Cities 72 (2018) 102–114

Contents lists available at ScienceDirect

Cities
journal homepage: www.elsevier.com/locate/cities

Urban vulnerability under various blast loading scenarios: Analysis using MARK
GIS-based multi-criteria decision analysis techniques
Yasser Ebrahimian Ghajaria,g,⁎, Ali Asghar Alesheikhb, Mahdi Modiric, Reza Hosnavid,
Morteza Abbasie, Ayyoob Sharifif
a
Malek Ashtar University of Technology, Lavizan, Zip Code: 15875-1774 Tehran, Iran
b
GIS Department, K.N.Toosi University of Technology, Tehran, Iran
c
Malek Ashtar University of Technology, Lavizan, Tehran, Iran
d
Malek Ashtar University of Technology, Lavizan, Tehran, Iran
e
Malek Ashtar University of Technology, Lavizan, Tehran, Iran
f
Global Carbon Project—Tsukuba International Office, National Institute for Environmental Studies, Tsukuba, Japan
g
Faculty of Civil Engineering, Babol Noshirvani University of Technology, Babol, Iran

A R T I C L E I N F O A B S T R A C T

Keywords: This paper examines physical vulnerability of District No. 6 of Tehran to blast damages and blast loads. The main
Physical vulnerability objectives are to enhance planners and decision makers' awareness of the extent of vulnerability of buildings in
Multi-criteria evaluation the study area (under different risk scenarios) and to provide guidance on how to enhance preparedness against
Geospatial Information Systems potential risks. Physical vulnerability of District No. 6 of Tehran was evaluated under three different blast
Blast loading
loading scenarios. An expert survey, involving specialists with expertise in urban development, passive defense,
Ordered Weighted Averaging
construction, and architecture was carried out to extract fourteen vulnerability criteria. Fuzzy criteria maps
corresponding to each of these criteria were produced in Geospatial Information Systems (GIS) environment.
Fuzzy analytic hierarchy processing was used to weigh the criteria. Criterion maps were combined using the
fuzzy weighted linear combination operator. Finally, Ordered Weighted Averaging (OWA) operator was applied
to generate final vulnerability maps. Results show that, for all scenarios, almost 70% of the buildings in the
district are of high vulnerability. Sensitivity analysis confirms the reliability of the model.

1. Introduction considered as an effective strategy for enhancing resilience and miti-

Urban vulnerability can be defined as the liability of cities and their
infrastructures to losses caused by disasters (Karashima, Ohgai, & Saito,
2014). Potentially existing vulnerabilities of an urban system are ex-
posed in the face of disruptive events. Four frequently mentioned types
of vulnerability are namely, physical, social, economic, and environ-
mental (Kim & Marcouiller, 2015). Vulnerability analysis can be con-
ducted for both natural and human-made hazards. The fact that cities
are loci of economic, social, and political activities and account for
more than half of world population makes conducting research on
urban vulnerability of high significance (Desouza & Flanery, 2013;
Parnell, 2016; Williams, Batho, & Russell, 2000). Doing research on
urban vulnerability is specifically important in the context of major
cities that are hubs for regional, national, and international activities.
Results of such research can be used by planners and decision makers to
enhance planning and preparation for disruptive events, that is widely
gating the impacts of disasters (Chmutina, Lizarralde, Dainty, & Bosher,
2016; Desouza & Flanery, 2013; Sharifi & Yamagata, 2016). De-
termining the extent of vulnerability of cities and their elements is an
essential step towards preparation for threats and disruptive events.
This study is focused on determining physical vulnerability of District
No. 6 of Tehran to blast loads, as potential human-made threats. Po-
tential explosion targets are identified and classified based on their
degree of importance. Determining vulnerability of explosion targets
can help planners and policy makers to develop appropriate planning
and preparatory measures for minimizing impacts of potential threats.
A brief literature review is presented in Section 2. The third section
provides further information about the particular threat analyzed in
this study and the characteristics of the study area. Section 4 explains
the research methods and provides a theoretical framework for ana-
lysis. In Section 5 vulnerability maps for different risk scenarios are
provided, results of the statistical analyses are presented, and a

⁎
Corresponding author.
E-mail addresses: y.ebrahimian@nit.ac.ir (Y. Ebrahimian Ghajari), alesheikh@kntu.ac.ir (A.A. Alesheikh), mmodiri@ut.ac.ir (M. Modiri), hosnavi@mut.ac.ir (R. Hosnavi),
mabbasi@mut.ac.ir (M. Abbasi), sharifi.ayyoob@nies.go.jp, sharifigeomatic@gmail.com (A. Sharifi).

