 266
probability data • 69 concept, of· 266
Case studies dose-based model· 266
companies profiles· 320 information, for· 268
review, of· 295
the project STRADA-DB· 319
transport study for the Canton Basle-
D
Landscape· 301 Damage categories
transport study for the Canton Uri • 311 indicators, for· 277
transport study for the Lyon area • 299 Damages· 278
transport study for the Vancouver area • Damages extent
295 procedure for assessing • 305
transport study MORIKA • 307 Damagesllosses to the environment • 304
Cause-consequence analysis • 22 Danger classes • 204
CEAM-(:enter· 244 Danger potential • 303
CHEMTREC • See Emergency center Dangerous goods • 4
Chernobyl • 286 BUWAL's list of liquids dangerous to
Chlorine· 10; 111; 144 water. 313
chlorine code • 197 classification, of· 204; 205
pipelines, for· 191 danger classes, for· 61
Swiss production, of· 211 method for selecting· 10
transport, of • 210 product categories· 139
Classification quantity restrictions for special products
of hazardous activities· 62 transiting through the Gotthard tunnel
of risk • 62 ·314
Cold rupture of gas carrier tank· 172 risk management· 32
Collision selection, of • 8
impact speed • 94 spills·4
Comparative risk assessment· 60 Swiss Federal Dangerous Goods
lack of knowledge· 270 Inspectorate· 77
potential use, of· 263 threshold quantities • 11
time dependence, in • 269 toxicity criteria· 11
uncertainty in data • 269 traffic statistics for the Gotthard tunnel •
uncertainty in models· 269 311
uncertainty in scenarios· 269 transport hazards • 65
uncertainty, in· 269 transports by road in the UK • 97
Consequence DAP • See Decision-aiding process (DAP)
assessment· 46; 262 DAP advanced techniques
data for analysis· 13 artificial intelligence models • 285
Continuous emissions decision conferencing • 286
comments, on • 276 decision support systems· 285
Cost-benefit analysis· 162; 283; 301 expert systems· 285
CQst-effectiveness analysis • 283 influence diagrams· 285
Costs knowledge based decision systems· 285
external costs· 275; 276 neural networks· 285
CRA • See Comparative Risk Assessment DAT • See Decision-aiding techniques
Criteria methodology· 280
SUBJECT INDEX 333
overview· 282 for hazardous materials transportation·

purpose of· 280 323
Data Decision-aiding process (DAP)· 281; 285
interpretation and comparability, of· 57 relative costs and benefits in DAP· 281
Data analysis Decision-aiding techniques (DAT) • 279
accidents study· 130 cost-benefit analysis· 283
Data bases· 17 cost-effective analysis· 283
ASHMIR·236 multi
CCINFO·238 criteria outranking technique· 283
CHEM-BANK • 237 multi-attribute utility technique· 283
CHEMDATA·238 Department of Transportation· See DOT
CHEMSAFE· 238 DERA interactive transport system· 253
CHEMSURE safety information system DERACS code· 252
·238 Disasters
Dangerous Goods· 238 fixed facilities· 194
DECHEMA • 238 pipeline· 188
demographic· 4 road· 66
description of selected data bases· 236 Dispersion of toxic vapors· 308
ECDIN·239 Distance criterion· 63
Energy Library· 239 Domino effect· 55
ENFLEX INFO • 239 Dose-effect relationship· 43
Environment Library· 239 Dow criteria list· 10
EXIS expert information system for
dangerous goods transports· 323
EXIS international dangerous cargo·
E
239 EDMZ, Swiss Federal Printing Office· 77;
for dangerous goods transport· 95 82
for equipment reliability· 243 Effects
for hazardous chemical~ and materials • cumulative/synergetic· 44
235 explosion overpressure· 54
geographic· 4 fatallnon-fatal • 41
HAZARD-MASTER· 240 heat radiation· 53
HAZDATA·238 short-termllong-term • 41
historical accident data· 236 Emergency Center
Material Safety Data Sheets· 240 for chemical transportation· 56
MHIDAS ·237 Emergency response· 6
NIOSHTIC • 240 Emission values method· 266
OHSIMSDS Material safety data sheets • EMPA Laboratories· 229
240 Environmental factors· 12
OSH-ROM • 241 criteria for route selection· 12
PC-Facts· 241 land use safety factors • 12
POISINDEX·241 Environmental Protection Agency· See
PSI severe accidents· 236 US EPA
SIGEDA·241 Environmental risk criteria· 43
SRADA-DB, a Swiss road data base • Equivalent safety concept· 6; 323
319 Ethylene· 189
STRADA-DB· 59 ETHZ Institute· 236
TOMES·241 European "Safety Information System" •
U.K. information on accident data· 72 321
Decision making tools
334 SUBJECT INDEX
European natural gas transportation Freight volume· 4

