Danish Hydraulic Institute

RED SEA HINDCAST STUDY CONTRACT NO. 45057/00/01

Study Report February 1997
SAER 5565

SAUDI ARAMCO

CSD-I.-261197

:.

, - .

,

.. _

cc:

V.P., Engineering Servi&s. _ . : Coord.; M&CED.;. . . SUPM., CEU/M&CED. _. . . MAM . .
. i

.

.

:.-:. . ... ;.
. '

. .

..
...
.'

,.

.i

.

.. . .
. . .

..
.
.:

UB

. .. .. .
.
,

..

,

:

.

.
. .

.,

..

,

:.
i

' .

' .

'

-.!

: s,

.

, .

'

,

,

.

0
.,

,.

, I I

..

. .

,

.

.......

. _ . ..
, i .i

. ...1
-

.:
1

:

. . .

' .
. .

. -

...
I .

. . ... . . . . ......

-.3 . ;

.. .
i:.,

CSD-L-261197

RED SEA EWDCAST STUDY
_I

./-.

. . 8. .

.?.: .

.. :_ . . . .. . ... _.-

DISTRIBUTION LIST

: . . . .. ,. .. .. .

~

. ;j
:

0

Manager, Distribution Operations::(We+th Reg.) Dept., Manager, Yanbu Gas/Ter;ninal,Operatio&: Dept.,.do.T-8 Manager, Jeddah Manager, Rabigh Manager, Maiine Dep Project Manager, West Project Manager, Jzan Port Captain, Terminal Pd General Supervisor,'Te General Supervisor, Enviio Work Director, Terminals SupeMsor, Hydrographic Sum Technical Information Center,
_ .
. .
, I ,

.

,:.

. I

,

. . .. . . . .

. .. . ,.. . . . . ..'<.. . . .

...I . ,
1.,T

. . . . .. . ,,
... .

.

.

-

. - . - r ..

. .. .

. ,,.

.... ... ..;..
: '
.

,>

:.:'?

,

I.L.

... . . . .. .. ... . . .

. . .. . .- . ... .,.;:: .
I :

.

i

.- .- . . . . '..> . . ,,.,...,-... .. .. ... .
1

....

Danish Hydraulic Institute Red Sea Hindcast Study
Final Report
Agern All6 5, DK-2970 Hersholm, Denmark Telephone: +45 45 76 95 55 Telefax: +45 45 76 25 67 Telex: 37 402 dhicph dk

Distribution: Saudi Ararnco x 20 DRK x 1 PHD x 1
12 February 1997

Client

Saudi Aramco

Dr. Motaz AI-Mashouk

'roject No

766815070
Jate

1 2 February 1 9 9 7
4pproved by

2
1

Final R e p ~ n including project extensions

A
(

A

1212.97

Final Report
Draft (issued for cornrnenrsl
Descti~tion

0
RBVISID"

ORK
BY

VJ
ChBCksd

3/4-96
APP.
DBtO

:ey words

Red Sea Hindcast Wave Modelling Current Modelling Wind Modelling
ill-0A.DRIMIC9608.ia0

0 Internal
Proprietary

i

CONTENTS
1 2 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . OVERVIEW OF THE RED SEA HINDCAST STUDY 1-1 2-1

..........

3
4
4.1
a 5

REVLEW OF AVAILABLE DATA . . . . . . . . . . . . . . . . . . . . . . . DIGITIZATION OF BATHYMETRIES

3-1
4-1 4-1
5-1

....................

General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . STATISTICAL ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . WINDSTUDY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Normal Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis of Normal Wind Conditions

6

6-1 6-1 6-2
6-4

6.1
6.1.1 6.2 6.2.1
7

.......................

Extreme Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis of Extreme Winds

.............................

6-5 7-1 7-1 7-1 7-1 7-2 7-13

CURRENT AND LEVEL HYDRODYNAMIC MODELLING . . . . . .
Overview

7.1 7.2 7.2.1 7.2.2 7.2.3 7.3 7.3.1 7.3.2 7.3.3

........................................

Model Set-up

.....................................

The Numerical Model - MIKE 21 HD . . . . . . . . . . . . . . . . . . . . . .
Bathymetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Normal Currents and Levels - Astronomical Tide Only . . . . . . . . . . . . . 7-13 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Calibration of Normal (Tidal) Currents and Levels Normal Current and Level Simulations 7-13

. . . . . . . . . . . . . . . . 7-15
7-23 7-25

.......................

0

7.3.4

Analysis of Normal (Tide) Water Levels . . . . . . . . . . . . . . . . . . . . . .

11

..

0

7.3.5 7.4 7.4.1 7.4.2 7.4.3 7.4.4 7.4.5

Analysis of Normal (Tide) Currents

.........................

7-26 7-27 7-27

Extreme Currents and Levels .30 Storms . . . . . . . . . . . . . . . . . . . . . Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Calibration of Extreme (Storm) Currents and Levels . . . . . . . . . . . . . . . 7-27 Extreme (Storm) Current and Level Simulations Analysis of Extreme (Storm) Water Levels Analysis of Extreme (Storm) Currents

. . . . . . . . . . . . . . . . . 7-29
7-29 7-31 8-1 8-1 8-1 8-1 8-2 8-3 8-3 8-3 8-3

....................

.......................

a

8
8.1 8.2 8.2.1 8.2.2 8.3 8.3.1 8.3.2 8.3.3 8.4 8.5 8.5.1 8.5.2 8.5.3 8.6 8.6.1 8.6.2 8.6.3

REGIONAL WAVE MODELLING . . . . . . . . . . . . . . . . . . . . . . .
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Numerical Model .OWI3G Model Outputs

..........................

.....................................

Wave Parameter Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . Modelsetup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

OWI3G Model Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Spectral Discretization

................................

Regional Wave Model Boundary Conditions . . . . . . . . . . . . . . . . . . . Calibration and Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Normal Conditions

8-5
8-5 8-5 8-5 8-6 8-6

..................................

Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Normal Wave Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results

.........................................

Extreme Conditions .30 Storms

..........................

Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8-6
8-7 8-7

iii

a

9

NEARSHORE WAVE MODELLING . . . . . . . . . . . . . . . . . . . . .
Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Normal Conditions Extreme Conditions

9-1 9-1 9-1 9-1 9-1 9-2 9-2 9-3 9-5 9-5

9.1 9.1.1 9.1.2 9.2 9.3 9.3.1

..................................

..................................
....................

The Numerical Model . MIKE 21 NSW

Modelsetup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General

......................................... ......................................

a

9.3.2 9.4 9.4.1 9.4.2 9.5 9.5.1 9.6 9.6.1 9.6.2 9.6.3 9.6.4

Bathymetries

Normal Wave Conditions

..............................

Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transformation from Offshore to Nearshore Extreme Wave Conditions Simulations

. . . . . . . . . . . . . . . . . . . 9-8
9-11 9-11 9-13 9-13 9-13 9-14 9-19
10-1

...............................

........................................

Verificatiodcalibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simulations Conclusion

........................................
........................................

10
10.1 10.2 10.3 10.4 10.5 10.6

DATABASE

......................................

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mapping and Grids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selection of Grid Points

10-1

10-1 10-2 10-2 10-2 10-3

...............................

Reliability of the Database Results . . . . . . . . . . . . . . . . . . . . . . . . . Discussion of Presented Results

.......................... ........................

File Locations. Types and Formats

iv

e

10.6.1 Data Directories . . . . . . . .

. . . . . . . . . . . . . . . . . 10-3
10-6
11-1 11-1

10.6.2 File Types and Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1 1

RED SEA EXTENSION .NEW BATHYMETRIES . . . . . . . . . . . . New Bathymetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effect of the new Bathymetries on the Hydrodynamic Conditions . . . . . .

11.1

11.2
11.3

11-4

Effect of the New Bathymetries on the Near-Shore Wave Conditions . . . 11-7

11.3.1 DUBA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11-7 11-11

0

11.3.2 JIZAN
12

..........................................

RED SEA EXTENSION - DIFFRACTION STUDY . . . . . . . . . . . . 12-1 Introduction

12.1 12.2 12.3 12.4 12.5 12.6

......................................

12-1 12-1 12-3 12-3 12-7 12-19 13-1 14-1

Study Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wave Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . REFERENCES

13
14

....................................

NOTATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

APPENDIX APPENDIX APPENDIX APPENDIX APPENDIX

A .Drawings B .Oceanweather. Inc . Final Report C .DHI’s EVA Program D .Short Description MIKE 21 HD E .Short Description MIKE 21 NSW
DATABASE USER GUIDE

ANNEX 1 .

Danrk Hydraulirk lnrtnutmnish Hydraulic lmlil~le

V

0

LIST OF FIGURES
Fig. Fig. Fig. Fig. Fig. 2.i 2.2 4.1 5.1 5.2 Location map of local operational areas. Location map of offshore reference points and regions. Digitized Red Sea bathymetry at a grid resolution of 75Om. EVA output. Example showing extreme winds offshore Duba. Comparison of normalized 1-year estimates of directional significant wave heights. Offshore Duba. Comparison of normalized 1-year estimates of directional significant wave heights. Offshore Yanbu. comparison of normalized 1-year estimates of directional significant wave heights. Offshore Rabigh. Comparison of normalized 1-year estimates of directional significant wave heights. Offshore Jeddah. Comparison of normalized 1-year estimates of directional significant wave heights. Offshore Jizan. Example of plot showing exceedence probabilities for northern and southern sector. Data from offshore point at Yanbu. Comparison of the predicted versus measured wind statistics at Jizan. Comparison of the predicted versus measured wind statistics at Yanbu. Duba local 1500 m bathymetry. Combined Yanbu, Rabigh, and Jeddah local 1500 m bathymetry, with individual areas indicated. Yanbu local 1500 m bathymetry (excerpt from "Cornbind" bathymetry). Rabigh local 1500 m bathymetry (excerpt from "Combined" bathymetry). Jeddah local 1500 m bathymetry (excerpt from "Combined" bathymetry). Jizan local 1500 m bathymetry. 6000 rn resolution Red Sea bathymetry Comparison of bathymetries at Bab-el-Mandeb. a) as digitized from charts b) modified bathymetry. Bab-el-Mandeb predicted water levels and Straits of Bab-el-Mandeb calibrated (model) water levels, 14 days comparison. Contour map of Manning resistance for regional Red Sea tidal model. Duba, Yanbu al Bahr, and Madinat Yanbu; comparison of model results from regional and local models (virtually coinciding), and comparison of model results to tidal station predicted water levels. Rabigh, Jeddah, and Jizan; comparison of model results from regional and local models (virtually coinciding), and comparison of model results to tidal station predicted water levels. Rabigh local area, January 1990. Calculated current field (example) and associated tidal variations. Rabigh local area, 9302 storm. Calculated current field (example) and time series of water level surge. Comparison of hindcast spectra versus JONSWAP spectra.