http://dx.doi.org/10.1016/j.cities.2017.08.006
Received 3 January 2017; Received in revised form 8 July 2017; Accepted 10 August 2017
0264-2751/ © 2017 Published by Elsevier Ltd.
Y. Ebrahimian Ghajari et al.
sensitivity analysis is conducted to test reliability of the results. Section
6 concludes the study by discussing the findings, providing policy im-
plications, drawing conclusions, and making suggestions for future re-
search.
2. Literature review
The literature on urban vulnerability is immense and still growing.
This literature deals with vulnerability and emergency management in
the context of a wide range of stressors and hazards such as flooding,
drought, energy shortage, wildfires, and terrorist attacks (Adger, 2006;
Coaffee, 2009; Gillen, 2005; Kapucu, 2012; Prior & Eriksen, 2013).
Research on urban vulnerability to invasions and terrorist attacks is
comparatively nascent, but rapidly growing (Coaffee, 2009; Williams
et al., 2000). In a recent paper, Matijosaitiene and Petriashvili (2017)
demonstrate that the extent of vehicular access to buildings, the extent
of mixture of private and public areas, and the distribution pattern of
buildings with identical functions are important factors that terrorists
may consider when selecting attack targets.
In terms of methods used for evaluating vulnerability, there are
many studies that utilize GIS-based, multi-criteria decision analysis
methods for evaluating disaster risk, prioritizing risks, and informing
urban planning. These studies are mainly dealing with vulnerability of
urban areas to natural hazards such as floods and earthquakes
(Ebrahimian-Ghajari, AleSheikh, Modiri, Hosnavi, & Nekouei, 2016;
Fernandez, Mourato, & Moreira, 2016; Hashemi & Alesheikh, 2011;
Zhang, Wang, & Lin, 2015). However, the use of GIS-based, multi-cri-
teria decision analysis techniques for evaluating risks and damages
related to human-made hazards such as blast loads and explosions is
limited. Emphasizing that improving robustness and blast resistance of
civilian structures is of critical importance, Tadepalli and Mullen
(2006) developed a model for assessing risk, and projecting and ana-
lyzing potential damages and losses. Building elements such as struc-
ture type, level of occupancy, number of stories, market value of the
building, glazing type, and maximum number of building inhabitants
are incorporated in their GIS model. The model is used to simulate
vulnerability of buildings located at the University of Mississippi. Re-
sults of the study are visualized in the GIS environment and can be
utilized for blast hazard mitigation. A web-based GIS was developed by
Kulawiak, Lubniewski, Bikonis, and Stepnowski (2009) for spatial vi-
sualization of the impacts that a simulated explosion may have on a
railway station in Gdansk, Poland. Results demonstrate the utility of
this technique for identifying potential physical, social, and economic
impacts of explosions on buildings and critical infrastructure in and
around the station.
A GIS-based risk analysis model was developed by Ma, Cheng, and
Li (2013) to simulate potential impacts of the explosion of natural gas
pipelines on buildings. The potential damages are classified as: “minor
damage, minor structural damage, major structural damage, and partial
demolition” (p. 1190–1191). Armenakis and Nirupama (2013) com-
bined GIS index modeling and risk assessment to develop a spatial risk
assessment method for estimating spatial hazard risks that propane
explosion can cause. Their results show that risk assessment using GIS
techniques is an effective method to prioritize spatial risks and estimate
potential damages depending on the level of proximity to the source of
explosion. Integrating such techniques into urban planning can make
significant contributions to disaster risk management in cities.
Limited research exists on vulnerability of Iranian cities to blasts
and explosions. Ebrahimian-Ghajari, Alesheikh, Modiri, Hosnavi, and
Abbasi (2017) developed a model to evaluate physical vulnerability to
intentional explosion hazards in Tehran. They classified buildings based
on their probability of being selected as targets for intentional explo-
sions. The findings showed that only about 25% of the buildings have
low vulnerability. Using GIS and Analytic Hierarchy Process (AHP)
methods, Azizi and Bornafar (2012) assessed vulnerability of the built
environment in District No. 11 of Tehran. The study uses 14 criteria for
103
Cities 72 (2018) 102–114
evaluating vulnerability. The authors assume that vulnerability to air
attacks and earthquakes can be modeled in a similar way. Due to this
assumption, specific criteria related to explosions are not used in the
study. The study also suffers from two other major shortcomings: the
uncertainty conditions are not considered, and sensitivity analysis is not
conducted.
This brief literature review reveals some major shortcomings of
existing research. Comparatively less attention has been paid to vul-
nerability of cities to human-made disasters and hazards. In specific,
multiple point explosion events have received little attention. Also,
experiments and simulations are mainly focused on small-scale, in-
dividual urban components such as buildings, street networks, and
railways. There is a clear lack of research on vulnerability to blast ha-
zard at larger scales such as neighborhoods, districts, and cities.
Management and mitigation of explosion hazards in cities warrants
further research. This study aims to fill the research gaps and build
upon existing research by demonstrating the utility of GIS-based MCDA
for estimating vulnerability at the district scale. Other noteworthy, new
contributions of this study are as follows: by using fuzzy sets theory,
uncertainty conditions involved in the decision making process have
been taken account of; the list of vulnerability criteria used in this re-
search is more comprehensive; the developed model is capable of
generating physical vulnerability maps for different scenarios and
under different risk conditions; and finally, the study provides re-
commendations and policy implications that can inform planners and
decision makers in Tehran about existing vulnerabilities and actions
that should be taken to address them.
3. Scope of the study
3.1. Explosion events as human-made threats
As discussed above, identification of potential threats and devel-
oping strategies to address them is of high significance in urban man-
agement and planning. In the context of urban environments, threats
can be divided into two main categories, namely, natural and human-
made. There are similarities and differences between these two types of
categories in terms of their underlying characteristics. It is, however,
beyond the scope of this study to elaborate on these characteristics.
Interested readers are referred to McDonald (2003).
Here, the focus is only on explosion incidents as common types of
human-made threats. The physical structure of cities is vulnerable to
blast waves produced by explosive devices. Often times, explosions
result in outward release of pressured gasses. The rapid expansion of
these gasses spreads the surrounding air outward (from the explosive
core) in a spherical pattern. The resulting layer of compressed air is
called blast wave and its pressures is known as overpressure
(Krauthammer, 2008). Blast waves fall into two categories: shock
waves, which are products of the positive phase of the blast and pres-
sure waves, which are attributed to the negative phases of the blast
wave. In general, range and shape of blast waves depend on the amount
of energy released and the distance from the explosion core. Since the
negative phase of the blast wave is less powerful than that of the po-
sitive phase, it is often ignored in designing structures. Therefore, it is
also ignored in this study.
In engineering calculations, various waves and their respective
pressures are approximated and used in the form of pressure-time
variations of the blast under different conditions. Variations of pressure
against time and the maximum amount of pressure in the positive phase
are of higher importance in calculations and are usually expressed by
exponential form (see Fig. 1). The variations shown in this figure are
referred to as pressure-time variations, and are expressed using the
following exponential equation (Friedlander Equation)
(Smith & Hetherington, 1994):
Y. Ebrahimian Ghajari et al.
Fig. 1. Free field pressure-time variation (Smith & Hetherington, 1994).
t −bt ⎞
p (t ) = ps ⎡1 − ⎤ exp ⎛ ⎜ ⎟
⎢
⎣ ts ⎥
⎦ ⎝ ts ⎠
where b is the characteristic of waveform, ts is the positive phase
duration and Ps is the peak side-on overpressure (Fig. 1).
Depending on the position relative to the ground surface, explosions
are divided into three categories: air blasts, surface blasts, and under-
ground blasts (Dobbs et al., 1984). This paper is focused on vulner-
ability of the physical components of cities to air blasts. Here, physical
refers to a wide range of components, including communication and
utility networks, buildings, critical infrastructures, and urban open
spaces. In an air blast, the explosive device is burst within a certain
distance from the target and the resulting waves cause damage on the
structure. This process is illustrated in Fig. 2.
3.2. Study area
The study area is District No. 6 of Tehran. It lies between 35° 42′ 02″
N 51° 22′ 37″ E to 35° 45′ 09″ N 51° 25′ 41″ E. The total area of the
district is 20 km2 (accounting for 3% of the total area of the city) and its
population is 242,302 persons (2.8% of the total population of the city).
Many ministries, embassies, universities, hospitals, corporate head-
quarters, and other governmental and private entities are located in this
area, making it one of the most sensitive and strategic districts in
Tehran. It is obvious that major damages caused by disruptive events at
this district can have significant ramifications well beyond the district,
at local, regional, and even national levels. The number of housing units
in this district (i.e. the number of alternatives in the multi-criteria
evaluation), is 28,228. The map of the study area is given in Fig. 3. The
Cities 72 (2018) 102–114
process taken for selecting the explosive targets, highlighted on the
map, is explained in Section 4.2.
4. Material and methods
Vulnerability is a function of several factors. Thus, multi-criteria
decision analysis models are suitable for its evaluation. Multi-criteria
decision analyses models are able to evaluate qualitative and quanti-
tative criteria simultaneously and are highly effective for dealing with
problems (such as the one investigated in this study) in which both
qualitative and quantitative criteria are involved (de la Fuente, Pons,
Josa, & Aguado, 2016). One substantial challenge in multi-criteria
evaluation is lack of standards for measuring qualitative criteria. An-
other challenge is conversion of the criteria rankings into a common
unit (Buchert, Neugebauer, Schenker, Lindow, & Stark, 2015). Evalua-
tion using multi-criteria analysis is not easy and straightforward. No
standardized process is followed during the process and the accuracy of
(1) evaluation is reduced due to its high dependence on subjective judge-
ments by experts and decision makers. In other words, subjective jud-
gements of decision makers significantly influence the evaluation re-
sults. As a result, a wide spectrum of results may be obtained. One of
the effects of subjective judgements of decision makers is observed
when they express their optimistic or pessimistic views. For a given
decision making problem, risk takers put emphasis on the positive
properties of the alternatives, while risk avoiders highlight the negative
properties (Kuin, Masthoff, Kramer, & Scherder, 2015). Another ex-
ample is use of linguistic variables to indicate the relative importance of
the evaluation criteria or to determine the extent of vulnerability of a
building based on each criterion. Therefore, in multi-criteria decision
analysis it is necessary to utilize techniques that are capable of gen-
erating results by capturing different risk perceptions (ranging widely
from an optimistic view (risk taker) to a pessimistic one (risk averter)).
It is also essential to address uncertainties and management ambi-
guities. In this study, multi-criteria evaluation of physical vulnerability
of cities to blast damage and blast loads is conducted in several steps as
shown in Fig. 4.
4.1. Vulnerability criteria and constraints
According to page 107 of Malczewski (1999), several factors should
be considered when selecting criteria for multi-criteria analysis. “Each
criterion must be comprehensive and measurable. A set of criteria
should be complete (the criteria should cover all aspects of a decision
problem), operational (they can be used meaningfully in the analysis),
decomposable (they can be broken into parts to simplify the process),
Fig. 2. Free air burst blast environment (Smith & Hethern-
gton, 1994).
104
Y. Ebrahimian Ghajari et al. Cities 72 (2018) 102–114