network· 182 Frequency· 278
Event estimation for ship transports· 163
frequency· 44 of cargo leakage· 315
initiating event • 44 of cargo release· 163
mean fatalities (per event) • 42 of ship striking • 164
probability· 44 of tanker explosion • 163
Event modeling Frequency analysis· 70
release of motor spirit· 132 failure frequencies • 71
Event probability· 261 of truck accidents· 70
data for estimating· 17 Fuel oil· 113; 114; 145; 146
estimation, of· 15 Fuzzy sets • 284
Event severity
calculation, of· 319
Event severity scales· 319
G
Event tree Gas cloud
for chlorine truck accident· 18 drift and dispersion, of· 170
for gas carrier· 164 explosion· 170
for LPG tanker accident· 20 explosion probability· 170
for petrol tanker accident· 19 Gas pipelines
Event tree analysis· 17 diameters used in Switzerland· 182
computer codes, for· 18 network in Switzerland· 184
Explosion· 308 operating companies· 186
estimation for ship transports· 165 reported incidentsldamages in
Exposure assessment modeling • 244 Switzerland· 185
Gasoline· 109
F Gauss dispersion model· 322
Gaussian puff dispersion model· 170
Failure frequency Geographic information system • See GIS.
tank wagon equipment • 130 See GIS
Failure probability data Geographic information systems
for tanks and components· 231 application related to transportation •
Failure scenarios· 68 245
Fatalities • 162 geographic data bases • 246
Fault tree analysis· 20 GIS applications for transportation· 255
computer codes, for· 21 GIS mapping and spatial analysis· 255
Filling ratio· 195 national transportation atlas databases •
Fire • 98; 308 249
combustible liquid· 98 natural resources information system·
flammable liquid· 98 254
liquefied flammable gas • 98 German catalog of substances potentially
Fire & Explosion Index· 242 dangerous to water· 313
Fire scenarios GIS ·245; 297; 298. See Geographic
for gas carrier· 167 information system
Fluid discharge GIS applications
adiabatic conditions· 169 selected papers and references· 254
isothermal flow· 169 GIS Center (DOT) • 248
F-N curves· 42; 49; 160; 279 GISIREGIS • 4
Freight traffic Global positioning system· See GPS
survey for dangerous goods· 96 GPS
SUBJECT INDEX 335
FINASAT for personal cars· 322 Heat radiation· 195

Groundwarer·103; 107; 108; 113 Highway tunnels· See Road tunnels
contamination scenario, for· 144 Historical data • 70
maps of resources· 137 HSC
zone classification· 103 Health and Safety Commission, GB • 40
Group dynamics· 287 Human errore 5; 17; 31
Guidebooks on TGG· 75 Hydrocarbon pipelines· 189
St. Fergus pipelines· 190
Hydrogen transport· 211
H
Hazard I
classification· 9
identification· 9 IAEA Agency· 62; 75; 79; 80
Hazard analysis Ignition sources· 162
CARA code· 251 Incident scenarios
compurer codes, for· 249 for the Baregg tunnel· 308
LPG-l code· 251 Incidents reporting cenrer
PHAST code· 251 for Switzerland· 86
SAFETI code· 250 Individual fatality risk
TECKJET code· 250 procedure for estimating· 48
WHAZAN-II code· 249 Individual risk • 43; 159
Hazard assessment crireria, for· 50
method, for· 262 defurition, of· 40; 42
objectives, of· 261 fatality risk levels· 50
Hazard identification tools· 261 probabilistic safety criteria, for· 42
Hazardous marerials transportation· See values adopted for Great Britain· 134
also Dangerous goods transportation, Industrial gases suppliers • 212
respectively, Transportation of Information services
dangerous goods for risk and safety· 240
estimation of the consequences, of· 13 Information sources
Hazardous product rating· 297 Canvey survey· 68
Hazards failure and event data· 68
of fICe/explosion • 43 failure rare data for fIXed installations •
Health 67
occupational· 277 for risk assessment· 202
public· 277 for Swiss statistics· 202
Health and Safety Commission· 40 ignition of leaks· 68
Health effects· 278 petroleum products· 216
chronic· 268 relared to dangerous goods· 321
indirect· 268 Inrernational Atomic Energy Agency· See
Health risks IAEA
comparative assessment, of· 268 International Maritime Organization· See
indicators· 268 IMO
Health risks indicators
acure fatalities· 268
delayed fatalities· 268
L
genetic effects· 269 Labeling· 75
incidence of injury. 268 Land use
relative utility loss· 269 implementation guidance, for· 24
years of lost life • 269 safety factors, for· 2
336 SUBJECT INDEX
Land use planing Mond Fire Explosion & Toxicity Index·