Fig. 5.3 Fig. 5.4 Fig. 5.5

Fig. 6.1 Fig. 6.2 Fig. 7.1 Fig. 7.2 Fig. 7.3 Fig. 7.4 Fig. 7.5 Fig. 7.6 Fig. 7.7 Fig. 7.8 Fig. 7.9 Fig. 7.10 Fig. 7.11 Fig. 7.12 Fig. 7.13 Fig. 7.14 Fig. 8.1

vi

a

Fig. 8.2 Fig. 9.1 Fig. 9.2 Fig. 9.3 Fig. Fig. Fig. Fig.
9.4 9.5 9.6

9.7

Fig. 9.8

0

Fig. 9.9 Fig. 9.10 Fig. 11.1 Fig. 11.2 Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig.
11.3 11.4 11.5 11.6 11.7 11.8 11.9 11.10 11.11 11.12 12.1

Fig. 12.2 Fig. 12.3 Fig. 12.4 Fig. 12.5 Fig. 12.6 Fig. 12.7 Fig 12.8 Fig 12.9 Fig. 12.10 Fig. 12.11

0

Grid discretization used by O W . (Grid spacing 0.25 degree latitude and longitude). Size and locations for the 12 models for Yanbu. Wave field, Yanbu. Hs, offshore = 1.Om Tm = 4s. Wave direction = 300". Wave Field, Yanbu. Hs, offshore = 1 . h Tm = 10s. Wave direction = 300" 50 year simulation. Wave direction 180" 50 year simulation. Wave direction 90" (wind from land) Measuring Positions at Yanbu. Timeseries of Hs and Tp at offshore boundary point. Verification period. Measured and calculated wave height and wave period at Shib Green. Measured and calculated wave height and wave period at A1 Nakla. Hs(m) in the Yanbu area for offshore wave conditions Hs, = 2Sm, Tp = 7.9 s, wave direction = 315" and wind speed = 12 d . s Difference between new and old bathymetry at Jizan. (Gridspacing =25Om). Difference between new and old bathymetry at Duba. (Gridspacing =25Om). Difference between new and old maximum current speeds at Jizan. Difference between new and old maximum current speeds at Duba. Wave Rose plot for normal conditions offshore of Duba. Hs,,, - HsOld(m), 50 year wave conditions, Wave Direction 120'. Hs,,, - HsOld(m). 50 year wave conditions, Wave Direction 210'. Hs,,, - HsOld(m), 50 year wave conditions, Wave Direction 30 . 0' Wave rose plot for the normal conditions offshore of Jizan. Hs,,, - HsOld(m), 50 year wave conditions, Wave Direction 90'. Hs,, - HsOld(m), 50 year wave conditions, Wave Direction 210' Hs,, - HsOld(m), 50 year wave conditions, Wave Direction 300" Bathymetry of the diffraction study area. (Gridspacing,dx =dy =7.51-4. Relative significant wave height using combined EMS results. Schematized bathymetry. Relative significant wave height using NSW fine grid model with no schematization. Relative significant wave height using NSW fine grid model using schematized bathymetry. Relative significant wave height using the BW model. Location map of line sections in model domain. Line section comparisons between the BW model and the fine grid NSW model for cross sections in the X-direction. Line section comparisons between the BW model and the fine grid NSW model for cross sections in the Y-direction. Relative significant wave height using the coarse grid NSW model. Line section comparisons between the BW model and the coarse grid NSW model including wind and dissipation for cross sections in the Xdirection.

vii

a

Fig. 12.12 Fig. 12.13 Fig. 12.14

Line section comparisons between the BW model and the coarse grid
NSW model including wind generation and dissipation for cross sections in the Y-direction. Line section comparisons between the BW model and the coarse grid NSW 'model results as prepared similarly as done for the database for cross-sections in the X-direction. Line section comparisons between the BW model and the coarse grid NSW model results prepared similarly as done for the database for cross sections in the Ydirection.

viii

a

LIST OF TABLES
Table 4.1 Table 5.1 Table 6.1 Table 6.2 Table 6.3 Table 7.1 Table 7.2 Table 7.3 Table 7.4 Table 7.5 Table 8.1 Table 9.1 Table 9.2 Table 9.3 Table 9.4 Table 12.1 List of charts use#_..r production a ligitized bathymetries Example of relative occurrence frequency of wind speed for 12 sectors and associated 50 year estimates at Duba. Summary List of the 30 Storms 50-Year extreme wind speeds (1-hour average, 10 meter above LAT) Wind gust factors Five local areas of interest. Defining comer coordinates (UTM Zone 37). Water level stations withiin local areas and at Bab el Mandeb. Station location and summary of four main tidal constants. Model time steps, maximum depths, and maximum Courant numbers. Comparison of the MEAN and RMS error between the MIKE 21 predicted water level and the Admiralty predicted tide levels at various locations in the Red Sea. The 30 storms, period overview. Recommended 50 year extreme significant wave height (m) Wave model grid dimensions. Model dimensions written in italic indicate waves coming from land, hence the only input parameter being wind speed and direction. Example of offshore to nearshore transformation Boundary data for extreme wave simulations at Jeddah. Numbers in brackets indicate direction where only wind is applied (wind from land). Measured and Simulated Wave Heights for verification period. Subgrid boundary conditions

Danrk Hydraulirk InnitullDaniih Hydraulic lnllit~le

1-1

1

INTRODUCTION
This report presents the results of a study aimed at providing environmental design
conditions for the Saudi Aramco concession areas in the Red Sea.

The study was carried out by Danish Hydraulic Institute (DHI) for Saudi Aramco (SA) under Contract No. 45057/00 between Aramco Overseas Company B.V. (AOC BV) and DHI. The contract was awarded to DHI and initiated by signing on February 27, 1995.
Bid specifications were described in the Invitation for Proposals dated November 10, 1994. Bidders were given an opportunity to submit and respond to questions by phone interview. The results of the phone interviews were compiled and distributed to prospective bidders. DHI submitted a Technical Proposal (DHI No. 94-2351), dated December 1994 based on the Terms of Reference in the Invitation for Bidders and the phone interview questions and answers. The contract covers the study of wind, waves, currents and water levels. No measured data exist that can provide the basis for design data, neither for noma\ nor for extreme conditions. Most of the effort has been directed toward a hmdcasting of these conditions from a 3-year continuous period representing normal conditions, and 30 storm events representing extreme conditions with subsequent analyses to derive the design values. The hindcast modelling of the coarse grid regional wind and wave conditions was performed by Oceanweather Inc., Conn., USA (OWI), acting as a subcontractor to DHI. Hindcast modelling of the currents, water levels, and h e grid local waves was performed by DHI. All data analysis as applies to the database was performed by DHI. The study began with a Kick-off meeting between DHI and SA at DHI’s offices in Hnrsholm, Denmark on 27 and 28 March 1995. Clarifications to the specifications and agreements for modification were made during this meeting. Minutes of the Kick-off meeting were distributed and submitted for approval. Nearing the completion of the study program, from January 6-8, 1996, a meeting at DHI was held between DHI and SA, where DHI presented the details of the study program and the state of the database. Further clarifications and modifications to the specifications were discussed. Minutes of Meeting were distributed, and they describe the agreed upon modifications to the specifications. As a consequence of the discussions between Saudi Aramco and DHI regarding the poor resolution of the bathymetries and the recently available charts from the Saudi Port Authority, it was decided to extend the project by updating the bathymetries at Duba and Jizan and rerun the numerical models and reanalyse the data for these two areas using the new, improved bathymetries. The Duba and Juan areas were updated because the Admiralty Charts for these areas showed the highest un-

1-2

certainty. At the same time it was decided to include an investigation of potential diffraction effects on the wave results for an area influenced by a large island. Both of the above tasks were carried out under Amendment no. 45057/01 dated 31 August 1996.

This report replaces the preliminary report issued June 1996 before the project
extension. It shall be noted that all results and data (tables, figures) and discussions in this final report refer to the revised bathymetries at Duba and Jizan. O l modest changes have been made to the text in the first 10 Chapters of the ny report. The changes to the results have been fully reflected and incorporated into the updated data base. Two chapters have been added to this final report, one, Chapter 11, describes and discusses the updated bathymetries and associated changes in results, the other, Chapter 12, presents the diffraction study. Deliverables include 20 copies of the final report, 3 CD-ROM disks containing the design database and interface software, and 3 copies of the database user manual.

DHI would like to thank Mr. Vince Cardone of O W for his cooperation and participation in the project study. The discussions and comments received from S.A. representatives during the project have been highly appreciated, especially as regards the detailed comments from Dr. M o m AI-Mashouk.

2- 1

2

OVERVIEW OF THE RED SEA HINDCAST STUDY
The overall objective of the study was to develop a graphical user interface (GUI) database which contains and gives users access to numerical model results and statistics for nearshore and offshore areas of the Red Sea. The database provides a higher resolution of data coverage in 5 local operational areas along the coastline of Saudi Arabia. The 5 local areas are located around the areas of Juan, Jeddah, Rabigh, Yanbu and Duba, see Fig. 2.1. The study consisted of hindcasting of winds, waves, currents and water levels based on available data and numerical modelling tools. The results of the hindcast study were based on 2 types of analysis. The first was an analysis of normal conditions and the second an analysis of extreme conditions. The normal wind and wave conditions were based on a 3 year continuous period from 1986 through 1988. The analysis provided various statistical information including fatigue design parameters. Normal current and water levels were based on astronomical tides. The extreme analysis was based on 30 storm events selected from the period 1975 through 1994. The hindcast analysis of the storm events provided directional extreme wind, wave, current and surge water level information for determination of extreme design values. The hindcast modelling was based mostly on historical measurements and observations of wind and air pressure throughout the Red Sea area. Measured data in the Red Sea is sparse for long periods of time, and the most reliable continuous period of measurement comes from ship observations in the offshore areas from the period after 1975 when the Suez Canal was re-opened. On-land wind station data have been used for calibration and verification of the wind modelling simulations. The specific numerical studies included:
1.

Regional (0.25 deg Lat/Long grid) wind field analysis of the entire Red Sea basin using kinematic wind model, by Oceanweather Inc., USA. Regional (0.25 deg Lat/Long grid) wave field analysis using 3’rd generation wave model, by Oceanweather Inc., USA. Regional (6 km) and local area (1.5 km) current and water level modelling using DHI’s MIKE 21 HD. Local (75Om x 125m grid) wave field modelling using DHI’s MIKE 21 NSW.

2.
3.

4.

2-2

.. ...... -...."

I...

e

Fig. 2. I

Location map of local operational areas.

2-3

The postprocessing of the data from the numerical modelling included fatigue analysis from the studies of normal wave conditions and joint probability statistics based on the combination of normal wind and wave data. For the extreme conditions with recurrence periods up to 100 years, the DHI programme package EVA was used. EVA combines a peak-over-threshold (POT) analysis and a statistical distribution function to derive the estimates on extremes.
Results from the analysis are stored in a GUI database for easy viewing and retrieval. Results are stored at every grid point at a resolution of 1500 rn x 1500 m in the 5 local areas, and also for 6 points representing general offshore regions of the Red Sea. The decision to store results for only 6 points representing the offshore regions of the Red Sea was made during the Kick-off meeting at DHI. The points which were selected were based on observations of the following. First, various literature identify 3 main regions of the Red Sea major. Generally speaking the Southern Red Sea, the Northern Red Sea and the Central Red Sea. There are general meteorological patterns which are typical for all regions, and other patterns which are distinct to each zone. For example, in the summer months the general wind pattern is from north to south for all regions. In the winter, however, the flow is from north to south in the northern regions, south to north in the southern regions, and in a transition between north to south and south to north and often calm in the central region. The central Zone "C" generally covers the typical transition region, while the zones, "A", "B" subdivide the southern region, and zones "D" and "E" subdivide the northern zone. Zone "F" represents a point in the Gulf of Aqaba. Fig. 2.1 shows the locations and extents of the 5 local operational areas and Fig. 2.2 shows the locations of the 6 reference points and regions for the Red Sea.
The report is organised as follows:

Chapter 3 Chapter 4 Chapter 5

summarises the available data for the Red Sea area. presents the digitization of the bathymetries, which is essential for the numerical modelling. details the statistical analysis that is common for all the extreme value analyses. Special analysis of winds and waves from the southerly sectors that was necessitated by the conditions for the area extending from south of Jeddah to Yanbu is also discussed. presents the wind simulations and analyses. The wind fields have been used to derive currents and waves in the respective numerical models (ref. Chapter 7, 8, and 9). The derived wind fields have also provided the basis for the normal and extreme wind conditions. presents the current and level study. The current and level are linked through a common numerical model and are thus presented together. cover the wave study, with the regional modelling in Chapter 8 and the local area modelling in Chapter 9.