Fig. 3. The study area.

nonredundant (they avoid problems of double counting) and minimal Based on studies conducted by his predecessors, analytical mod-
(the number of attributes should be kept as small as possible)”. Fol- ification of previous equations, and experimental studies, Heinrich
lowing these selection factors, literature review and expert survey (14 developed Eq. (3) through to (5) and used the scaled distance in order
participants) were used for extracting vulnerability criteria and con- to calculate blast overpressure (kg/cm2) (Ngo et al., 2007):
straints. As can be seen from Table 1, selected participants were scho-
14.072 5.54 0.375 0.00625
lars, municipality officials, planners, and practitioners with expertise in Ps = + 2 + − (0.05 < z < 0.3)
z z z3 z4 (3)
passive defense (four experts), structural engineering (three experts),
architecture (two experts), urban planning (three experts) and explo- 6.194 0.326 2.132
sion engineering (two experts). Initially 17 criteria were extracted Ps = + + (0.3 < z < 1)
z z2 z3 (4)
based on literature review and expert opinions. However, since it was
necessary to keep the number of criteria as small as possible 0.662 4.05 3.288
Ps = + 2 + (1 < z < 10)
(Malczewski, 1999), sensitivity analysis was performed to select the z z z3 (5)
most important criteria. Ultimately, 14 criteria were extracted for
Using expert opinions, the vulnerability criteria are divided into five
vulnerability evaluation (see Table 2). The constraints implemented on
classes (from ‘very low’ to ‘very high’) according to their values (see
the 1:2000 base maps of Tehran were exclusion of vacant lots and ex-
Table 2).
traction of buildings within the selected district. In addition, parks,
green spaces and any polygon features other than those representing
buildings were excluded. Overall, 28,228 parcels were extracted as 4.2. Scenario determination
selected features of vulnerability analysis.
The most important criterion is the intensity of explosion. This is Two parameters should be identified when defining scenarios. These
because the main threat stems from the wave produced immediately are location of explosion and strength of the explosive charge used
following the blast. The strength of this wave is a function of weight (equivalent weight of TNT explosive). As mentioned earlier, this study
equivalent of TNT explosive charge in kilograms (W) and distance of is focused on intentional explosions. Obviously, these targets are not
the core of explosive charge to the target building in meters (R). In this selected randomly. Instead, they are selected based on their degree of
paper, blast pressure in the air (overpressure (Ps)) is regarded as a importance. In this study information provided in the Passive Defense
criterion for measuring ‘intensity of explosion’, which is calculated chapter of the Iranian National Building Code (BHRC, 2009) is used for
according to Heinrich's equations. The main variable in these equations classifying buildings and other urban uses based on their degree of
is scaled distance (Z), calculated from Eq. (2). It can be seen that the importance. According to the Passive Defense chapter, buildings and
scaled distance is a function of W and R (Ngo, Mendis, urban uses can be categorized into five groups, based on their degree of
Gupta, & Ramsay, 2007): importance. These are, namely, ‘critical’, ‘sensitive’, ‘important’, ‘of low
importance’, and ‘of very low importance’ (see Table 1 of the Appendix
R
Z= for more details).
W1 3 (2)
Only vulnerability to explosions (with explosive charges of 1 k, 4 k,

105
Y. Ebrahimian Ghajari et al.
and 16 k TNT) targeting critical buildings and land uses is modeled and
evaluated in this study. These are the most common explosive weights
used by typical and frequently used explosive devices. It is worth noting
that modeling can be carried out by using different targets and for any
desired weight of TNT. However, this will result in many different
scenarios. In order to keep the study within manageable limits, here,
only the most feasible scenarios are analyzed. In other words, explosive
targets are critical building shown in Fig. 3. Vulnerability of these
targets to explosion events with equivalent explosive weights of 1 k,
4 k, and 16 k TNT is modeled and evaluated.
4.3. Criterion maps
Criterion maps were generated for each of the 14 vulnerability
criteria. Each of these maps show vulnerability, of the buildings in the
district, in terms of one of the 14 criteria shown in Table 2. Each cri-
terion map includes 28,228 building units derived from urban maps at a
scale of 1:2000. For each vulnerability criterion, vulnerability of the
buildings was classified into five levels, ranging from “very low” to
106
Cities 72 (2018) 102–114
Fig. 4. Steps of vulnerability evaluation.
“very high” (Table 2). It is obvious that using such linguistic variables
adds a certain level of uncertainty to the analysis, as it may not always
be easy to make a clear distinction between two different levels of
vulnerability. For instance, it may be possible to evaluate the vulner-
ability level of a specific building, to a given criterion, as either “high”
or “very high”. As a case in point, buildings that are 30-45 m high were
considered ‘highly’ vulnerable and buildings higher than 45 m were
considered ‘very highly’ vulnerable. Obviously, buildings that are 45 m
high belong to two different classes of vulnerability and a certain de-
gree of uncertainty will be inevitable when assigning the building to
either of these classes. These uncertainties can be, to some extent, ad-
dressed by using fuzzy sets. Utility of fuzzy sets can be explained by the
fact that the modeling phase of this analysis requires quantification.
This means that quantitative variables should be used instead of lin-
guistic (qualitative) variables. Quantitative variables used in this re-
search are triangular fuzzy numbers. Triangular fuzzy numbers are
particular types of fuzzy numbers that are defined using ordered triples
(l, m, n). In each fuzzy number, l is the lower bound, m is the medium
value and u is the higher bound, and membership function μ is
Y. Ebrahimian Ghajari et al. Cities 72 (2018) 102–114

Table 1
Characteristics of survey participants.