risk assessment criteria, for· 51 242
Legislation concerning TOO MSE, "Management system for highway
international regulations· 74 maintenance • 319
RED book on TOO transportation· 75 Multi-attribute utility technique • 284
Swiss regulation· 76 Multi-criteria outranking technique· 284
UN Orange Book transportation
regulations • 75
UN transport code· 74
N
US Federal transportation regulations • National agencies
75 Switzerland
Legislation concerning TOO by air EGI, Swiss Federal Dangerous Goods
IATA-OOR regulation· 75; 77; 78 Inspectorate • 77
ICAO-11 technical instruction· 75; 78 SVDB, Swiss Association for the
Legislation concerning TOO by rail Control of Pressure Vessels· 77
RID regulation· 74 USA
Legislation concerning TOO by road DOT·75
ADR ordinance· 74 USEPA·75
Legislation concerning TOO on the Rhine National economic index, CH • 209
ADNR regulation· 79 National risks
Legislation concerning transportation of rail transportation (GB) • 134
radioactive substances . Natural gas • 189
IAEA regulation • 75 import and consumption in Switzerland·
Uquid spill radius· 169 182
Ust of liquids dangerous to water NRC· 300
German catalog· 313 Nuclear power plants· 79
list by the Fire Prevention Office for
Industry and Trade· 313
Loss prevention· 287 o
cost of losses • 288 Occupational risk • 40
hazards control • 288 Oil pipelines
LPG· 10; 189 diameters used in Switzerland· 183
explosions· 159 Operating cost· 34
spills· 159
LPG spills· 95
LPG transport· 210
p
Packaging· 75
M danger classes and corresponding
packaging groups and codes • 205
Material safety data sheets· 222 Pathway model
Matrix representation for consequence for pollutants· 265
analysis· 319 Petroleum products· 204
Mean accident rate· 160 distribution and consumption by region·
Mean fatalities per event· 160 214
Meteorological conditions • 162 distribution, of • 213
Miscellaneous information geographic distribution· 215
software, companies involved with importation, of· 212
transportation/risk assessment • 320 importation I exportation· 217
Models inland sales, of· 221
for pollutant pathway· 267 market share· 213
SUBJECT INDEX 337
sales by categories· 220 procedure for estimating· 64; 65