Chapter 6

Chapter 7

Chapters 8-9

2-4

+

+

+
W

T
W

W

o\
E *

9

?

$

r)

+

+

+

+

+
4-

+

+
+

+
4-

+
-t

+
4-

c

,I

'0

Fig. 2.2

Location map o oflhore reference points and regions. f

1 W 2 7 1 7 M 6 1 .drWhicSt&.AB

2-5

Chapter 10 Chapter 11

covers an overview description of the database, covering general use, data formats, and file storage. reviews and discusses the updated bathymetries for Duba and

Jizan and the consequences for the current and wave data as
compared to the original bathymetries. Chapter 12 contains the diffraction study carried out to assess the effect of diffraction around islands.

For Chapters 7-9, a common structure of the chapters has been adapted so far as possible: presentation of the basic models and input, calibration and verification, review of normal condition modelling, and finally the extreme modelling and analysis. The report contains only few examples of the results since all results are delivered in the database in a user-friendly format. The Database User Guide is issued as a separate document and is attached to this report as Annex 1. When using the database, it is important to note that all data have been based on numerical models. Even though state of the art models have been used and a successful verification has been obtained for the extremely scarce data, measurements are generally more accurate and provide a more reliable source for deriving data for a specific location.

In consideration of the scarce verification possibilities, field data campaigns and the use of satellite data are recommended for obtaining improved confidence in the data. Assumptions and procedures leading to slightly conservative estimates have been introduced where possible in the recommended design data. The design engineer should, however, treat the data with caution and make use of any additional information that may become available to assist in defining design criteria.
This report was furnished by Danish Hydraulic Institute to Saudi Aramco (with input from its sub-contractor Oceanweather Inc.) at the request of Saudi Aramco in fulfilment of contract no. 45057/00.
Neither Danish Hydraulic Institute, O W ,nor any person acting on behalf of these parties assumes any liabilities with respect to the use of, or for damage resulting from the use of the information, method or process disclosed in this report.

3-1

3

REVIEW OF AVAILABLE DATA
Although it was generally found that there existed a lack of direct measurements useful to this study program, numerous sources of information were used. The information and data which was used is as follows: a. Bathymetry Data 1. British A d m i i t y Charts 2. Saudi Aramco survey charts 3. Saudi Port Authority b. WindData 1. Direct measurements at various land based stations 2. Ship observations 3. Climatic Summaries 4. Synoptic weather charts 5. Measurements and statistics from past study programs c. Wave Data 1I Ship observations 2. Measurements and statistics from past study programs (very limited data) Current and Water Level Data 1. Measurements and statistics from past study programs

d.

Bathymetry data was based mostly on information obtained from British Admiralty charts supplemented with a few Saudi Aramco charts. Near the end of the study program the existence of the Saudi Port Authority(SPA) charts was realized by Saudi Aramco. Comparisons to the Admiralty charts were made and it was determined that the SPA charts offered greater detail of depth soundings. The project extension utilized the new SPA charts to opgrade the bathymetries in the Jizan and Duba areas. Data used for the wind study is thoroughly described in the appended report prepared by Oceanweather Ipc. (OW), Appendix B. The major source of information was from ship observations mostly along the center of the basin from the years 1975 to 1994. Other information used for the wind study includes time series of wind measurements at various land based stations along the Saudi coast, synoptic charts, and various reports available at DHI and provided by Saudi Aramco. Very little direct measurement information was available for the wave study. The only source of reliable wave measurements comes from a study performed by TetraTech at Yanbu for the period 1977 to 1980. Wave height observations from ship reports were generally not useable for calibration or verification purposes.

3-2

Information regarding water levels and currents was generally limited to the TetraTech study at Yanbu, and to some very general information provided in various texts. Tide level information was limited mainly to tidal constituent information from the British Admiralty and International Hydrographic Organization (IHO) databases. Other tidal data was provided by Saudi Aramco, and was found to agree with information published by the Admiralty and IHO sources. A list of all reference material found relevant to this project is listed in the reference list given in Section 13, and in OW’Sreport, Appendix B.

41

4
4.1

DIGITIZATION OF BATHYMETRIES

General
The modelling study is complicated by the fact that the bathymetry of the Red Sea is very irregular. The bathymetry varies from a very deep and long offshore region (below 2000 meter depth) to very irregular nearshore areas of large shallow reef formations, often surrounded by deep dropoffs (300-500 meter depth), sometimes located very close to the shore. Beach slopes in many local areas are found to be very steep, in many areas reaching slopes of 1:4 to 1:3. The digitization of the Red Sea bathymetry was a large comprehensive effort. Approximately 28 navigational charts and survey maps were used. Generally, all available depth information, contours and soundings, were used to synthesize the digital map for the numerical models. See Table 4.1 below for a summary list of all the charts which were used in the study. As noted in Chapter 3, the charts provide a somewhat poor resolution and infomation in many areas.

Table 4.1 List of charts used for production o digitized bathymetries f
~~

CHART

DESCRIPTION

SCALE

I ADM-6
ADM-8 ADM-63

I Gulf of Aden
Gulf o Aqaba and Suez to El Akhawein f
El Akhawein to Gezirat el Dibia

I I

(1.750,OOO)
750,000) (1:

I

(1:750,000) (1:750,000) (1: 750,000)
(I :400,000)

I ADM-138 I Gezirat el Dibia to Masamirit Islet
ADM-141 ADM-143 ADM-321 Masamirit Islet to Zubair Islands Jazirat at Ta’ir to Bab-el-Mandeb Abu Duda to Sirrain and approaches to AI Lith

I

(1:200,000)

I

ADM-322

I

Sirrain to Qadimbal Islet and approaches to AI Qunfzdha Hanish Islands to Bab-el-Mandeb Bab-el-Mandeb to Aden Red Sea (1:200,000) (1:200,000) (1.2,250,000)

ADM-1925 ADM-3661 ADM-4704

4-2

~

Aak'itionalty is listed below any fine scale charts and surveys used to provide more detailed informationfor the local areas of particular concern to the study:
Duba

I SUR-8371 I Saudi Aramco commercial survey
SPA-12 SPA-13 Saudi Ports Authority Saudi Ports Authority

I

(1:2000)
(1:150,OOO)

I I

(I:150,OOO)

I Yanbu
ADM-63A ADM-327 Southern approaches to Madinat Yanbu Northern approaches to Yanbu (1:160,OOO)
(I : OOO) 75,

I ADM-328 I Madinat Yanbu as Sinaiyah
ADM-328 Approaches to Yanbu Industrial Port

I I

(1:25.OOO) (1:so,000)

I I I I I

I ADM-63B I Rabigh I ADM-2659 I Shib Nazar to Qita Kidan
ADM-2577 Jiddah Harbour

(1:100,000) (1:200,000)

I
I

(I:lO,OOO)
(1:75,000) (1:75.OOO) (1:200,OOO)

I ADM-2599 I Approaches to Jeddah
ADM-2658 ADM-2659 Outer approaches to Jeddah Shib Nazar to Qita Kidan

I Jizan
SUR-8299 SUR-8299A Saudi Aramco commercial survey

(I :20,000)

Saudi Aramco commercial survey

(I :IO,rn)

I SPA-26
SPA-27

1 SAUDI Ports Authoriv
SAUDI Ports Authority

I

(1:150,OOO) (1:I so,OOO)

I

4-3

program interpolates the digitized x,y,z data points into a two-dimensional fmed grid map. A mother bathymetry was interpolated to a grid resolution of 750 meter from which all other model specific bathymetries were derived. After the initial interpolation was performed, the digitized maps were then overlaid on the original charts, and final checking and editing was performed manually. Further discussions of the bathymetry as it applied specifically to current, water level and wave modelling are presented in their respective chapters.

f Fig. 4.1 shows the entire digitized Red Sea bathymetry at a grid resolution o 75om.

4-4

Depth (m)

:.

Fig. 4. I

Digitized Red Sea bathymetry a! a grid resolution of 7SOm.

15%c6nnwi r w l r 9 6 a s ~ ~ d

Danik Hydraulak InPtulIDamh nydravlr lnilil~le

5-1

5

STATISTICAL ANALYSIS
The directional distributions of both extreme winds and waves have been determined based on a combination of extreme value analysis of the storm events and an analysis of the normal conditions found in the 3 years continuous data, excluding the lowest wind speeddwave heights.

The basic assumption is that the wind speeddwave heights from all directional sectors follow the same probability distribution as the omnidirectional values, but the frequency of occurrence of storms vary for each sector, or in other words the storm intensity varies for the different directional sectors.
The omni-directional distribution of wind speedslwave heights is found from the peak values of the 30 storms included in the hindcast. From the results of the hindcasts it was evident that the dominant direction at the southern part of the Red Sea is from South to South-East, whereas for the remaining part of the Red Sea, the dominant directions are from North to North-West. The omnidirectional wind speeds and wave heights were determined by taking the peaks from the hindcasted storms and analysing them using the DHI statistical analysis package EVA. This package makes use of a peaks-over-threshold (POT) analysis which can be combined with a number of statistical distribution functions (see Appendix C) to produce estimates of extreme values. For the present study the POT method and Weibull distribution function as recommended by the Work Group under the International Association of Hydraulics Research has been applied. The aim of the storm selection for the hindcast was to define the most severe events over the 19 years covered by the required meteorological data. As there is some uncertainty associated with the storm selection, especially when an area as large and varied as the Red Sea is concerned, only the selected events that actually resulted in high values at a given location should be used in the subsequent statistical analysis. In the selection of data for the statistical analysis at any given location it is important not to mix normal events with the real storm events, and it is important to ensure that the population of extreme events inciuded in the analysis contains all storms within the time period considered, that have yielded significant wave heights higher than the selected threshold. There is no explicit procedure available for selecting the threshold value that will ensure that these requirements are fulfilled. In the present study the data in a number of grid points were exercised and evaluated, and it was concluded that a threshold equal to 60 percent of the maximum value resulting from the 30 storms at any given location provided an adequate threshold for the subsequent analysis.

5-2

The storm intensity per year is then determined as the number of events exceeding the threshold (0.6 times Uw,max) or Hs,max divided by 19 (the number of years covered by the hindcast).

The estimates of the values corresponding to return periods of 1, 10, 50 and 100 years are determined automatically by running the EVA programme with the input being the peak values (those exceeding 60% of the maximum value) and the number of years covered (19 years in this study). Fig. 5.1 shows an example of the EVA output for extreme winds at the location offshore Duba.
In addition to the central estimates of the extreme values, the EVA programme also produces estimates of the standard deviations of the estimates based on the input data and a Monte Carlo simulation. The standard deviation is a measure of the uncertainty of the estimates as induced by the limited number of data. The standard deviations are also shown on the output, see Fig. 5.1. For the present project it has been decided to use the central estimate plus one standard deviation as the recommended values. Some account is in this way taken of the uncertainty in the extreme value estimations.