No Expertise Education level Age Affiliation

1 Urban planning M.Sc. 40 Tehran municipality

2 Urban planning M.Sc. 45 Road, Housing & Urban
Development Research Center
3 Urban planning Ph.D. 42 University of Tehran
4 Structural Ph.D. 59 K.N.Toosi University of
engineering Technology
5 Structural Ph.D. 49 Babol Noshirvani University of
engineering Technology
6 Structural M.Sc. 31 Tehran municipality
engineering
7 Explosion M.Sc. 45 MUT University
engineering Fig. 5. Membership function for vulnerability measurement (Vahdat, Smith, & Amiri,
8 Explosion M.Sc. 41 MUT University 2014).
engineering
9 Passive defense Ph.D. 31 Shahid Beheshti University
10 Passive defense Ph.D. 32 Shahid Beheshti University
criteria using experts' opinions (Table 3). Since uncertainties associated
11 Passive defense M.Sc. 38 MUT University with experts' subjective judgments should be taken into consideration,
12 Passive defense Ph.D. 35 Babol Noshirvani University of weight calculation using this technique is specifically important. Ex-
Technology perts used linguistic variables for pairwise comparison and for in-
13 Architecture M.Sc. 65 University of Tehran
dicating the relative importance of each criterion. As mentioned above,
14 Architecture Ph.D. 50 Iranian Construction
Engineering Organization this involves a certain degree of uncertainty that can be handled by
using the fuzzy set theory (Celotto, Loia, & Senatore, 2015). Fuzzy
Analytic Hierarchy Processing (AHP) was used for the purpose of this
calculated from Eq. (6) (Vahidnia, Alesheikh, & Alimohammadi, 2009): research (see Section 2 of the Appendix for more details on fuzzy AHP).
The experts (see Table 1) individually completed the pairwise ma-
⎧ (x − l) (m − l) l ≤ x ≤ m trix using linguistic variables given in Table 2 of the Appendix. After
M = (u − x ) (u − m) m ≤ x ≤ u
μ∼ collecting the expert opinions, a matrix composed of fuzzy geometric
⎨
⎩0 otherwise (6) means of elements corresponding to each expert's pairwise comparison
matrices was obtained. The basic fuzzy operators used in this study are
As illustrated in Fig. 5, in this study triangular fuzzy numbers
given in Eqs. (1) to (3) of the Appendix. Table 3 shows the final fuzzy
(ranging from ‘very low’ to ‘very high’ vulnerability), rather than real
pairwise comparison matrix.
numbers, are used to determine vulnerability levels of buildings.
There are various methods to weigh the criteria in a fuzzy AHP. Of
Vulnerability was modeled for three scenarios of 1 K, 4 K, and 16 K
those, fuzzy geometric mean was used in this research. The major ad-
equivalent weight of TNT explosive charges. The only criterion map
vantage of using geometric mean over arithmetic mean is that the in-
that varies in these scenarios is ‘explosion intensity’, which is a function
fluence of maximum and minimum values is reduced (Shepard, 2005)
of explosive charge and distance to targets with high likelihood of ex-
(see Section 2 of the Appendix for more details on fuzzy AHP).
plosion occurrence (these targets have been identified in previous
Final weights of the criteria are given in Table 3. Inconsistency ratio
stages and are shown in Fig. 3). The criterion maps for ‘explosion in-
was calculated for all pairwise comparison matrices, and it was always
tensity’ and also the other criterion maps are illustrated in Fig. 6.
lower than 0.1. Also, the inconsistency ratio for the final pairwise
comparison matrix (obtained by combining pairwise comparison ma-
4.4. Assigning weights to the criteria trices of the experts) was 0.074, indicating an acceptable level of con-
sistency between the experts' opinions.
Techniques for assigning weights to the criteria are generally di-
vided into four different types, namely, ranking, rating, pairwise com- 4.5. Combination of criterion maps
parison, and examining trade-offs (Malczewski, 1999). In this paper,
pairwise comparisons were used to calculate the relative weights of the After generating criterion maps, they should be aggregated using a

Table 2
Vulnerability criteria selected for analysis (Adapted with changes from Ebrahimian-Ghajari (2017)).

No Vulnerability criteria Vulnerability levels

Very low Low Moderate High Very high

1 Intensity of explosion (kg/cm2) < 0.7 0.7–2.8 2.8–4.9 4.9–7 >7

2 Surface area of the site (m2) > 500 400–500 300–400 200–300 < 200
3 Distance to gas station (m) > 500 400–500 300–400 150–300 < 150
4 Height of buildings (m) < 10 10–20 20–30 30–45 > 45
5 Number of adjacent buildings 0 1 2 3 >3
6 Distance to gas storage facilities (m) > 400 300–400 200–300 100–200 < 100
7 Building density (%) < 60 60–120 120–180 180–240 > 240
8 Distance to electrical facilities (m) > 250 200–250 150–200 75–150 < 75
9 Quality of buildings New house Under construction Common Restorable Old house
10 Distance to water facilities (m) > 200 150–200 100–150 50–100 < 50
11 Age of buildings (year) <5 5–15 15–25 25–35 > 35
12 Construction materials Concrete Metal Brick-Iron Masonry Others
13 Lot coverage ratio (%) 0–20 20–40 40–60 60–80 80–100
14 Width of pathways (m) > 45 30–45 15–30 10–15 < 10

107
Y. Ebrahimian Ghajari et al. Cities 72 (2018) 102–114

Intensity of explosion (1K TNT) Intensity of explosion (4K TNT) Intensity of explosion (16K TNT)

Age of buildings Quality of buildings Height of buildings

Construction materials Width of pathways Surface area of the site

Fig. 6. Criterion maps of vulnerability (Adapted with changes from Ebrahimian-Ghajari (2017)).

decision rule in order to obtain the final vulnerability map. For this 4.6. Defuzzification
purpose, the Fuzzy Weighted Linear Combination (Fuzzy WLC) was
used in this paper (Eq. (7)). Vulnerability of each building is calculated as the weighted mean of
the vulnerability levels of that building in term of the 14 selected cri-
n
∼ ∼ ∼ teria. Since both criteria weights and vulnerability levels are fuzzy
Vi = ∑ Wj ⊗ Rij ,
j=1 numbers, the overall vulnerability levels of the buildings will also be
i = 1, 2, …, m (7) fuzzy numbers. Rating buildings based on their vulnerability level is of
high interest for planners and policy makers. In order to generate the
∼ ∼
where (Vi ) is general vulnerability of building i, Wj is weight of criterion final vulnerability map and categorize buildings into five vulnerability
∼ classes, it is necessary to present the vulnerability levels as real num-
j, and Rij is vulnerability level of buildings i in criterion j. All of the
above calculations are conducted using basic fuzzy operators from Eqs. bers within the interval [0, 100], and defuzzification is needed for this
(4)–(7) of the Appendix. purpose. Defuzzification can be done using various methods. Ordered

108
Y. Ebrahimian Ghajari et al.
Distance from gas station Distance from gas storage facilities
Distance from water facilities
Building density
Weighted Averaging (OWA) operator is one of those methods that has
been used in this study (Yager, 1988). For more details on the OWA
operator, we refer the reader to Section 3 of the Appendix.
The behavior of the OWA operators can be explained using two
characteristics: (1) the degree of ORness (or risk-taking) and (2) the
range of trade-offs that occur between the criteria. The ORness degree
shows where the OWA operator stands between the min (AND) and max
(OR) operators. It indicates the extent of emphasis of decision makers
on higher or lower values of a set of criteria (Ebrahimian-Ghajari, 2017;
Vahedi, Alesheikh, & Honarparvar, 2014). In other words, it indicates to
what extent decision makers tend to take or avoid risk.
The higher the ORness, the more optimistic or risk taking is the
decision maker, and the lower the ORness, the more pessimistic or risk
Cities 72 (2018) 102–114
Distance from electrical facilities
Number of Adjacent buildings
Lot coverage ratio
Legend
Fig. 6. (continued)
avoiding is s/he. In general, high ORness values (> 0.5) imply that the
decision maker is risk taker or optimist, ORness value equal to 0.5 refers
to the neutrality of the decision maker, and low ORness values (< 0.5)
indicate that the decision maker is pessimistic or averts risk.
As mentioned earlier, defuzzification was performed using OWA
operators. In defuzzification, triangular fuzzy numbers should be con-
verted to crisp numbers (indicative of the vulnerability level of each
building) through weighted combination of the numerical values for l,
m, and u. Therefore, three ordered weights (v1, v2 and v3) are needed to
be calculated for different risk conditions. Vulnerability maps in this
study were produced for five different risk conditions varying from
ORness = 0 to ORness = 1. Table 4 shows ordered weights associated
with each ORness value.
109
Y. Ebrahimian Ghajari et al. Cities 72 (2018) 102–114

Table 3
Fuzzy pairwise matrix (see Table 3 of the Appendix for an extended version of this matrix).