Swiss statistics • 216 Probability of occurrence· 306
transit pipeline, for· 218 Probability of release· 232
transport by rail • 221 Probability of transport accidents· 295
transport by road· 221 Probit·45
transport by ship· 219 Probit Analysis· 263
transport capacity of road tankers· 213 Propane· 110; 143
transportation, of· 212; 218 PSA • See Probabilistic Safety Assessment
Pipeline companies PSC • See Probabilistic Safety Criteria
Gasnat· 182 PSI Institute· 236
Gasunie·181 Public health risk· 40
Gasverbund Mittelland AG· 182 Public risks • 40
Gasverbund Ostschweiz AG· 182 Puncture force models· 124
Ruhrgas· 181 Puncture frequency
Swiss Gas· 181 rail cars
Pipeline network liquefIed gases • 130
petroleum products pipelines· 215 motor spirit· 130
Switzerland· 181
Western Europe· 181
Pipeline surveys· 189 Q
Pipeline transport in Switzerland· 181 QuantifIed risk assessment· See QRA
Poisson distribution· 58; 160; 251 Quantitative risk analysis
Pool fIre • 132; 161 limitations concerning DGT systems· 86
Population density· 61; 103; 162; 201; 300
Population risk • 23
Population vulnerability model· 6 R
Ports of both Basle RA • See Risk assessment
dangerous goods traffIc· 177 Radioactive substances
goods traffIc • 177 legislation, concerning· 80
Potential impact area· 14 Rail accident rates • 121
PPR&S, Poly Project Risk & Safety of Rail accidents· 66
technical systems· 320 Rail cars fIre • 122
Probabilistic risk assessment Rail containers
for ship· 166 mechanical damages, to • 124
for waterways pollution· 174 Rail link • 298
Lyon case study· 299 Rail tankers
Probabilistic safety criteria· 40 chlorine leakage, from· 127
Probabilities for liquefIed gases, number of • 231
for waterway vessel casualty· 161 statistics, on • 230
of gas cloud explosions· 170 wall thickness· 230
Probability· 278 Rail transportation
estimation for fIxed installations· 63 Swiss methodology for assessing the
estimation for hazardous materials risks, of· 136
transport· 63 Railroad
Probability density function (pdf) • 274 networks· 4
Probability number transportation· 4
conversion, of· 65 Railroad accident scenarios· 126
Probability of ignition· 131 RAlRS accident reporting system· 71
Probability of major accidents Reach circle· 297
for storage or fIxed installations· 64 Regional risk assessment
338 SUBJECT INDEX
pipeline network· 37 of water pollution· 157

rail· 37 public acceptance, of· 282
road· 37 ranking
ship· 37 matrix approach· 266
transportation of dangerous goods· 37 Risk analysis • 311
Regional risks model uncertainty· 274
classification, of • 62 of ports· 159
Relative accident rate· 306 uncertainty, in • 273
Release frequencies XENVIS environmental infortnation
for rail • 227 system ·253
for road· 227 Risk analysis model
Release scenarios for transportation· 71
cold gas release • 172 Vancouver code· 298
cold release of flammable liquids Risk and cost analysis
(atmospheric tank)· 172 models for TOO· 251
cold release of flammable liquids (full RISKMOD code· 252
, bore-leak)· 172 Risk assessment· 40
cold release of flammable liquids (pipe chlorine transport by rail • 126
leak)· 172 comparative
cold release of liquefied gas from methods & tools· 278
pressurized tanks • 171 computer support, for· 249
cold release of liquefied gas from DAT results· 285
refrigerated tanks • 171 DAT tools· 283
for tanker-ship· 171 methodology for ship transportation.
hot release of liquefied flammable gas 159; 161
from pressurized tanks • 172 of tunnels· 101; 307
hot release of liquefied flammable gas pathway approach· 34
from refrigerated tanks • 172 quantitative
Release type· 163 elements· 270
Rhine river results (severe accidents)· 279
navigation· 175 setting boundaries· 264
RID list· 303 special studies, on • 256
RID list of hazardous products· 9 system boundaries, for· 264
RID regulation· 77 Risk assessment criteria· 38
RID/ARD regulation· 204 implementation, of· 43
Risk Risk assessment methodology
acceptability· 44 by Swisselectra· 301
acceptability of risk assumptions· 275 Risk assessment techniques
categories· 276 for ship transportation
damage categories indicator· 278 objectives· 158
data for risk quantification· 276 Risk categorization· 41
environmental· 265 Risk classification· 63
from transport operations· 41 Risk communication· 290
impact categories, of • 277 priorities· 291
impacts· 265 public acceptance· 290; 291
integration of risk elements· 275; 279 tasks· 291
mitigation measures· 61 Risk comparison
of accident· 61 rail vs. road transportation· 61
of explosion· 157 Risk computations· 15
of spill· 61 Risk criteria· 38; 39; 40
SUBJECT INDEX 339
acceptable risk • 38 Road tankers