The storm intensity for the omnidirectional winddwaves can be estimated from the 30 storms. The wind/wave directions of most of the storm peaks fall within narrow directional bands, ie data for only a couple of the 12 directional sectors are available. The storm intensity for the individual sectors can therefore not be estimated from the limited storm hindcast dataset. However, the 3 years of continuous hindcast can give some indication of the expected relative intensity. By neglecting the smallest values (wind speeds lower than 4 m/s and significant wave heights less t a 0.5 m), the remaining data have hn been analysed with respect to directions to yield the relative frequency of occurrence of winddwaves within each of the 12 directional sectors at all calculation points. Subsequent to this analysis, the storm intensity of the worst direction, ie the most frequent direction, was adjusted to be identical to the.omni-directional, and all other directions were adjusted accordingly. The storm intensity for the 11 remaining sectors was thus relative to the worst direction. With this storm intensity combined with the omni-directional statistical Weibull distribution, the extreme values for the 11 other directions have been determined. Table 5.1 shows the normalized frequency of occurrence of winds for the location offshore Duba together with the resulting 50 year estimates of wind speeds.

Table 5.1 Example o relative occurrence frequency of wind speed for I 2 sectors f and associated 50 year e s t i m e s at Duba.

5-3

. I

I

1

:IF
. . . . . ..-. .....
"....A:

...
I.

I

- ,-.-,*.., .,.. .,., , c-.,
-r.
.-.t
."..I,.,.

..-.... ..... .-..
1 . " .

1

....... ........
L

... .... ...;:\.; ............

,.

I-.-,
11".

I..:,.\.. ........;............ :..~ ; ;. ........ ............ j ........... 'II
~

I

.-..<-,,
,.I
~

..

..,.....

.i

........................ < ............ :...........:............ :........... i
;

...........

~

1

Fig. 5.1

EVA outpur. Example showing atreme win& offshore Duba.

IPPM16lln668-l.drMh.AB

5-4

For most areas and most directions the applied procedure for deriving directional values is expected to produce conservative estimates, partly because the worst direction has been .equated to the omnidirectional and the other directions have been increased similarly, partly because the more normal conditions lead to more frequent occurrences on the directions deviating from the main wind and wave directions (N-NW and S-SE) of the Red Sea. The assumption regarding identical probability distribution functions for all directions is normally rather robust and is implicitly applied when omni-directional data are derived by lumping all data into one population (or when data from non-directional wave measurements form the basis). To test the applicability of the assumption for the present study a comparison was made between the results of the extreme value analysis and normal wave conditions. The maximum significant wave height within each 30 degree sector from each of the 3 years of continuous hindcast were determined at the five offshore boundary grid points at Duba, Yanbu, Rabigh, Jeddah and Jizan. The average value of the three yearly maximum values were used as an estimate on the 1 year values. The directional distributions of these are illustrated in Figs. 5.2 - 5.6, where the values have been normalized with the maximum value. Also shown in these Figures are the 1 year values derived from the extreme value analysis (also normalized with the maximum value). It is seen that the directional variation is quite similar in the two different data sets. At Duba, Fig. 5.2, and at Jizan, Fig. 5.6, the extreme value analysis results in overestimation of the values for the sectors with very infrequent winds and waves, i.e. the sectors different from the northerly and the southern directions, respectively. For these sectors, the data from the extreme value analysis may seem somewhat conservative. At Yanbu, Rabigh and Jeddah, the comparison between the two datasets are remarkably favourable. It is thus concluded that the assumption regarding identical distribution functions for all directions at a given location is acceptable, however, with comments and modifications discussed next. In the central zone extending from south of Jeddah to Yanbu, the northerly winds and waves are dominating, but some very rare meteorological conditions can result in severe Southerly storms. Due to this, a special analysis has been implemented for the extreme conditions here.

5-5

Fig. 5 . 2

Comparison of normalized 1-year estimates of directional significant wave heights. Ofshore Duba.
-

direction

Fig. 5 . 3

Comparison of normalized 1-year estimates o directional sign@cant wave heights. f Ofshore Yanbu.

5-6

Fig. 5 . 4

Comparison of normalized I-year estimates of directional signflcant wave heights. mshore Rabigh.

i
.........
i i j

i
j

i i i

j
? .

i

..........
. .
j

............... j
~ ~

j j
j j
i j ! i

j
j

i i
~

j j

j i

i
i j

j

*
I
i j
~

j i

<
: j
j

i
+
!
.-c-- --

+
j j

+

j

; i

'

i

j
!

i i i i i

1
)

j j 150 180 2 I 2

i i

I

)

270 300 330 360 3 1 0

direction

Fig. 5.5

Comparison of normalized I-year estimates of directional significant wave heights. mshore Jeddah.

5-7

Fig. 5.6

Comparison o normalized I-year estimates o directional sign$cant wave heights. f f

Offshore Jizan.

After the initial 30 storm events had been produced, the north to north-westerly directions were clearly the dominant for the central zone locations, however, with single very large values from south to south-easterly directions. To assess whether this single event was an outlier, Oceanweather was asked to perform a second round of storm evaluation, focusing on the rare events with southerly storms around the central zone area. This exercise-resulted in the selection of three more events, one of which proved to yield severe wind and wave conditions in the area, the other two producing more moderate conditions. None of the three additional storms had any impact at Jizan and Duba. With the inclusion of the 3 storms, it is unlikely that more extreme events of this kind have occurred in the 19-years period. A period of 19 years is normally considered representative as a typical period from which it is possible to derive extreme values for hydrographic data. It can not be fully excluded that a very rare event may occur should a different time period be considered. Evidently, the 100 year event is bound to occur some day. However, for the analysis applied to derive the extreme values, any 19 year period would be considered representative. With the rather irregular storm distribution and few events with south to southeasterly directions it was decided to perform a special analysis for the areas in the central Red Sea zone: for each calculation point, the EVA result for the northwesterly direction was used to produce the directional distributions for the ten 30 degree sectors, in a way identical to the procedure described previously. The

5-8

northwesterly storm peaks were plotted as exceedance probabilities together with the values from the southeasterly directions. Fig. 5.7 shows an example of such a plot for the wind speeds at a location offshore Yanbu. It is evident from Fig. 5.7 that most of the extreme events come from the northerly directions, but a few severe events are from the southerly sector. Based on this and similar plots for the other calculation locations, the results for the southerly sector could not be reliably assessed using the standard approach. The exceedance probability plots have aided in this reassessment. From inspection of the directions associated with the southerly peaks it is found that the rare events come from only two sectors, the ones centred around 150 and 180 degrees. The extreme values for wind speeds and wave heights have been adjusted to take these observations into account: The recommended design values for these two directional sectors have thus been increased, mainly for the 100 and 50 year return period. For the 1 year return period no adjustments have been made, and for the 10 year return period the values have been adjusted as necessary to keep inline with the data.

0.0001

0.0010

0.0100 exceedance prob (%)

0.1ooo

A

NWstorms

SEstorms

Fig. 5.7

Example o plot showing exceedence probabilities for northern and southern secf tor. Data from offshore point at Yanbu.

6- 1

6

WIND STUDY
The purpose of this phase of the project was

.

to investigate available wind measurement data for areas of the Red Sea to develop 2-D wind fields for use in the modelling program to determine wind statistics from.the derived wind fields for selected areas of interest.

Using the available measurements and observations, and the derived wind fields from OWI's wind modelling program, statistics for selected areas of the Red Sea were compiled and entered into the database. Within the local operational areas, winds are specified as either "nearshore " or "offshore". For the "offshore" reference, a grid point was analyzed at the local area model offshore boundary. For the "nearshore" reference, a grid position over water but nearest to 1,and was analyzed. As there is a land-sea breeze which may not be fully resolved spatially in the wind model, it was not reasonable to try to present results between one region to the other. The nearshore zone attempts to represent the region of the local area where the land-sea breeze is most developed, and the offshore zone represents the region of the local area where the land-sea breeze is least developed. It was not possible to resolve this any further without more measurements and using a much finer grid spacing in the regional model. Wind statistics based on long-term reliable measurements were only available for a few land based locations. The largest dataset available for synthesis of the wind fields comes from ship observations reported from the years 1975 to 1994. However, other sources were used to complement this dataset and are discussed in detail in OW'Sreport, Appendix B. The wind study, like the wave study, consisted of two main programs: The study of a 3-year continuous period for investigation of so-called "normal conditions", and the study of storm events to represent "extreme conditions".

6.1

Normal Conditions
A 3 year continuous period was selected from 1986 through 1988 to represent the typical "normal conditions". The selection of this period was based mainly on the 3 year period with the most observations and with the least amount of gaps from the available data sources. Refer to OWI's report, Appendix B, for a full discussion.

6-2

O W delivered the 2D wind field time series to DHI in digital form. The winds were representative of 1-hour averages at a height of 1Om above LAT (Lowest Astronomical Tide).

611 ..

Analysis of Normal Wind Conditions
Various levels of processing were performed by DHI: 1. Converting 0.25 degree Lat/Long O W grid coordinates to 6.0 km and 1.5 km UTM grids for use in the DHI hydrodynamic current and level models. (note that UTM Zone 37 was utilized throughout the study) Extracting specific grid point time series of wind speed and direction for statistical processing. Converting O W 1-hour average wind speeds to World Meteorological Organisation (WMO) standard 10-minute averages. Statistical analysis including: - scatter diagrams of Wind Speed (Ws) vs Direction - exceedance diagrams of Ws vs Direction - monthly and seasonal distribution of Ws - monthly and seasonal exceedance of Ws - persistence diagrams (to account for duration)

2.

3.
4.

Conversion of 1-hour average wind speeds to 10-minute averages was performed by applying a factor of 1.08. Refer to Section 6.2.2 for details of the formulation for wind gust scaling factors. Figs. 6.1 and 6.2 show comparisons of the predicted versus measured wind statistics at Jizan and Yanbu, respectively. Note the good agreement between the airport measurements and predicted nearshore winds. Computed wind fields do not vary significantly from model grid point to model grid point, but instead vary regionally. For this reason, wind analysis was limited to a selected number of grid points. The analysis was performed for 5 grid points along the center of the Red Sea basin, one grid point in the middle of the Gulf of Aqaba, and for an offshore and onshore grid point for each of the 5 operational areas. Refer to Fig. 2.1 for the locations of these points. Drawings 6.1 - 6.11, Appendix A, show scatter diagrams of the normal wind statistics compiled for the offshore points of each of the 5 local operational areas and the 6 Central Offshore reference points. This information has also been stored in the database.

6-3

Jizan
All Wind Speeds
35

,

j

270

31 5

1 Offshore 0 0 Airport

Nearshore

1

Fig. 6.I

Comparison of rhe predicred versus measured wind statistics ar Jizan.

6-4

Yanbu
All Wind Speeds

70
60

50

8
40

0 30

3

s

20
10

0

0

45

90 135 180 225 Wind Direction (Coming from)

270

315

1 Offshore 0
Fig. 6.2

Airport

Nearshore

I

Comparison of the predicted versus measured wind statistics at Yanbu.

62 .

Extreme Conditions
From the period of 1975 through 1994,a thorough and lengthy process of selecting 30 storms to represent the most severe events was undertaken. This selection process was performed by O W . The aim was to find at least 30 storms which produced the most severe events throughout the entire basin while at the same time considering the directionality of the storms to aid assisting in the determination of the directional statistics. The selection of the storms, based on the wind and air pressure only, had to be considered not only for the maximum speed and direction, but also for duration and area of coverage, since these factors combined determine the expected wave growth, current development, and surge water levels. See Owl’s report, Appendix B, for a full description of the storm selection process. Table 6.1 s~&arizes the 30 storms which, based on a thorough evaluation of all information, were initially selected by O W for the wind and wave hindcast.

6-5

e

6.2.1

Analysis of Extreme Winds
The analysis of the storm events was performed using the methodology described in Section 5. The following steps describe the general procedure for carrying out the extreme value analysis of each offshore and nearshore reference point:

1.