No Vulnerability criteria Criteria number Criteria

fuzzy weights
1 2 … … 14

1 Intensity of explosion (1,1,1) (8.56,8.72,8.87) … … (8.56,8.72,8.87) (0.028,0.357,0.454)

2 Surface area of the site (0.11,0.11,0.12) (1,1,1) … … (0.26,0.35,0.55) (0.014,0.023,0.042)
3 Distance to gas station (0.12,0.13,0.14) (1.93,2.96,3.97) … … (1.30,2.33,3.34) (0.055,0.094,0.143)
4 Height of buildings (0.11,0.11,0.12) (1.93,2.96,3.97) … … (1.09,1.25,1.36) (0.031,0.054,0.093)
5 Number of adjacent buildings (0.11,0.11,0.12) (0.26,0.35,0.55) … … (0.11,0.12,0.14) (0.009,0.014,0.021)
6 Distance to gas storage facilities (0.12,0.12,0.14) (2.72,3.83,4.89) … … (1.09,2.10,3.11) (0.047,0.085,0.135)
7 Building density (0.11,0.12,0.12) (1.09,2.10,3.11) … … (0.26,0.35,0.55) (0.014,0.024,0.047)
8 Distance to electrical facilities (0.11,0.12,0.12) (1.19,2.21,3.22) … … (1.09,2.10,3.11) (0.045,0.079,0.126)
9 Quality of buildings (0.11,0.12,0.12) (1.09,2.1,3.11) … … (1.15,1.19,1.22) (0.029,0.049,0.086)
10 Distance to water facilities (0.11,0.12,0.12) (1.09,2.1,3.11) … … (0.19,0.24,0.31) (0.038,0.069,0.11)
11 Age of buildings (0.11,0.11,0.12) (0.32,0.47,0.92) … … (0.19,0.24,0.31) (0.012,0.019,0.033)
12 Construction materials (0.11,0.11,0.12) (6.12,7.12,8.12) … … (1.09,2.10,3.11) (0.038,0.063,0.107)
13 Lot coverage ratio (0.11,0.11,0.12) (1.09,2.10,3.11) … … (0.30,0.45,0.84) (0.014,0.023,0.047)
14 Width of pathways (0.11,0.11,0.12) (1.83,2.85,3.86) … … (1,1,1) (0.030,0.049,0.084)

Inconsistency ratio = 0.074.

Table 4 5.2. Statistical analysis of vulnerability

Ordered weights associated with various risk conditions (calculated by Authors).
As expected, the more destructive the explosive incident, the higher
ORness U M L
is vulnerability of buildings. Also, for a fixed amount of explosive
0 = AND 0 0 1 charge, increasing the risk-taking degree results in higher levels of
0.25 0.037 0.2593 0.7037 vulnerability. The degree of risk taking is the same as the degree of
0.5 0.3333 0.3333 0.3333 proximity between the OWA operator and the OR operator. Following
0.75 0.6934 0.1802 0.1264
gradual increase of the risk-taking degree, the number of buildings in
1 = OR 1 0 0
the ‘very low’ and ‘very high’ vulnerability level decreases and in-
creases, respectively. Fig. 8 shows statistical information on the extent
For example, as a fuzzy number such as v = (0.182,0.356,0.581) is of vulnerability of buildings under different scenarios (1 K, 4 K, and
defuzzified under various risk conditions, the following results are ob- 16 K TNT).
tained: As already mentioned in the introduction, give the growth in fre-
quency and intensity of human-made events, passive defense principles
ORness = 0: C = (0 ∗ 0.581) + (0 ∗ 0.356) + (1 ∗ 0.182) = 0.182 should be taken into consideration in different fields, including urban
ORness = 0.25: C = (0.037 ∗ 0.581) + (0.2593 ∗ 0.356) + (0.7037 design and planning. Disaster risk management activities, including
∗ 0.182) = 0.199 vulnerability evaluation, should be carried out in a way that prepares
ORness = 0.5: C = (0.3333 ∗ 0.581) + (0.3333 ∗ 0.356) + (0.3333 communities to deal with the worst-case scenarios. Such an approach
∗ 0.182) = 0.373 enhances the possibility of minimizing urban vulnerability. Our statis-
ORness = 07.5: C = (0.6934 ∗ 0.581) + (0.1802 ∗ 0.356) + (0.1264 tical analyses, depicted in Fig. 8, show that for ORness = 1, vulner-
∗ 0.182) = 0.49 ability is maximum. From the diagrams, it is clear that under this
ORness = 1: C = (1 ∗ 0.581) + (0 ∗ 0.356) + (0 ∗ 0.182) = 0.581 condition, most buildings in District No. 6 are of high vulnerability and
authorities in charge are warned to address this issue by taking ap-
propriate mitigation actions. This is further explained in the
5. Results and discussions
Conclusions section.

The main results of this research are: 1) maps showing physical

vulnerability of the study area to explosions with different blast load-
ings and under various risk conditions and 2) statistical analysis of the
percentage of vulnerable buildings and their various levels of vulner-
ability. Results of vulnerability modeling were presented for three dif-
ferent blast scenarios and under five different risk conditions.
Accordingly, a total number of 15 vulnerability maps were generated
(see Fig. 7).
5.1. Vulnerability maps
Based on equations mentioned in preceding sections, vulnerability
maps can be generated for various risk conditions (from the most pessi-
mistic condition to the most optimistic one). Vulnerability maps were
generated for three blast scenarios (with explosive charges of 1 K, 4 K, and
16 K TNT) and under five different risk conditions (ORness = 0, 0.25, 0.5,
0.75, 1) (see Fig. 7). In these maps, vulnerability level of each building is
shown using a real number that lies within the interval [0, 100]. Finally,
buildings are classified into five vulnerability levels: very low [0−20],
low [20–40], medium [40–60], high [60–80] and very high [80–100].
5.3. Sensitivity analysis
Sensitivity analysis is one of the most important steps in a multi-
criteria evaluation. It is performed to determine how making changes to
the inputs of the model affects its outputs (Ikonen, 2016). Obviously,
there are a variety of error sources that may affect the final outputs of
evaluation models and sensitivity analysis is conducted to check the
accuracy of the models. In a multi-criteria evaluation, sensitivity ana-
lysis is carried out by systematically modifying weights and values of
the criteria and analyzing how these modifications may affect the final
rating of the alternatives (Gal, Stewart, & Hanne, 2013). In other words,
when sensitivity of a multi-criteria evaluation model is analyzed,
variability of final outputs (robustness of outputs) would be evaluated.
Since the criteria are weighed using experts' opinions (involving sub-
jective judgments), errors are most likely to occur. If altering criteria
weights does not result in considerable changes in the final outputs of
the model, the model results can be trusted (Sadeghi-Niaraki,
Varshosaz, Kim, & Jung, 2011).
A commonly used method for sensitivity analysis is to increase or

110
Y. Ebrahimian Ghajari et al. Cities 72 (2018) 102–114

Fig. 7. Vulnerability maps in different and

under various risk conditions.