for industrial complex· 43 statistics, on • 230
for toxic concentration' 55 Road transport· 4
for toxic effect exposure· 55 information, on • 4
indices of industrial injury risk· 39 Road transport statistics
Risk defmitions • 42 for the Baregg tunnel' 308
Risk from pipelines Road transportation
experience in Switzerland' 192 gasoline' 94
experience in the USA' 187 risk of transporting gasoline by truck'
Risk from ship and barge transport· 157 95
Risk indicator Swiss methodology for assessing the
environmental· 268 risks, of' 10 1
overall' 275 Road tunnels
single' 275 accident scenarios, for' 97
Risk levels fire frequency· 98
acceptability· 49 fire prevention and control' 101
fatality· 43 risk assessment, of' 96
fmal assessment, of • 56 societal risk, of' 100
for accident propagation· 55 Swiss risk study "MORIKA" • 101
for explosion overpressure' 55 Roadway elements· 61
for heat radiation' 55 Routing analysis' 11
for injury' 53 alternative routes' 27
for property damage' 55 criteria, for' 12
individual fatality· 43 environmental factors' 3
Risk management· 281; 282 mandatory factors' 3; 11
IRIMS system' 253 subjective factors' 3; 23
X1RIM system· 244 traffic signals • 27
Risk matrices' 44 travel speed • 27
Risk matrix' 65 travel time' 27
Risk mitigation· 62 type of vehicle flow' 27
Risk quantification
methodology, for' 46
Risk representation· 63
s
Risk scaling relationship' 296 SAFETI code' 172
Road SAFETI code' 162
classification' 28 Safety guidelines
level of service· 25 for ammonia' 198
network capability· 2 SBB • See Swiss Federal Railroad
service volume' 26 Sensitivity analysis· 281; 284
structuraVgeometric adequacy, of' 26 Service stations' 221
Road classification' 58 Ship
Road link • 296; 297 gas carriers' 163
defmition, of' 296; 297 tankers' 163
Road risk calculation' 298 type of accidents· 163
total risk' 298 Ship collisions' 163
Road tanker Societal fatality risk
fire and explosion risk in a tunnel· 98 procedure for estimating' 47
fire in tunnels· 100 Societal fatality risk levels
Road tanker accidents procedure for estimating· 46
Western Germany· 222 Societal risk· 162
340 SUBJECT INDEX
analysis, of· 41 of gas under pressure· 193

definition, of· 40; 42 risk of storage tanks • 194
estimation for fixed installation· 65 segregation· 195
procedure for estimating· 46 separation distance, for· 195
road tunnels· 100 standards, codes, guidelines· 194
Societal risk levels· 49 storage tanks maloperation· 193
Software type of storage tanks· 193
"Classification of Dangerous Goods Storage of special products
Transports" • 321 ammonia storage· 197
"Dangerous Goods Module for chlorine storage· 197
Radioactive Substances" • 321 gaseous hydrogen storage· 198
ALPHAI teaching tool for emergency liquid hydrogen storage· 198
and rescue actions· 322 LPG storage· 197
for communal and cantonal Storfallverordnung, StFV· 38; 87; 101;
Environmental Protection Services· 136; 147
321 Surface waters· 104; 107; 108; 138
fpr estimating automotive emission Swiss Authorities
concentrations· 323 responsible for dangerous goods
PCMLUS-92 for simulating emissions transportation. 79
from automotive vehicles· 323 Swiss Federal Office for Environment,
population vulnerability model· 323 Forestry and Landscape, BUWAL •
railway timetable· 322 102; 121; 136
Swiss national maps on CD-ROM· 321 Swiss Cantonal Authorities involved with
SOLAS agreement· 78 TDG·324
Spill Swiss Federal Pipelines Inspectorship· 186
LPG· 95 Swiss Federal Railroad· 6
Spill frequencies Swiss methodology (for rail)
methods for determining· 72 accidents history· 139
Spill Probability· 97 definition of severe damages· 140
Spreading pool· 73 environmental aspects· 137
Standard accident estimation of the probability of incidents
defmition, of· 296 ·147
Standard commodity volume units· 298 frequency of incidents with severe
Standard release volume· 296 damages· 141
Standard release zone· 296 likelihood of incidents· 140
State databases· 70 representative incident scenarios· 140
Statistical data representative scenario "Groundwater" •
for Switzerland· 58 144
on reported accidents· 57 representative scenario "Surface waters"
tank trucks accidents and fires· 100 ·146
Statistical methods· 70 scenario "Explosion"· 143
Statistics scenario "Fire" • 143
on rail accidents· 6 scenario "Release of water-polluting
Storage· 192 liquids" • 145; 147
atmospheric storage tanks· 193 scenario "Release of mineral oil
calculations concerning separation products" • 145; 146
distances • 195 scenario "Toxic gas release" • 144
effect of tIre on storage tanks • 194 technical aspects· 136
heat received by pressure vessels· 196 traffic composition· 138
jumbo storage tanks • 193 Swiss methodology (for road) • 102
SUBJECT INDEX 341
accident statistics· 106 TOO on the Rhine river· 84