Extraction of the time series of the wind speed and direction at each reference position from the 2-D field, for each storm. Identification of the peak wind speed and direction for each storm event. Sorting the peaks by directional sectors. Mainly from N-NW, S-SEand OMNI directional. Peak over threshold (POT) analysis including the Weibull probability of exceedance to derive values corresponding to 1, 10, 50 and 100 year return periods. Determination of the extreme values for the other directions based on information from the normal condition distribution. Computing and applying gust factors for 3-seconds, 1-minute and 10minutes to 1-hour average results.

2.
3.
4.

5.

6.

The POT and Weibull analysis was performed using DHI’s EVA software package. See Appendix C for a detailed description of the EVA software. Generally speaking, the extreme storm peaks found from the hindcast of the 30 storms tend to come from the directions of N-NW or S-SE, along the main axis or of the Red Sea basin. This was also found to be the case for the extreme wave hindcast, and even more so for the waves. In fact, because of the predominance of these main wind flow directions, it would likely take many hundreds of storms to get the entire directional distribution of storm peaks for all 12 directional sectors based on a rigorous statistical analysis. This was not a feasible option for this study. Instead, a method combining the results of the 30 storm hindcast analysis with the 3-year continuous period analysis has been adopted to perform the complete directional extrema1 analysis. For details of the statistical analysis procedure, please refer to Chapter 5 .

6-6

e

Table 6.1 Summry List of the 30 Storms

a

a.

e

6-1

Modifications to the general method were implemented at Jeddah, Rabigh and Yanbu. At these locations, very few severe storms from S-SE occur. However, when they do occur, they can be the most severe of all. Due to the very small number of severe storms from S-SE, an alternate directional analysis was performed for winds from the 150 and 180 degree sectors for these 3 locations. Generally speaking, the fitting of the 10-year, 50-year and 100-year return period were based on the exceedance probability "best fit", while the 1-year return period extreme value is based on the general method. For the areas affected by the rare but severe storms, the omnidirectional values were adjusted by the maximum values found from the combination of the general analysis and the special analysis for the 150 and 180 degree sectors.

The EVA analysis also provides information regarding the standard deviation of the
central estimates of the predicted extremes. To account for some of the uncertainty inherent in the hindcast data and in the extreme value analysis the standard deviation has been added to the central estimates to yield the recommended values. Table 6.2 shows the recommended 50-year directional extreme wind speeds found for the offshore boundary of the five operational areas and the 6 Central Red Sea locations.

Table 6.2 50-Year extreme wind speed (I-hour average, 10 meter above U T )
RECOMMENDED 50 - YEAR EXTREME W SPEED m I (I-hour average, 10 meter above L4Tj
(mh)

Region E Region F

16.2 13.0

13.9
10.8

11.9 9.6

9.6 6.9

12.2 9.4

14.2
10.0

13.4 9.1

16.7 9.9

16.7 9.9

15.3

18.1

19.6 13.7

19.6
13.7

11.3

12.6

6-8

The ratios between the values for individual sectors found for the 50 year return period were used also for the 100 year, 10 year and 1 year values. Extreme winds are presented in the database for averaging intervals of 3-seconds, 1-minute, 10-minutes and 1-hour. Conversion of averaging intervals has been done using the industry standard application of gust factors. Wind gust factors were computed following the relationship given by DNV (Det Norske Veritas, 1989). The relationship is as follows: U(t,z) = U (q,q) [ 1 where, z

+ 0.137 In (z/z,)

- 0.047 In (t/q) ]

zr
t

= height above still water sea surface level = reference height ( 10 m above LAT) = averaging time interval

tr = reference time interval U (t.2) = average wind speed at the specified t and z. U ( q , ~ ) = reference wind speed
Table 6.3 gives the wind gust factors which were used for the conversion of 1-hour wind averages to the 3 second, 1 minute and 10 minute average values.

Table 6 . 3 Wind gust factors
Wind Speed Averaging Interval Gust Factor

3 seconds 1 minute 10 minutes 1 hour

1

1.33

7-1

7
7.1

CURRENT AND LEVEL HYDRODYNAMIC MODELLING
Overview
The hydrodynamic current and level modelling was carried out in 2 parts. The study of normal conditions was based on astronomical tidal forcing. The study of extreme conditions was based on the combination of astronomical tide, wind stress and atmospheric pressure variations. The modelling was carried out using DHI’s MIKE 21 HD numerical model. Results of the modelling were processed and entered into the design database. The processing of results was carried out in a way that horizontal spacial variations of water levels and currents were generalized into zones. Each grid point classified within a particular zone is assigned the same result for either water level or current speed and direction. This was done to reduce the computation time for processing, and to keep the database storage requirements within a reasonable size.

7.2
7.2.1

Model Set-up
The Numerid Model MIKE 21 HD

-

MIKE 21 is a comprehensive modelling system for 2-dimensional free surface flows where stratification can be neglected, ie it is a depth-averaged one-layer
model.

MIKE 21 HD is the basic module of the entire MME 21 system; most other modules, for advection-dispersion modelling, sediment transport modelling etc etc. must be based on a MIKE 21 HD application.
For present purposes, however, the HD-module has been applied as a stand-alone modelling tool, to produce information about currents and water levels for the entire Red Sea as well as for the five local areas around Duba, Yanbu, Rabigh, Jeddah, and Jizan in particular. MIKE 21 HD solves the three depth-averaged partial differential equations describing conservation of overall mass and momentum in two horizontal directions, (x;y). The equations are solved by an implicit finite difference numerical scheme. The following physical effects are included in the fundamental equations and are thus described by the model:

7-2

bathymetrical variations water level variations wind shear stress barometric pressure gradients convective and cross momentum bottom friction Coriolis forces momentum dispersion ("eddy viscosity") Bathymetrical variations inter alia cause acceleration of the current over shallow areas associated with a local increase of the friction loss and a possible wind set-up relative to deeper areas. Water level variations cause pressure gradients which accelerate flow towards areas with surface depressions. The wind forcing at the surface is proportional to the wind speed squared. The proportionality constant (the wind friction factor) is a calibration parameter. Atmospheric pressure gradients tend to be balanced by water level variations such that 1 millibar added pressure corresponds to a depression of the water level by 1 cm. The bottom friction, usually expressed by the Manning resistance formula, expresses that the current is decelerated by a force proportional to its velocity squared. The Manning number, from which the proportionality factor may be derived, is a calibration parameter. Large variation of the Manning number within a model area is particularly relevant in connection with flooding (and/or drying) of overgrown areas and flow over bathymetrically very irregular areas. The Coriolis force compensates for the fixed coordinate reference system being tied to the surface of the Earth, thus undergoing accelerations due to the Earth's rotation. This force is proportional to the current velocity and deflects it to the right (on the northern hemisphere) when facing in the direction of the current. The Coriolis force may effect water level variations of several centimetres along a section perpendicular to the current direction.
7.2.2

Bathymetries

A general discussion of the production of digitized bathymetries is given in Section 4. The following specific details relate to the bathymetries as pertains specifically to the hydrodynamic current and water level modelling. A total of 4 grids have been produced for the current and level modelling portion of the hindcast study. A coarse resolution 6 km grid regional model and 3 fine resolution local area models of 1.5 km grid spacing were used. The local models were produced for the following areas:

7-3

Duba Yanbu, Rabigh, and Jeddah - All combined within a larger local area. Due to the relative proximity of these three local areas it proved more efficient to construct one extended model area, comprising and combining them all during (local) modelling, but generating results for each separately for further processing and statistical analysis

Jizan
It may be noted that as an intermediary step all bathymetrical information was compiled and interpolated to a single 750 m resolution 'mother' grid covering the entire Red Sea. This fine grid was never used directly for modelling purposes, but only served as a well-defined basis from which all the regular modelling bathymetries could be extracted.

The 750 m Red Sea mother bathymetry was produced from position and depth
information collected from various sea charts. The charts were digitized and then interpolated to the fvred grid. Refer to Section 4 for a general description of the bathymetrical processing. With respect to these local area models it should be noted that except for certain areas of only a very limited extent covered by commercial surveys, even the available detailed chart bathymetrical information (as listed) is in some respects flawed. Generally, reefs are only incompletely charted, and the indicated depths will tend to be in error on the side of caution from a mariners point of view. For Rabigh, it is also noted that no very detailed charts have been available. For Yanbu and Jeddah, the most complete and satisfactory bathymetrical information has been available, relatively speaking. Comer coordinates defining the areas of interest, as specified by Saudi Aramco for each of the five local areas, Duba, Yanbu, Rabigh, Jeddah, and Jizan are listed in Table 7.1. Figs. 7.1 through 7.6 show contour plots of the digitized 1500 m resolution local area model bathymetries, and within the plots are indicated extent of the defined areas of interest. Fig. 7.7 shows a contour plot of the 6000 m resolution regional bathymetry, which was used for the regional hydrodynamic model of the entire Red Sea.

The pronounced shelf character in all the local areas is evident from the bathymetry plots, the offshore water depth increasing from less than 100 m to more than 500 m within a relatively short distance. At Jizan the shelf is seen to be much wider, and to be more complicated than the other areas, in terms of the topography.

Danrk Hydraulsk Inllil~I/Ddnlh Hydraulic lnrlllule

7-4

For simplicity and to preserve accuracy in the modelling, all model bathymetries have been kept aligned, for which reason they have been made larger than the mere areas of interest in order to completely encircle those areas. Also, at least 15 km (10 grid points) have been added around all areas of interest to secure results within those areas from contamination with any possible spurious boundary effects. In particular the seaward model boundary of the bathymetrically very complex Juan local area model have been pulled away some distance from the area of interest to reach out to deep water conditions off the shelf. In the other local areas the seaward model boundary naturally reaches out beyond the shelf when the respective areas of interest are to be contained by aligned rectangular model grids. Note that for the hydrodynamic modelling, no minimum depth limitation is imposed in any of the bathymetries. However, it should be understood by the users that the resolution of the bathymetry will not reflect local changes which may occur within a single grid point of the model.

7-5

(UTM Zone 37).

AREA Comer
Duba NE Duba NWl Duba SE Duba SW Yanbu NE Yanbu NW Yanbu SE Yanbu SW Rabigh NE Rabigh NW Rabigh SE Rabigh SW Jeddah NE Jeddah NW Jeddah SE Jeddah SW Jizan NE Jizan NW Jizan SE Jizan SW

Longitude 35.69 35.51' 35.91' 35.79 3 7.820 37.430 38.720 38.330 39.19 38.800 39.19 38.800 39.27' 38.830 39.27' 38.830 42.58 41.960 42.820 42.260

Lutitude
2 7.480 2 7.380

Eusting 172884.4 154771.7 193678.1

Northing 3044007.2 3033406.4 3004644.8 2995067.7 2705702.5 2668441.4 2606683.2 2571387.1 2543560.9 2543567.0 2458324.5 2458330.5 2414066.1 2414051.3 2341015.9 2341001.5 1912903.3 1894142.4 1814832.8 1796064.4

27.130

27 . 0 8
24.460 24.120 23.57' 23.250 23.00' 23.000 22.230 22.230 21.830 21.830 21.17' 21.17' 17.27' 17.11' 16.380 16.220

I 77549.o
380400.8 340442.0 471425.9 431460.7 515373.0 479502.7 515458.9 479388.3 52 7904.0 482431.1 528030.2 482351.6 876479.4 81501 7.0 908199.0 848582.1

Danrk Hydraulirk Inslilul/Danish Hydraulic lnlfil~te

7-6

(GridsDacina 1500 rn)

42

40

3e

36

34

32

%

>, .<

P. .

%

DEFINED AREA OFINTEREST

,

%
1t

0

1E

0

2

4

6

8

10

12

14

16

18

20

15 krn

Fig. 7.1

Duba local 1500 m bathymetry.