111
Y. Ebrahimian Ghajari et al. Cities 72 (2018) 102–114

1K TNT Very Low Low Moderate High Ver High

91

83.5
Very Low Low Moderate High Very High

94

91

69
54.5

54
68

64
53.5

42

36
44

25.5
34
30

10.5
6.5

5.5
4.5
3.5

3.5
3.5

2.5
1.5

1.5

1.5
0.5

0.5

0.5
5

2
2

1
0
0

0
0

0
0
Orness=0 Orness=0/25 Orness=0/5 Orness=0/75 Orness=1 Orness=0 Orness=0/25 Orness=0/5 Orness=0/75 Orness=1

Fig. 9. Statistical analysis of vulnerability in the 4 K TNT scenario, when ‘explosion in-
tensity’ increased by 10%.
4K TNT
Very Low Low Moderate High 86.5 Ver High Very Low Low Moderate High Ver High
94

93

88
68.5
54.5

68
53

55.5
44
43

45.5
51
27.5

42

28.5
9.5
2.5

2.5
1.5

1.5

1.5

1.5
1.5

1.5
0.5

4
1
0

0
0

4.5

3.5
1.5

1.5

1.5

8
0.5
Orness=0 Orness=0/25 Orness=0/5 Orness=0/75 Orness=1

2
1

1
0

0
0
Orness=0 Orness=0/25 Orness=0/5 Orness=0/75 Orness=1
16K TNT Fig. 10. Statistical analysis of vulnerability in the 4 K TNT scenario, when ‘explosion
intensity’ decreased by 10%.
Very Low Low Moderate High Very High
84

highest amount of variation is observed for the ‘intensity of explosion’

73
72
60.5

criterion (it is worth noting that this variation is within an acceptable

55

range, indicating the robustness of the model). Therefore, ‘intensity of

33.5
35

explosion’ is the most influential criterions, followed by ‘building

19
15

density’, ‘age of buildings’, ‘height of buildings’, and ‘lot coverage ratio’.

11
3.5
2.5

0.5

0.5
9

8
5
3
2
0

0
0

From the above results, it is clear that, the highest variations oc-
Orness=0 Orness=0/25 Orness=0/5 Orness=0/75 Orness=1
curred for ORness = 0.5. However, these variations are negligible.
Altering weights of for the other criteria resulted in far less variations,
Fig. 8. Statistical analysis of vulnerability to explosions with different explosive charges suggesting that the degree of stability of the model is acceptable and the
and under various risk conditions.
model is reliable.

decrease the weight of a given criterion by P%. Since the sum of criteria
weights should be equal to one, if the weight of one criterion is in-
creased (decreased), weights of the others should be decreased
(increased).These calculations are made using Eq. (8) (Eldrandaly,
2013).
n understand where the city stands in terms of vulnerability to potential
W (P ) = ∑ W (ci , P ) = 1
1 (8)
where, W(ci, P) is the weight of the ith criterion when the amount of
variation is by P%. If the mth criterion varies by P%, then its new value
is obtained by Eq. (9):
W (cm , P ) = W (cm , 0) + W (cm , 0) × P , 1 ≤ m ≤ n (9)
And weights of other criteria are derived from Eq. (10):
W (c i,P ) = (1 − W (cm , P )) × W (ci , 0) (1 − W (cm , 0), i ≠ m , 1 ≤ i ≤ n
(10)
Forty two sets of sensitivity analyses were performed using Eqs. (8)
to (10). For each of the three scenarios, all criteria (14 criteria) were
modified by 10%, resulting in new vulnerability maps with new
weights. The highest and lowest variations, as compared to primary
conditions, were observed for the ‘intensity of explosion’ and the
‘number of adjacent neighbors’, respectively. To save space, only results
of sensitivity analysis for ‘intensity of explosion’ (under the 4 K TNT
scenario) are presented here (Figs. 9 and 10).
As mentioned, according to the results of the sensitivity analysis, the
6. Conclusions
Vulnerability analysis and mapping are widely recognized as es-
sential tools for enhancing awareness and knowledge of planners and
decision makers about baseline conditions. These maps can be used to
disasters and what factors contribute to vulnerability. By analyzing
vulnerability maps planners and decision makers can develop strategies
and take appropriate actions to improve conditions, minimize vulner-
ability, and enhance resilience. Developing such strategies and action
plans facilitates improved preparation and planning for disaster risk
management. Vulnerability maps can also be used for site selection and
for developing land use management plans. In the absence of vulner-
ability maps it will be difficult, if not impossible, for planners and de-
cision makers to: 1) enhance resilience and adaptive capacity of critical,
sensitive, and important structures and facilities, 2) facilitate urban
crisis management, 3) ensure community safety and well-being, 4)
develop disaster mitigation plans, and 5) minimize costs and maximize
efficiency of actions.
Vulnerability evaluation of buildings in District No. 6 of Tehran
shows that, assuming optimistic risk perceptions (ORness = 1), for all
blast scenarios almost 70% of the buildings are highly vulnerable. This
clearly indicates that the district is highly prone to accidental and/or
intentional explosion incidents. This paper argues that it is essential to
reduce the physical vulnerability of the district through various pre-
paratory efforts. The results can be used to enhance awareness about