defmition of severe damages· 107 Swiss regulations
environmental aspects· 103 changes and revisions since 1993·79
frequency of incidents with severe concerning the protection against
damages· 108 incidents· 82
likelihood of incidents· 106 concerning transportation of radioactive
representative incident scenario for substances
groundwater contamination· 113 atomic law· 80
representative incident scenario for radioprotection ordinance· 80
surface waters· 114 SDRIRSD ordinance· 80
representative incident scenario for the Environmental Protection law· 82
population· 109 for dangerous goods transportation by air
representative incident scenarios· 107 ·77
scenario "Explosion" • 110 for dangerous goods transportation on
scenario "Fire" • 109 the Rhine river· 79
scenario "Release of water-polluting for dangerous goods transportation on
liquids" • 113; 115 waterways· 78
scenario "Release of mineral oil for rail transportation· 77
products" • 113; 114 for road transportation· 76
scenario "Toxic gas release" • 111 other Swiss regulation for DGT by air •
structural and technical aspects of roads· 78
102 other Swiss regulations· 76
traffic composition· 105 other Swiss regulations concerning TOO
traffic information· 105 by rail· 77
Swiss railroad RSD regulation for railroad
accident statistics· 148 transportation of 00 • 77
incidents frequency· 148 SDRIRSD regulation· 76
network configuration· 134 Waters Protection Law· 82
statistics· 135 Swiss reports
traffic structure· 136 assessment of transportation risks· 321
Swiss railroads Swiss statistics
defmition of damage indicators· 149 accident frequency for the Canton Uri·
event severity scale· 150 316
incident frequency· 151 general traffic • 315
Swiss regulation concerning incidents road tanker accidents· 315
Assessment of the • 85 transport accidents involving leakage of
duties of the executive authorities· 85 dangerous goods· 316
duties of the Swiss cantons· 85 Swiss waterways· 175
duties of the Swiss Confederation· 86 ports of both Basle • 175
handbook III, prescriptions for traffic ship traffic, on • 176
ways· 83 System sensitivity· 61
information and alarm in case of
incidents· 86
legal basis· 82
T
main topics· 83 Tank car failures· 124
preventive safety measures· 84 Tank cars • 4
tasks and duties of the owner of traffic Tank truck fires· See Road tanker fires
ways· 83; 85 Tank trucks
TOO by rail • 83 accident frequency, of· 101
TOO by road· 84 Tanker accident· 168
342 SUBJECT INDEX
fluid discharge hypothesis· 169 IMDG code· 229