7-7
(Grid spacing

1500 rn)

-G

260- -6
0

250240230220-

210200-

190r:

180170-

g

160-

E 150-

‘D 140-

.E?

E 130Ln

.3

v

6 1201101009080-

7060%

504030-

=.-I
0

10

20

30

40

50

60

70

80

Fig. 7.2

Combined Yanbu. Rabigh, and J e w local 1500 rn bathymetry, with individual a r e a indicated.

7-8

(Grid spociw 1500 m)

39

44

49

I
Fig. 7.3

15km

'

54

59

,

64

69

74

79

04

Yanbu local I500 m bathymetry (excerpt from "Combined"bathymetty).

7-9

(Grid spacing 1500 rn)

140

135

130

125

12c

Fl l5
'D
0 0

._ FllC
0 0

: :
0 L

10:

1 oc

9 :

9 (

8 f

8(

48

53
15km

58

63

68

73

78

83

87

I

I

Fig. 7.4

Rabigh local 1500 m bathymetry (excerpt from "Combined"bathymetry).

7-10

(Grid swcinq 1500 rn)

60

55

50

25

20

15

1c

1 0

15
15 k m

20

25

30

35

40

45

50

I

I

Fig. 7.5

Jeddah local 1500 m bathymetry (excerpt from "Combined"bathymetry).

7-11

H 15km

Fig. 7.6

Jizan local 1500 rn bathymetry.

19WQl-l7r76661.~kiEMCI.~

7-12
(Gridspacing 6000 m)

0

20

40

60

Fig. 7.7

6000 m resolution Red Sea bathymetry

7-13

e

7.2.3

Boundary Conditions
In the case of the entire Red Sea, regional model boundary conditions refer to; - (a) tidal water level variations forcing the behaviour of the model at Bab-el-Mandeb (formerly Perim Island) to the South; and - (b) the meteorological wind and pressure fields over the entire area covered, which modifies the pure tidal behaviour of the Red Sea. In the case of the three local models (Duba, Jizan, and Combined Yanbu, Rabigh, and Jeddah) boundary conditions refer to; - (a) local flux or water level variations at the three open model boundaries (west, north, and south, see Figs. 7.1 through 7.6),as extracted from the regional Red Sea model and transferred to the various local models; as well as - (b) the meteorological wind and pressure fields over the different local areas covered by the models. In all the local models a combination of water level and flux boundary transfer data have been applied in the modelling, since that is generally preferable over relying solely on one type of transfer data. In all cases the offshore west boundary has been a level boundary and the north and south boundaries have been applied as flux boundaries.

7.3
7.3.1

Normal Currents and Levels Astronomical Tide Only
Boundary Conditions
In the "normal" case, where wind and pressure effects are small and may be disregarded (or at least averaged out with respect to their expected value), there is really only one independent function determining the flow in the Red Sea, namely the astronomical tidal forcing at Babel-Mandeb. Within the Red Sea all variations in levels and flow are determined from that, in accordance with the physical laws and principles laid down in the modelling system. Thus, the modelling procedure for tide only conditions (pure tide) has been to f m t calculate time series of tidal variations at Bab-el-Mandeb and at an array of tidal stations within the Red Sea from published tidal constants, M2, S2, K1 and 01, with associated phase lags. This initial pre-processing was accomplished separately by a tidal prediction program. Then the 2D model was run, and a satisfactory fit of the model-internal tidal station variations was sought by making adjustments to model parameters such as bathymetry, bottom friction etc, as well as by tuning the Eab-el-Mandeb tidal constants. This was considered proper in view of the fact that the literature values are not really authoritative and f m l y established, and furthermore taking into account, that the tidal boundary variations are not associated with one, very specific location. They must be considered to be more of a Straits of Bub-el-Mundeb average tidal variation. Adjustment of model parameters and tuning of Bab-el-Mandeb tidal parameters were the essence of the model calibration process.

-

Danrk Hydraulirk InlitutlDan&h Hydraulic lnllilule

7-14

Published tidal constants for Red Sea water level stations withiin each of the five defined local areas, as well as at Bab-el-Mandeb, are summarized in Table 7.2. For Bab-el-Mandeb both the initial literature values and the final calibrated values are given. It is seen from Table 7.2 that Red Sea tides are generally small, in part very small, due to the MITOW entrance at the Straits of Bab-el-Mandeb. At Yanbu the tide may be characterized as mainly semidiurnal, otherwise it is mixed, albeit with a preponderance of semidiurnal components. At Jeddah tidal variations are particularly small. The station is not far from an amphidromic point offshore at deep water.
Table 7.2 Water level stations within local areas and at Bab el Mandeb. Station location and summaly of four main tidal constants.(source. Admiralty Tide Tables).
Station

Station name

1

Lat.

I

Long.

I
I

M2'. m

S2'. m

Kl', m
0.02

O' m l,
0

431 38.07 I 11
Madinat Yanbu Sherm Rabigb

2.8 40'

0.14

0.03

I!

!

1

(0' 21)

I

(4' 29)

1

(8' 17)

I

(4' 22)

1

I

Bab-el-Mandeb3 Notes:

I q m
0.36

(2' 26)

0.17 (4' 26) 0.17 (6' 21)

0.35 (359

0.18 (339 0.18 (339

I I

0.17 (359

(1) Amplitude (phase parameter parenthesized) (2) Literature values (formerly Perim Island) (3) Calibrated values for Srraits of Bab-el-Mandeb model b o u n d q values

Finally it may be noted, that in order to spin-up the model to a point when conditions everywhere in the Red Sea is consistent with the tidal forcing at Bab-el-Mandeb, and conversely when it is virtually independent of the arbitrary initial conditions applied inside the model, it was found to be necessary to run the model for a period of 10 days. In all regional simulations, where tidal forcing was included, an initial spin-up period of 10 days was applied, from which all model results were discarded. Where tidal forcing was not included, that is for pure surge storm modelling, it was found that a 5 day spin-up period was sufficient.
The models have been run with time steps ranging from 450 s for the regional model down to 90 s for the combined Yanbu, Rabigh, and Jeddah model as shown

7-15

in Table 7.3. The table also gives the maximum value of the Courant number anywhere within the model. The Courant number, C, is defined as:
a = c.At/hx 2

- where c is the celerity, At is the time step, and A x
c
=

is the grid spacing. The celerity is traditionally represented by the tidal wave celerity: (g.h)'

- where g is gravity and h is the water depth. MIKE 21 is designed for Courant numbers up to about 20. Courant number values of five (or less) may be considered wholly unproblematic. Values in the range 5-20 are commonly permissible, and it may be noted that in areas where the water depth is less than a quarter of the maximum water depth, the Courant number is less than half the model maximum Courant number.
Table 7.3 Model time steps, maximum depths, and maximum Courant numbers.
Time step

6)
Regional model, Ax=6 km Duba local model, Ax= 1.5 km Yanbu, Rabigh, and Jeddah Combined local model, Ax = 1.5 km Jizan local model, Ax= 1.5 km

Courant

450
180
90

120

app. 1600

10.1

732 ..

Calibration of Normal (Tidal) Currents and Levels
Predominantly, the model calibration process has concerned water levels, since few suitable current measurements have been available for that purpose. Initially, the regional model has been calibrated by comparison over a full springheap tidal cycle of calculated water level variations to predicted variations at tidal stations within the Red Sea. Subsequently, it has been c o n f i i e d that the obtained fit persisted in another, later period. Additionally, it has been verified that tidal station variations calculated by regional and local model simulations are - for all intents and purposes - identical.

19n01-17fl6661.mVEMCl.~

DanrX Hydraulirk lnilitul IDantrh Hydraulr lnrlilule

7-16

The main calibration parameters have been:

1.
2.

modification of the bathymetry at the Straits of Bab-el-Mandeb modification of the tidal constituents at the Straits of Bab-el-Mandeb adjustment of Manning bottom roughness

3.

The modification of the bathymetry at the model open boundary consisted of smoothing and widening of the opening to allow the M2 component of the tide to propagate through the model. The digitized bathymetry of the southern portion of the Red Sea was extremely irregular and its proximity to the open boundary was responsible for an excess dissipation of the M2 wave. The final bathymetry configuration was found through an iterative process, adjusting the bathymetry until the M2 component throughout the model was best represented. Fig. 7.8 shows the modified bathymetry at Bab-el-Mandeb. Modification of tidal components at the open boundary at Bab-el-Mandeb are made further to the adjustment of bathymetry. Published tidal constants concern a particular site, rather than a line, as it is applied in the model. It was found that the M2 component was about a factor of 2 low throughout prior to the modification of the bathymetry. Simply doubling the M 2 component at the boundary produced strange behaviour in the model, and it was found that the modification of the bathymetry near the boundary was a more effective way of allowing the M2 component to propagate. This bathymetry modification however had the effect of amplifying the K1 component throughout the model. With this, the amplitude of the K1 component at the model open boundary was reduced by nearly a factor of 2. In Fig. 7.9 is shown a comparison of "raw" Bab-el-Mandeb predicted water levels and resulting Straits of Bab-el-Mandeb calibrated water levels. It is seen from the figure that in calibration the extreme diurnal tidal variations were reduced and the semi-diumal tidal components were enhanced.

7-17

a) as digitized from chsrts

b) modified bathymetry

Fig. 7.8

Comparison of bathymetries at Bab-el-Mandeb. a) as digitized from charts b) modified bathymetry.

81-L

7-19

near the shoreline, and extend them constantly out to the nearest reported offshore depths. The third assumption was to assume a constant shallow depth throughout the uncharted region (both -5 m and -10 m were tested). Although local differences in currents and levels in the close proximity of the uncharted zone were noticed, there were no noticeable changes in currents or levels in any of the local operational areas of interest. Overall, the obtained model fit of water level variations are shown in Fig. 7.11 (Duba, Yanbu a1 Bahr, and Madinat Yanbu) and Fig. 7.12 (Rabigh, Jeddah, and Jizan). At Duba there is no tidal station to compare the model results to, but results from both the regional and local models are shown and coincide so closely that they are difficult to discern. The same is true for model results from the other four local areas. The best apparent fit between modelled and predicted water levels is found at Jizan, where tidal variations are relatively highest, and the least good fit is at Rabigh. At Rabigh, however, tidal amplitudes are so small that even though some deviation is evident, it only concerns a few centimetres. Despite extremely small tidal variations at Jeddah, a good fit is apparent from Fig. 7.12. Finally, an extended tidal simulation was performed to confrm that the model fit to tidal station variations persisted well after the calibration period. Results from this simulation are plotted on Drawings 7.1 to 7.2.

OanrX Hydraubk InrIfluflDanirh Hydrauk Inmule

7-20

Fig. 7.10

Contour map of Manning resistance for regional Red Sea tidal model.

19pb~17ns8-1 drm~cm AB

Dansk Hydraulak Ins18luIIDdnrsh Hydraulr InSlllUle

7-21

0

. .. ... . .
0.2
0.1

I _ -

Modirwt Y o h u WL. locd mod4 M o d k t Yahu WL, regional nwdd

0.0

-0.1 -0.2
-0.3

0000
01/11 1990

0o:oo
01/13

0000
01/15

0000
01/17

0000
01/19

0000 01/21

0000 01/23

0000 01/25

Fig. 7.11

Duba. Yanbu a1 Bahr, and Madinat Yanbu; comparison of model results from regional and local models (virtually coinciding), and comparison of model results to tidal station predicted water levels.