112
Y. Ebrahimian Ghajari et al.
the extent of vulnerability (of buildings in the study area) under dif-
ferent risk scenarios. Further analysis of the results provides insights
about the underlying causes of these vulnerabilities. Building density of
the district is higher than the city average and this is argued to be a
major reason behind the high levels of vulnerability. Other noteworthy
reasons are the; dominance of old building stock; lack of urban open
spaces; failure of the city officials to regulate building construction
activities through enforcing building codes (Sarkheyli, Sharifi, Rafieian,
Bemanian, & Murayama, 2012); and lack of attention to passive defense
principles in urban planning and design. Therefore, integrated efforts
across different sectors should be taken to develop and implement ac-
tion plans for reducing vulnerability of buildings to a wide array of
threats including explosion hazards. Below, further policy implications
are discussed in light of study findings.
6.1. Policy implications
An important issue highlighted in this study is the high concentra-
tion of sensitive and critical places and structures such as embassies,
ministry buildings, hospitals, corporate headquarters, and higher edu-
cation centers in the study area (see Fig. 3). Such a high degree of
centralization potentially increases the likelihood of occurrence of ex-
plosion incidents in the district. Current land use plan should be reex-
amined with the objective of developing strategies for achieving a more
balanced distribution of building and uses with critical importance
across the city. This may require adjustment of existing land use plans
and relocation of some of the uses and activities to other districts in the
city.
Results of the sensitivity analysis indicate that ‘intensity of explo-
sion’, followed by ‘building density’, ‘building age’, ‘building height’,
and ‘lot coverage ratio’ are the most determining criteria that con-
tribute to physical vulnerability of the study area. Some policy mea-
sures that need to be taken in order to deal with vulnerability related to
these criteria are discussed below.
Policy measures designed to deal with ‘intensity of explosion’ can be
either preventative or impact-minimizing. These may differ depending
on the origin of explosion. Explosions can be of either intentional or
accidental origin. Preventing intentional explosions that occur because
of air raids and missile attacks requires installing military defense
systems which is not the responsibility of urban planners and decision
makers. What planners and urban authorities can do is enhancing the
likelihood of preventing detonation of explosive devices planted by
terrorists through improving urban surveillance. This could be achieved
through promoting Jane Jacobs' “eyes on the street” design principles
(Jacobs, 1961) and/or monitoring public spaces using closed-circuit
security cameras. Preventing unintentional explosions (e.g. gas sta-
tions) requires regular maintenance and performance monitoring. This
is again not the responsibility of urban planners and designers. While
having limited authority to prevent explosions, urban planners and
designers can play an important role in minimizing potential adverse
impacts of accidental explosion incidents. Adequate security buffer
distances should be provided when siting critical facilities such as gas
stations. Other noteworthy measures include provision of adequate
open spaces and refurbishment and reinforcement of infrastructures to
increase their resistance. These are further explained below in the light
of study findings.
High building density and lot coverage were found to be important
factors contributing to physical vulnerability of the district. Given the
frequency of explosion targets in the district, high-density development
would result in the exposure of a large population to risk. It may also
hamper emergency services following potential explosions. A more
moderate building density can provide opportunities for incorporating
more open and green spaces. The important role of open spaces for
absorbing shocks by functioning as a buffer against the spread of risk
and providing multi-functionality (e.g. sheltering and evacuation pur-
poses) has been emphasized in the literature (Ishikawa, 2002;
113
Cities 72 (2018) 102–114
Leon & March, 2014; Villagra, Rojas, Ohno, Xue, & Gomez, 2014). More
open spaces can be provided by complying with zoning regulations
related to lot coverage ratio. In Tehran, these regulations and also
density and Floor Area Ratio (FAR) regulations are often violated by
developers (Sarkheyli et al., 2012). Various socio-economic and in-
stitutional factors contribute to non-compliance with zoning and
building regulations in Tehran. These include, but are not limited to,
developers' lack of awareness of building regulation, prevalence of
speculative activities in the housing market, loose control over con-
struction activities, and municipality's dependence on revenues gener-
ated from the so-called ‘density selling’ policy (Sarkheyli et al., 2012).
This policy allows municipality officials to waiver regulations and
award high bonus densities to those developers who wish to purchase
development rights. To ensure compliance with zoning and building
regulations, Tehran Municipality should improve monitoring and su-
pervision, enhance awareness about regulations, develop strategies to
deal with speculative activities, and develop sustainable sources of
revenue (to replace the ‘density selling’ policy which lures developers to
violate regulations).
Building age was another major factor contributing to physical
vulnerability. The study area is one of the oldest districts in Tehran with
many old buildings. Such old buildings are less resistant to blast waves.
Urban authorities should prioritize reinforcing old buildings to make
them more resistant to blast waves produced by explosive devices.
This study used scenario building for vulnerability analysis.
Scenario building is an effective instrument for simulating performance
in the event of disaster. It informs decision makers about preparatory
actions that need to be taken under different risk conditions. In this
research scenarios were built for explosions with different levels of
strength (i.e. caused by 1, 4, and 16 K TNT as the most common ex-
plosive charges). Physical vulnerability to these explosions was esti-
mated depending on the risk perceptions of decision makers (ranging
from most pessimistic to most optimistic). This approach allows us to
simulate different levels of damage that the district may experience
during potential explosion incidents and to take proportional pre-
paratory actions. Ideally, cities need to get prepared to deal with the
most devastating scenario. That is the scenario when the most powerful
explosive charge (16 K TNT) is used and it is expected to result in
maximum damage (ORness = 1). Preparatory action plans developed
based on this scenario would minimize potential impacts of explosion in
the district. However, due to time and budget constraints, maximum
preparation may not always be possible. Access to results of other
scenarios enables planners and decision makers to prioritize actions and
choose appropriate planning and preparation strategies depending on
resources at their disposal. For instance, due to extreme shortage of
resources, it may be decided to use the most optimistic scenario (ex-
plosive charge of 1 k TNT and ORness = 0). As Fig. 7 shows, even
under this scenario, there are buildings that are highly vulnerable. The
priority should be given to reinforcing these buildings. In a similar vein,
it would be possible to use less optimistic scenarios (between the most
optimistic and most pessimistic scenarios, see Fig. 7) to identify other
buildings that should be reinforced when more resources become
available.
6.2. Suggestions for future research
As already mentioned, research on modeling vulnerability to
human-made threats in general, and explosion incidents in particular, is
limited. This study is an initial effort towards addressing this gap in
research. Of different urban physical components, only buildings were
studied in this study. Future work should also evaluate vulnerability of
other vital physical components such as transportation, communica-
tion, and other critical infrastructure. In addition, social, institutional,
economic, and environmental vulnerability to human-made threats
should also be investigated in the future. This research was highly re-
lied on experts' opinions for extraction of criteria and assigning weights
Y. Ebrahimian Ghajari et al.
to them. As a result it should be acknowledged that, as with all research
relying on data extracted from expert judgments, the study involves a
certain extent of subjectivity and uncertainty. Other methods such as
those developed based on group decision making techniques can help
achieve better and more actual results. Uncertainty is a broad concept,
and more comprehensive research should be conducted in order to
obtain results that are more accurate. Further research on uncertainty
theories can provide strategies for reducing uncertainty. For instance,
recently there have been new developments in the area of fuzzy logic
systems and concepts such as Type-2 fuzzy sets and intuitionistic fuzzy
sets have emerged. It is recommended that further research should be
undertaken to investigate utility of these new concepts for multi-criteria