hazards from bulk substances· 169 U.S Code of Federal Regulations· 229
Tanker explosion· 168 UNO recommendations, for· 229
Tanker terminal wall thickness, of· 229
hazards, of· 233 Transport information· 201
Tankers dimensions· 231 Transport of chemicals· 66
Tanks and containers hazard, from· 66
description, of • 204 Transport statistics· 96; 201
Tanks safety· 229 by BAY· 206
Target levels· 266 bySBB ·206
Target loads· 266 by SBB-CARGO • 206
TOG • See Transportation of dangerous by the Swiss Society of Chemical
goods, respectively, dangerous goods Industry· 208
transportation byUNO·207
TERA list· 303 dangerous goods· 206
Tetrachlorethen· 113; 115; 145; 147 distribution of the goods volume (rail,
Threshold event probability· 167 road, ship) • 209
TNT equivalent· 170 for dangerous goods of class 2 • 210
Toxic gas release· 73 for dangerous goods of class 3 • 212
Toxic products dangerous to water· 313 for dangerous goods of class 6.1, 8 and 9
Toxic products transport· 216 ·216
Toxic releases for rail, by product category· 206
assessment, of· 44 Transportation
Toxicity economics· 2
acute/chronic· 43 environmental factors· 1
Toxicity classes applicable to water information gathering· 4
pollution· 313 safety factors· 1
Toxicity indices' list (fERA-list) • 10 Transportation accidents
Traffic probability, of· 14
AADT volume· 5 Transportation costs· 34
factors influencing transportation· 25 Transportation of dangerous goods by ship
volume· 25 ·157
volume (for urban roads)· 26 Transportation of dangerous goods
Traffic accidents statistics UN transport code· 60
for Western Germany· 222 Transportation of radioactive substances·
Traffic control systems· 116 79
Traffic redistribution· 61 inside a nuclear facility· 80
Traffic statistics max. allowable surface contamination·
for Switzerland· 312 81; 82
Train accident scenarios· 123 on public roads in Germany· 81
Train accidents outside a nuclear facility· 80; 81
British experience· 128 Transportation risk
U.S. experience· 121 factors defining· 302
Train derailment accident simulation· 252 Transportation risk analysis
1RAIN Il- database· 71 overview· 37
Transport accidents Transportation risk assessment
consequence calculation, of· 297 methodology used by the Canton Uri •
defming the consequences· 296 318
probability, of· 295 Transportation risks
Transport containers· 228 comparison of alternative routes· 21
SUBJECT INDEX 343
computation, of • 22 concerning incorporation· 271

information and data, for· 56 concerning population exposure· 271
Transportation routes concerning toxicological response· 272
analysis procedure, for· 3 guidelines on how to deal with
assessment, of • 3 uncertainty· 273
Transportation statistics in databases· 275
USA· 242 in environmental RA· 270
Transportation systems Markow models· 270
comparison of alternative routes· 59 measurement error· 272
comparison, of· 59; 60 of the release quantity· 271
Travel time· 61 of toxic release probability· 271
Truck accident rate· 94 on policy implications· 272
factors influencing· 94 overall uncertainty· 272
Truck accidents probability theory· 270
estimation of accident rates· 28 range, of • 273
factors affecting· 93 techniques for estimating uncertainty·
in the USA· 91 270
rate data • 30 the role of risk managers· 272
rates, of· 91 types, of • 284
severity· 30 variability in environmental risk
severity, of • 92 assessment· 273
statistics· 94 UnconfIned Vapor Cloud Explosions· 67
type/categories· 29 UNEP program· 62
Truck fIre UNIDO Organization· 62
gasoline· 95 US Environmental Protection Agency· See
Truck spills/releases USEPA
gasoline· 96 UVCE·189
LPO·96
Trucks
statistics· 5
v
volume data· 30 Validation· 275
Trucks spills/releases Vapor Cloud Explosions (VCE) • 73
Canadian study, of· 95 Vaporization rate· 169
Tunnel fIre Vehicle control systems· 116
duration, of • 99 Vinyl chloride transport· 211
structural damage· 99 VSS-Norm • 306
temperature, during· 99 Vulnerability model· 45
Tunnel risks Vulnerability models· 262
assessment, of • 307
effects on ecosphere· 310
effects on infrastructure· 310 w
effects on people· 309 Waterways· 84
characteristics, of· 161
u pollution by dangerous cargo· 173
pollution scenarios· 174
Ultratoxic substances· 67 risk management, of· 175
UN list of hazardous materials· 10 Weighting factors· 300
UN transport code· 60 WHO Organization· 62
Uncertainty
about concentration· 271
TOPICS IN SAFETY, RISK, RELIABILITY AND QUALITY
1. P. Sander and R. Badoux (eds.): Bayesian Methods in Reliability. 1991

ISBN 0-7923-1414-X
2. M. Tichy: Applied Methods of Structural Reliability. 1993 ISBN 0-7923-2349-1
3. K.K. Aggarwal: Reliability Engineering. 1993 ISBN 0-7923-2524-9
4. G.E.G. Beroggi and W.A. Wallace (eds.): Computer Supported Risk Management.
1995 ISBN 0-7923-3372-1
5. M. Nicolet-Monnier and A.V. Gheorghe: Quantitative Risk Assessment of Hazardous
Materials Transport Systems. Rail, Road, Pipelines and Ship. 1996
ISBN 0-7923-3923-1
KLUWER ACADEMIC PUBLISHERS - DORDRECHf I BOSTON I LONDON

Quantitative Risk Assessment of Hazardous Materials Transport Systems

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Quantitative Risk Assessment of Hazardous Materials Transport Systems

Uploaded by

Copyright:

Available Formats

QUANTITATIVE RISK ASSESSMENT OF

HAZARDOUS MATERIALS TRANSPORT SYSTEMS

Editorial Advisory board

SPRINGER-SCIENCE+BUSINESS MEDIA, B.V.