7-22

0
E

........
0.5 0.25

Jizon WL. local model JizM WL. regional model

0.0
-0.25

-0.5
0000 01/11
1990

0000 01/13

0000 01/15

0000
01/17

00:00 01/19

00:00 01/21

0000 01/23

00:00 01/25

Fig. 7.12

Rabigh, Jeddah. and Juan; comparison of model results from regional and local models ( v i m l l y coinciding), and comparison of model results to tidal station predicted water levels.

7-23

Criteria for the calibration acceptance was based on the following:

RMS error between MIKE 21 predicted and Admiralty predicted levels less than or equal to 0.05 m.
The mean of the absolute value of the difference between MIKE 21 and Admiralty predicted levels be less than or equal to 0.05 m. Table 7.4 summarizes the final calibration acceptance criteria RMS and Mean error for the various locations in the Red Sea, including Jizan, Jeddah, Rabigh and Yanbu.

Table 7.4 Comparison of the MEAN and RMS error beiween the MIKE 21 predicted water level and the Admiralty predicted tide levels at various locations in the Red Sea.
STATION Sawakin Port Sudan Quseir Aqaba Umm Qusur Yanbu a1 Bahr Madinat Yanbu Sherm Rabigh Jeddah Jizan

MEAN* (m)

RMS (m)

0.04

0.06

0.05 0.01

0.03
0.04

0.02

0.04 0.04 0.02 0.02

* average [abs (MIKE 21-ADM)I

733 ..

Normal Current and Level Simulations
A purely tidal (no wind) simulation has been performed which is basically the same as the model calibration run. Results of both regional and local simulations have been extracted and fitted in the data base, which is the main outcome of the investigation. An example current field from the Rabigh local area is shown in Fig. 7.13 with associated tidal variations.

7-24

140

135

130

125

120

0 0
Lo

Lo

E 115

.- 1 1 0 e

8 a
U ._
v

5

105

100

95

9c

85

8 C
48

---+ 0.2 rn/s
53
58

63

68

73

78

82

0.12 0.08

O.M

00 .
-0.04

-0.08
-0.12 -0. 16

0000 01/22 1990

OD00 01/23

0000 01/24

OD00 01/25

W:00 01/26

0000 01/27

Fig. 7.13

Rabigh local area, January 1990. Calculated current field (example) and associated tidal variations.

mt.osnntta.mwm

AB

1-25

7.3.4

Analysis of Normal (Tide) Water Levels The analysis of normal water levels is based on harmonic analysis of the time series of predicted tidal water levels. From an analysis of the water level at a point, the magnitude and phase of the 4 main constituents, namely M2, S2, K1 and 01, magnitude and phase, are computed. This analysis is performed using DHI’s standard tidal analysis software. Since the analysis of the constituents for each model grid point would be a computationally huge effort, and because the variation over a local operational area is relatively small, a strategy for reducing the number of analysis points was adopted. The analysis for implementation into the database was performed as follows:
1.

Create a map of maximum predicted water levels for an entire springneap cycle for each local area. Present the maximum water levels rounded up to the next highest increment of 0.10 m. Identify a representative grid point for each level increment, based on the largest maximum value in that increment band. Perform harmonic analysis for each representative grid point and apply to all grid points in that increment band.

2.

3.

This adaptation reduces the number of analysis points from thousands, to about 2-4 points for a single operational area. This analysis methodology introduces water level amplitude errors up to 10 cm, but the error will be on the conservative side. This method also assumes that there is no phase difference within a local area. This is not so serious considering that the areas are relatively small, and the uncertainties of the gauge station predictions and modelling predictions introduce errors of the same order. Since the analysis of normal conditions in this study does not take into account the effects of wind stress and atmospheric pressure, the Mean Sea Level (MSL) reference level was determined from published sources, where available, and linearly estimated for areas where there is no available information (namely Duba). The source used was the British Admiralty Tide Tables. Various other reference levels were computed based on the amplitudes of the constituents, and are given by:

MHWS = MSL + (M2 + S2) MLWS = MSL - (M2 + S2) MHWN = MSL + (M2 - S2) MLWN = MSL - (M2 - S2) MHW = (MHWS MHWN) I 2 MLW = (MLWS MLWN) I 2

+ +

(mean high water spring) (mean low water spring) (mean high water neap) (mean low water neap) (mean high water) (mean low water)

The highest astronomical tide (HAT) and the lowest astronomical tide (LAT) were determined from the maximum and minimum statistical values found from a long time series synthesized from the constituent values, respectively.

7-26

7.3.5

Analysis of Normal (Tide) Currents
The analysis of normal currents was based on harmonic analysis of the time series of predicted tidal currents. From an analysis of the current speed and direction at a point, the magnitude and phase of the main constituents, speed and direction, was computed. The analysis was performed using DHI’s standard tidal analysis software. Since the analysis of the constituents for each model grid point would be a computationally huge effort, and because the variation over large areas of an entire local operational area are relatively small, a strategy for reducing the number of analysis points was adopted. The analysis for implementation into the database was performed as follows: 1. Create a map of maximum current speeds and a map of their associated directions for an entire spring-neap cycle for each local area. Present the maximum current speeds rounded up to the nearest increment of 0.10 m/s and present directions in directional sectors of 30 degrees. Identify a representative grid point for each current speed and direction pair increment, based on the highest mean of the current speed in that increment band. Perform harmonic analysis for each representative grid point and apply to all grid points in that increment band.

2.

3.

This adaptation reduced the number of analysis points from thousands, to between 10-40 points for a single operational area.
The currents predicted from MIKE 21 represent the depth-averaged velocities. The currents results presented in the database are given for the depth-averaged velocity, the surface velocity and the velocity 0.5 meter above the sea bottom. The industry standard practice is to assume a velocity profile distribution based on a 1/7’th power law tit (det Norske Veritas, 1989), and is given as:

where; vtide(z) = tidal current velocity at level z h = water depth taken to the still water level (taken positive) Z = distance from still water level, positive upwards

7-27

7.4
7.4.1

Extreme Currents and Levels 30 Storm
Boundary Conditions

-

A total of t i t short (72-168 hours, exclusive of spin-up) storm periods were hry identified, as listed in Table 7.5. The models, both regional and local, were run for each of the 30 storms, both with meteorological forcing alone (pure surge) and with
meteorological forcing in combination with Bab-el-Mandeb tidal forcing in the period concerned. In Table 7.5 is additionally listed, as a rough classification of the storms, whether the resulting calculated main current direction along the coast is mainly northgoing or mainly southgoing. It is seen that southerly directions outnumber northerly directions by almost 3:l (22:8). Fortunately however, the two most severe storms, 8512 and 9302, each have different directions, northgoing and southgoing, respectively. The meteorological forcing, wind and pressure fields for the entire Red Sea area, were provided by Oceanweather, Inc. for each of the 30 storms with a resolution of 0.25 deg latitude by 0.25 deg longitude spatially, and 3 hours temporally. This coarser resolution meant that the meteorological fields had to be interpolated before being compatible with the 6 km regional and 1500 m local model resolutions. Ax=Ay=O.25 degree is considered quite sufficient to secure that any significant features of the wind and pressure distribution were resolved.

To secure model consistency when variable wind/pressure fields and Bab-el-Mandeb water level variations were applied simultaneously in the modelling, the model system is constructed such that during computations the specified pure tidal variation at the south boundary was automatically modified by an offset in accordance with the instantaneous local pressure deviation from "neutral" (defined as 1007 mbar).
7.4.2

Calibration of Extreme (Storm) Currents and Levels
There has been no basis for performing a calibration of the model to special storm characteristics. Thus, the storm periods have been simulated entirely based on the calibration obtained for tidal conditions. In particular, it would have been preferable to have had a site specific foundation for the choice of model wind friction factor, but since the extent of wind-czused surges is generally restricted in the Red Sea, this want of data is not considered a problem.

Danrk Hydravlirk Inrlm/Danirh Hydraulic lnllil~le

1-28

e

Table 7.5 The 30 storms, period overview.

0

e

a

l) Note: Classification by main current direction along coast in model calculations.

7-29

7.4.3

Extreme (Storm) Current and Level Simulations The 30 storms have been modelled, both as pure surges (no tides) and in combination with the predicted tidal forcing of the Red Sea. Results of both regional and local simulations have been extracted and fitted in the data base, which is the main outcome of the investigation. An example current field from the 9302 storm in Rabigh local area is shown in Fig. 7.14 with associated water level variations through the storm period. The figure concerns the pure surge modelling. The corresponding tide and surge computed model result field is shown on Drawing 7.30. Similar example current fields at Rabigh from the combined tide and surge modelling are shown in Drawings 7.3 through 7.32, - one from each of the 30 storms. 7701 to 9401. For the 9302 storm corresponding example currents fields (tide+surge) are also shown from the four other local areas, Duba, Yanbu, Jeddah, and Jizan, cf Drawings 7.33 through 7.36. It is seen from the plotted current fields how depth averaged current speeds are negligible at deep water, but significant (albeit not very high) over shallow areas. Likewise is seen how at Rabigh (Dwgs. 7.3-32) southgoing currents along the coast are generally associated with low water levels, so that the most common storm type(s) do not cause any significant surge. Conversely, high water levels occur during northerly storms, cf Dwg. 7.32 (9401 storm).

7.4.4

Analysis of Extreme (Storm) Water Levels
Analysis of extreme water levels was based on performing an extreme value analysis (EVA) of the peak levels found from the 30 storm simulations. The variation of extreme water levels over a local area are, in general, not so large. So to reduce the huge computational effort to analyze each model grid point, an alternate approach was adopted. The analysis for implementation into the database was performed as follows:
1.

Create a map of maximum water levels for each of the 30 storm events at each grid point. Create a map of the absolute extreme water level for each grid point and identify sub areas based on water level increments of 0.10 m. Identify a representative grid point for each extreme level increment. This was determined by finding the grid point with the highest mean of the 30 maximum values for each incremental sub area. For each level increment, perform an EVA analysis based on the 30 maximum levels for the representative grid point.

2. 3.

4.

7-30

(Grid spacirw 1500 m)

140

137
Y
#

135
I

125

120

a

0
0

E 115

'"

._ FllO
(0

t i
105

2

v

&

1oc
95

9c

8 :

a
0.2 0.1 0.0
-0.1

0 2 m/s .
00W 02(02
DO00 02/03 DGDD 02/04 0000 02/05 00:00 02/06

4.2
4.3 0O:W DZ/DI 1993

Fig. 7.14

Rabigh local area, 9302 storm. Calculated currentfield (example) and time series of water level surge.

i%&nneai.&.rnw.~~

7-3 1

This adaptation reduces the number of analysis points from many thousands, to about 2-10 points for a single operational area.
7.4.5

Analysis of Extreme (Storm) Currents
The analysis of extreme currents was based on the EVA (extreme value analysis) of the highest current speeds from the 30 storm simulations. To reduce the huge computational effort of performing the EVA analysis for each grid point, a strategy for reducing the number of analysis points was adopted. The analysis for implementation into the database was perfomed as follows:
1.

Create a map of maximum current speed and a map of their associated directions for each of the 30 s t o m events. Create a map of the maximum of all maximums, and present the speeds in increments of 0.10 m/s and direction increments of 30 degrees. Identify a representative grid point for each current speed and direction pair increment, based on the point with the highest mean of the 30 maximum values. Perform an EVA analysis for each representative grid point and apply to all grid points in that speedldirection increment band.

2.

3.

4.