evaluation of urban vulnerability.
The model introduced in this paper can be adjusted and extended in
future work by incorporating additional parameters to estimate how
proposed new developments in the district (e.g. constructing new
buildings and infrastructures) can reduce or increase vulnerability of
adjacent buildings. Such information would be useful for making de-
cisions about whether or not to permit proposed new developments.
This study was only focused on physical aspects of urban vulnerability.
Future research should also take social and economic aspects into ac-
count. More parameters can be added to the model to estimate potential
social (e.g. number of people exposed to risk) and economic (e.g.
monetary losses) impacts associated with physical vulnerability of
buildings and develop strategies to plan/prepare for and absorb them in
an appropriate way.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.cities.2017.08.006.
References
Adger, W. N. (2006). Vulnerability. Global Environmental Change, 16, 268–281.
Armenakis, C., & Nirupama, N. (2013). Estimating spatial disaster risk in urban en-
vironments. Geomatics, Natural Hazards and Risk, 4, 289–298.
Azizi, M. M., & Bornafar, M. (2012). Assessing urban vulnerability due to air attacks:
District 1 of region 11 of Tehran municipality (in Persian). Journal of Advanced
Defence Science and Technology, 3, 127–137.
BHRC (2009). Iranian national building code (part 21): Passive defence. Iran: Building and
Housing Research Center.
Buchert, T., Neugebauer, S., Schenker, S., Lindow, K., & Stark, R. (2015). Multi-criteria
decision making as a tool for sustainable product development – Benefits and ob-
stacles. Procedia CIRP, 26, 70–75.
Celotto, A., Loia, V., & Senatore, S. (2015). Fuzzy linguistic approach to quality assess-
ment model for electricity network infrastructure. Information Sciences, 304, 1–15.
Chmutina, K., Lizarralde, G., Dainty, A., & Bosher, L. (2016). Unpacking resilience policy
discourse. Cities, 58, 70–79.
Coaffee, J. (2009). Terrorism, risk and the global city towards urban resilience. Surrey:
Ashgate.
de la Fuente, A., Pons, O., Josa, A., & Aguado, A. (2016). Multi-criteria decision making in
the sustainability assessment of sewerage pipe systems. Journal of Cleaner Production,
112, 4762–4770.
Desouza, K. C., & Flanery, T. H. (2013). Designing, planning, and managing resilient
cities: A conceptual framework. Cities, 35, 89–99.
Dobbs, N., United States, Department of Defense, & Explosives Safety Board (1984).
Structures to resist the effects of accidental explosions. Picatinny Arsenal, N.J.: U.S. Army
Armament Research, Development, and Engineering Center.
Ebrahimian-Ghajari, Y. (2017). Development and implementation of a GIS-based model to
evaluate urban physical vulnerability to human-made threats (in Persian) (Unpublished
doctoral dissertation)Tehran, Iran: Malek Ashtar University of Technology.
Ebrahimian-Ghajari, Y., Alesheikh, A., Modiri, M., Hosnavi, R., & Abbasi, M. (2017).
Spatial modelling of urban physical vulnerability to explosion hazards using GIS and
Fuzzy MCDA. Sustainability, 9(7), 1274. http://dx.doi.org/10.3390/su9071274.
Ebrahimian-Ghajari, Y., AleSheikh, A. A., Modiri, M., Hosnavi, R., & Nekouei, M. A.
(2016). Modeling of seismic vulnerability of urban buildings in geographic in-
formation system environment: A case study in Babol, Iran. Journal of Rescue and
Relief, 7, 12–25.
Eldrandaly, K. A. (2013). Exploring multi-criteria decision strategies in GIS with linguistic
quantifiers: An extension of the analytical network process using ordered weighted
averaging operators. International Journal of Geographical Information Science, 27,
2455–2482.
114
Cities 72 (2018) 102–114
Fernandez, P., Mourato, S., & Moreira, M. (2016). Social vulnerability assessment of flood
risk using GIS-based multicriteria decision analysis. A case study of Vila Nova de Gaia
(Portugal). Geomatics, Natural Hazards and Risk, 7, 1367–1389.
Gal, T., Stewart, T., & Hanne, T. (2013). Multicriteria decision making: Advances in MCDM
models, algorithms, theory, and applications. Springer Science & Business Media.
Gillen, M. (2005). Urban governance and vulnerability: Exploring the tensions and con-
tradictions in Sydney's response to bushfire threat. Cities, 22, 55–64.
Hashemi, M., & Alesheikh, A. A. (2011). A GIS-based earthquake damage assessment and
settlement methodology. Soil Dynamics and Earthquake Engineering, 31, 1607–1617.
Ikonen, T. (2016). Comparison of global sensitivity analysis methods – Application to fuel
behavior modeling. Nuclear Engineering and Design, 297, 72–80.
Ishikawa, M. (2002). Landscape planning for a safe city. Annals of Geophysics, 45.
Jacobs, J. (1961). The death and life of great American cities. New York: Vintage books.
Kapucu, N. (2012). Disaster and emergency management systems in urban areas. Cities,
29, S41–S49.
Karashima, K., Ohgai, A., & Saito, Y. (2014). A GIS-based support tool for exploring land
use policy considering future depopulation and urban vulnerability to natural dis-
asters – A case study of Toyohashi City, Japan. Procedia Environmental Sciences, 22,
148–155.
Kim, H., & Marcouiller, D. W. (2015). Urban vulnerability and resiliency to natural dis-
asters: An integrative tourism planning perspective. Cities at Risk: Planning for and
Recovering from Natural Disasters, 159.
Krauthammer, T. (2008). Modern protective structures. CRC Press.
Kuin, N., Masthoff, E., Kramer, M., & Scherder, E. (2015). The role of risky decision-
making in aggression: A systematic review. Aggression and Violent Behavior, 25(Part
A), 159–172.
Kulawiak, M., Lubniewski, Z., Bikonis, K., & Stepnowski, A. (2009). Geographical in-
formation system for analysis of critical infrastructures and their hazards due to
terrorism, man-originated catastrophes and natural disasters for the city of Gdansk.
Information fusion and geographic information systems, proceedings (pp. 251–262). .
Leon, J., & March, A. (2014). Urban morphology as a tool for supporting tsunami rapid
resilience: A case study of Talcahuano, ChileD. Habitat International, 43, 250–262.
Ma, L., Cheng, L., & Li, M. C. (2013). Quantitative risk analysis of urban natural gas
pipeline networks using geographical information systems. Journal of Loss Prevention
in the Process Industries, 26, 1183–1192.
Malczewski, J. (1999). GIS and multicriteria decision analysis. John Wiley & Sons.
Matijosaitiene, I., & Petriashvili, A. (2017). Urban planning and design for terrorism re-
silient cities. Journal of Sustainable Architecture and Civil Engineering, 18, 27–38.
McDonald, R. (2003). Introduction to natural and man-made disasters and their effects on
buildings. Routledge.
Ngo, T., Mendis, P., Gupta, A., & Ramsay, J. (2007). Blast loading and blast effects on
structures–An overview. Electronic Journal of Structural Engineering, 7, 76–91.
Parnell, S. (2016). Defining a global urban development agenda. World Development, 78,
529–540.
Prior, T., & Eriksen, C. (2013). Wildfire preparedness, community cohesion and social-
ecological systems. Global Environmental Change-Human and Policy Dimensions, 23,
1575–1586.
Sadeghi-Niaraki, A., Varshosaz, M., Kim, K., & Jung, J. J. (2011). Real world re-
presentation of a road network for route planning in GIS. Expert Systems with
Applications, 38, 11999–12008.
Sarkheyli, E., Sharifi, A., Rafieian, M., Bemanian, M. R., & Murayama, A. (2012). An
investigation of the reasons for non-compliance with FAR regulations in Tehran.
Cities, 29, 223–233.
Sharifi, A., & Yamagata, Y. (2016). On the suitability of assessment tools for guiding
communities towards disaster resilience. International Journal of Disaster Risk
Reduction, 18, 115–124.
Shepard, R. B. (2005). In R. B. Shepard (Ed.), Quantifying environmental impact assessments
using fuzzy logic. New York: Springer.
Smith, P. D., & Hetherington, J. G. (1994). Blast and ballistic loading of structures.
Butterworth-Heinemann.
Tadepalli, T., & Mullen, C. L. (2006). Simplified blast simulation procedure for hazard
mitigation planning. Structures under shock and impact IX. 87. Structures under shock
and impact IX (pp. 55–63).
Vahdat, K., Smith, N. J., & Amiri, G. G. (2014). Fuzzy multicriteria for developing a risk
management system in seismically prone areas. Socio-Economic Planning Sciences, 48,
235–248.
Vahedi, B., Alesheikh, A. A., & Honarparvar, S. (2014). Quantitative assessment of
pragmatic quality of volunteered geographic information using fuzzy linguistic
quantifiers and OWA operator. Journal of Geomatics Science and Technology, 3, 65–76.
Vahidnia, M. H., Alesheikh, A. A., & Alimohammadi, A. (2009). Hospital site selection
using fuzzy AHP and its derivatives. Journal of Environmental Management, 90,
3048–3056.
Villagra, P., Rojas, C., Ohno, R., Xue, M., & Gomez, K. (2014). A GIS-base exploration of
the relationships between open space systems and urban form for the adaptive ca-
pacity of cities after an earthquake: The cases of two Chilean cities. Applied
Geography, 48, 64–78.
Williams, G., Batho, S., & Russell, L. (2000). Responding to urban crisis - the emergency
planning response to the bombing of Manchester city centre. Cities, 17, 293–304.
Yager, R. R. (1988). On ordered weighted averaging aggregation operators in multi-
criteria decisionmaking. IEEE Transactions on Systems, Man, and Cybernetics, 18,
183–190.
Zhang, X. R., Wang, Z. B., & Lin, J. (2015). GIS based measurement and regulatory zoning
of urban ecological vulnerability. Sustainability, 7, 9924–9942.