ISBN 978-90-481-4666-6 ISBN 978-94-017-2821-8 (eBook)

Printed on acid-fru paper

Ali Rights Reserved

CHAPTER 1: ANALYSIS AND ASSESSMENT OF TRANSPORTATION

CHAPTER 2: ASSESSMENT OF TRANSPORTATION RISKS IN A

2.2.1. QUANTITATIVE RISK CRI1ERIA 38

2.7.6. REMARK CONCERNING THE BUWAL METHOD 87

CHAPTER 3: RISK ANALYSIS OF HAZARDOUS MATERIALS

CHAPTER 4: RISK ASSESSMENT OF HAZARDOUS MATERIALS

4.2.1. MAIN CAUSES OF SPILLAGE 128

CHAPTERS: TRANSPORTATION OF DANGEROUS GOODS BY

5.4. Swiss Waterways and the Ports of Both Basle 175

CHAPTER 6: PIPELINES AND STORAGE TANKS FOR GASES AND

CHAPTER 7: TRANSPORTATION OF DANGEROUS GOODS-

CHAPTER 8: DATA BASES AND COMPUTER SUPPORT FOR RISK

8.2. U.S. Transportation Statistics and Information on Commodity

CHAFfER 9: TOWARDS INTEGRATED RISK ASSESSMENT AND

9.3. Some Methodological Issues in CRA for Severe Accidents 278

CHAPTER 10: CASE STUDIES IN REVIEW AND MISCELLANEOUS

Industrial development is essential to improvement of the standard of living in all coun-

It was felt that existing hazard management techniques need to be supplemented

Integrated Regional Risk Assessment and Sa/ety/Hazard Management (IRRASM)

regional risk assessment and safety management guideline manual complemented by

API : American Petroleum Institute

ANALYSIS AND ASSESSMENT OF TRANSPORTATION RISK

1.1. A Methodological Overview

result in death or injury to people, property damage or damage to the biophysical

Analysis and Assessment of Transportation Routes

Consideration of each of the above factor may on its own or in combination

1.1.1. GATHERING INFORMATION ON THE REGION OF INTEREST

Environmental Resource Mapping. It is of good practice to identify and map the

Demographic and Geographic Data Base. A demographic data base should be

Railroad Networks. Railroad transportation of dangerous goods has to be assessed as

Road Transport. The transportation of dangerous goods may be done by several

b) Using statistics: The second method is to use statistics based on a trucking

A CCI·dent rate Accidents per year

Generally due to a lack of good statistics (based on accident categories), accidents

The Equivalent Safety Concept and the Population Vulnerability Model

1.1.2. SELECTING DANGEROUS GOODS FOR REGIONAL RISK ASSESSMENT

For the assessment of transportation risk, the objective is to first develop a

1.1.3. THE NATURE OF THE MAJOR RISK TO BE ASSESSED

Hazard Identification. The process of assigning a chemical to a hazard category

Class Product Hazardous Product

Method for Selecting Dangerous Goods in the Study Area

CriteritJ list for Dangerous Goods and Products

1.1.4. MANDATORY ROUTING FACTORS

1.15. ENVIRONMENTAL AND LAND USE SAFETY FACTORS

Designating Routes for the Tmnsportotion of Hazardous Materials

Estimations of the Consequences of Hazardous Materials Transportation

Data Needed for Estimating the Consequences of Hazardous Materials Tmnsportadon

Accident scenarios usually includ:

Class of Hazardous Material ImpadArea

Estimations of the Probability of Transportation Accidents

The above three steps (Figure 1.2) may be summarized as follows:

All vehicles accidents Hazardous material accidents

- Hazardous materials accidents I.e th f link (km)

Event Probability Estimation. Careful definition of the events' sequence to be quanti-

Types of Data for Event Probability Estimation

TABLE 1.3. Computer codes available for event tree analysis

~- ~ ~ J; ~