This adaptation reduces the number of analysis points from thousands, to between 10-20 points for an entire area.
The currents predicted from MIKE 21 represent the depth-averaged velocities. The currents presented in the database are given for the depth-averaged velocity, the surface velocity and the velocity 0.5 meter above the sea bottom. The velocity profile was based on industry standard recommendations. For the tidal component it is based on the 117’th power law as described in Section 7.4.2. The total current profile, as normally assumed in standard industry practice (det Norske Veritas, 1989). is a sum of an assumed tidal component and the wind driven component, and is given as v(z) = vtide(z) = vWind(z) Vwind(z) where = v(z) Z =
Vtide

vtide [ (h+z) I h ] I n for z < = 0 for 0 > = z > = - vwind[ (b+z) l h ] = for z < -b

+ vwind

-b

vtide

P
b

total current velocity at level z distance from still water level, positive upwards = tidal current velocity at the still water level = wind-generated current velocity at the still water level = water depth taken to the still water level (taken positive) = reference depth for wind-generated current = 50 meters

1-32

For this application, the wind-generated surface current must be assumed to solve
the above equation for the total profile. The industry standard assumed wind-generated surface current is normally given as;
'wind

=

0.015U( 1 hr, 1Om)

where;

U

= the I-hour average wind speed, relative to 10 meters above the still water level.

Due to the simplicity of the industry standard current profile model, it was not justified to apply more than a single typical extreme wind speed applied to all current profiles. A wind speed of 13.3 m/s was applied, which is very typical of what may be considered as average extreme winds for all areas. This gives the following formulation for the surface current speed.
'wind

= 0.20 m/s

Danrk nydraulirk InrmrlDanhh Hydrauk lmutule

8- 1

8
8.1

REGIONAL WAVE MODELLING Overview
The regional wave modelling was performed by Oceanweather Inc., USA (OWI). This section of the report describes the general details of the work performed by O W and the analysis of that data performed by DHI. For a full description of the work performed by O W , the reader is directed to read Owl’s report, Appendix B. The purpose of the O W regional wave study was to develop wave fields covering the entire Red Sea which could be used directly to develop offshore wave statistics, and to provide boundary conditions for the local area nearshore wave modelling performed by DHI. Using the O W predicted 2D wind fields, as described in Chapter 6, applied to OW’S 3’rd Generation wave model, OW3G, wave fields were simulated for a 3 year continuous period, representative of normal conditions, and 30 storm events representative of the most severe events covering a 19 year period to be used for extrema1 analysis for determining extreme design conditions.

8.2

The Numerical Model - OWI3G
The wave hindcast model, OW3G, is a fully-discrete spectral wave model. The wave spectrum is resolved in discrete frequency and direction bins. In the model the Red Sea basin is represented by a discrete grid of water depths. The solution of the wave field is obtained based upon the integration of the spectral energy balance equation. This process is successively solved at each model grid point and each time step to simulate the physical processes of wave growth and dissipation (through source terms in the energy balance equation), and wave propagation.

8.2.1

Model Outputs
Standard output results of the OW3G model are time series o f m0 - zero’th spectral moment ml - fiist spectral moment m2 - second spectral moment HmO - Significant wave height (4*sqrt(mO)) (=H,) TP - Peak spectral period VMD - vector mean direction

8-2

a

8.2.2

Wave Parameter Relationships
Outputs from the OWI3G model are derived from the directional spectra. The spectral shape predicted during significant wave events was found to be closely represented by a JONSWAP spectrum. Fig. 8.1 shows an OWI3G hindcast spectra compared to a standard JONSWAP spectrum (ua = 0.07, ob = 0.09, y = 3.3). The JONSWAP spectral formulation was used throughout the study to relate statistical wave period parameters, namely:

TdTz TdT,

=

1.29 1.20

=

where Tz

=

upcrossing period

=

4 2

T,,, = mean period

10

16
14 12

10
0 6
4 2

0

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Frequency (Hz) Fig. 8.1
Comparison of hindcu.fr spectra versus JONSWAP spectra.

8-3

The maximum wave height is estimated from the relation:

H , ,

=

1.92

* H, ,

Other relationships were derived from correlation of parameters found in the hindcast, such as the relationship between the peak period and the significant wave height.

T ,= 3.5 * ,
with
%o

in m and

T in sec. ,

8.3
8.3.1

Model Setup
OWUG Model Grid The O W model grid is the same as was used for the regional wind field study. The O W grid is based on a latitude-longitude array as follows: Latitude range: Longitude range:
10 - 30 deg N (spacing of 0.25 deg) 32 - 45 deg E (spacing of 0.25 deg)

The OW1 grid is shown in Fig. 8.2

8.3.2

Spectral Discretization Direction: Frequency:

24 bands, at 15 deg intervals 0.039 Hz to 0.32157 Hz, increasing in a geometric progression

8.3.3

Regional Wave Model Boundary Conditions The OWI3G model is driven by the time history of surface wind speed and direction (1-hour average at 20 meter height) at 30 minute intervals. The duration of each simulation was long enough to ensure that the wave response was fully spunup before peak conditions were specified in all m a s of interest in the basin, and also spun-up in source regions of swell which eventually propagate to the areas of interest.

84

30' 28'

22' : 20' ; 18' :

IS'
14' 12'

32' 33' 34' 35' 36' 37' 38'

39'

40'

41'

42'

43'

44'

45'

Fig. 8.2

Grid dircretizafion used by O W . (Grid spacing 0.25 degree larirude and longitude).

8-5

8.4

Calibration and Verification
Refer to Appendix B for Owl’s discussion of the OW3G model validation. With the type of data available to this study, very little direct verification was possible, however, refer to Section 9.6 for a discussion of the combined offshore-nearshore wave verification performed at Yanbu.

8.5
8.5.1

Normal Conditions
Simulations The simulation period selected was the 3 years from January 1, 1986 through December 31, 1988. Refer to Appendix B for Owl’s discussion of the selection of the 3 year period to represent the normal conditions. Using the 3-year continuous wind field produced from Owl’s kinematic analysis, as described in Section 6.1 and in Owl’s report, the OW3G wave model was run for the same period. Results were saved to disk at 3 hour intervals for the entire period. 3 years of data were thus available throughout the Red Sea for m m,, m,, H ,,, . , T,,VMD.

8.5.2

Normal Wave Conditions
Normal waves were analyzed much the same as for normal winds. For each offshore grid point of interest, Le. the five points at the boundaries of the local models and the six points representing the conditions in the middle of the Red Sea and in the Gulf of Aqaba, the time series of wave parameters for the 3 year period was processed into various statistical tables. The analysis consisted of composing the following tables:

1. 2. 3. 4. 5. 6.

7.
8. 9.

directional distribution of significant wave height, HmO. monthly, seasonal and annual distribution of HmO. directional exceedance of HmO. monthly, seasonal and annual distribution of HmO. table of most probable wave period and the associated range of periods. total number of waves versus direction. total number of waves versus period. Joint occurrence of HmO vs Peak period, Tp. Joint occurrence of H m O vs Wind speed, Ws.

For the normal waves in the five local models, the time series at the boundaries were stored for the subsequent analysis and application in the database, cf. Section 9. Drawings 8.1 - 8.6 in Appendix A show the directional distribution of significant wave height at the offshore locations.

8-6

8.5.3

Results
The results of the regional normal wave study are used to provide statistics (including fatigue) of offshore areas, and to provide statistics and time series for use in the nearshore wave modelling. Various statistical analysis were performed from the results of these simulations to produce the relationships mentioned above. The procedure applied to derive the scatter diagrams for seastate parameters, eg monthly distribution of significant wave heights, directional distribution of significant wave heights, scatter diagrams of significant wave heights vs peak period, is straight forward: each data of the timeseries were stored in its appropriate bin in a multidirectional array (height vs month, period, direction), and the summing up was subsequently made to produce the relevant 2-dimensional arrays. For the individual waves, the height distribution was assumed to follow the Rayleigh distribution: p(H) = 1 - exp -(H/Hm0)**2 With a seastate duration of 3 hours, the number of waves can be found as: Nwave = 3*60*60/Tz = 10800/Tz where Tz is the mean zero-crossing period. For a zero-crossing period of 10 s, a 3-hour seastate contains roughly 1000 waves. Combining the number of waves and the probability distribution given above, the number of waves within a given range of heights can be determined for each set of Hs and Tz. Each wave has been assigned a period equal to Tz. This is a simplification, but when summing up over many seastates (a total of almost 6000 for the 3 years), an acceptable estimate on the wave period distribution will result.

This calculation has been performed for each seastate and the results stored in a
3-dimensional array containing wave height, wave period and wave direction. The scatter plots of heights versus period and versus direction have been found by summing up over the alternate parameter (ie summing up over all directions to produce the height vs period distribution and vice versa).

8.6
8.6.1

Extreme Conditions 30 Storms
Simulations
Within the 19 years for which acceptable data information was available, 30 events were selected for hindcasting. Refer to Appendix B for OW’S discussion of the selection of the 30 storm events used to represent the most severe events from 1975 through 1994. Using the wind

-

8-1

fields produced from OW’Skinematic analysis of the same 30 storm events, as described in Section 6.2 and in OW’S report, the OW3G wave model was run for the same period. Results were saved to disk at 1 hour intervals for each storm period. A summary list of the storms is shown in Table 8.1. Each storm is represented as a time series of data similar to the ones produced for the 3 years continuous conditions. However, for the extreme analysis only the peak value of the seastate and associated parameters were of concern.
8.6.2

Analysis Procedures
The analysis of the 30 storm events was performed similarly to the extreme wind analysis, using a peak-over-threshold model combined with the Weibull distribution. The full directional distribution was determined using the methods as discussed in Chapter 5. The following steps describe the general procedure for canying out the extreme value analysis of each offshore and nearshore reference point:
1.

Extraction of the time series of the significant wave height, peak period, mean upcrossing period and mean wave direction at each reference position from the 2D field, for each storm. Identification of the peak significant wave height and associated period and direction for each storm event. Sorting the peaks by directional sectors. Mainly N-NW, S-SE and OMNI directional. In general, the NW and SE directions are the main directions which produce significant extreme events. Perform extreme value analysis based on DHI’s software package, EVA, using the distribution and POT analysis with Weibull using threshold value of 0.6*(maximum peak value) for the N-NW, S-SE sectors and OMNI directional. Determination of the extreme values for the other directions based on information from the normal condition distribution, cf. Chapter 5 .

2. 3.

4.

5.

See Appendix C for a detailed description of the EVA software.
863 ..
Results
The results of the regional extreme wave study are used to provide long term return period statistics of offshore areas, and to provide long term return period statistics for use for the nearshore wave modelling boundary conditions. Table 8.1 shows the recommended 50 year values at the offshore boundary of the five local operational areas and for the six Central Red Sea reference points.

8-8

Table 8.1 Recommended 50 year extreme significant wave height (m)

50 - YEAR EXTREME SIGhTFICANT

WAVE HEIGHT (m)

Recommended Values

RegionE Region F

2.9
1.7

2.3

1.9

1.4

1.8

2.8

2.5

2.2

2.3

2.8
1.1

3.8

4.7

4.7

0.6

0.6

0.6

0.6

06 .

0.8

I

1.0

0.9

1.3

1.8

I;8

9- 1

9 9.1
9.1.1

NEARSHORE WAVE MODELLING
Methodology
Normal Conditions
The requested data for normal waves in the local areas requires information to the same extent as that used for the offshore areas. As for the offshore areas, the nearshore data material is scarce and is in no way adequate for providing the basis for the analysis. Therefore, the hindcast data was transferred to the nearshore areas. The transfer was made using DHI’s MIKE 21 NSW model. Instead of simulating the nearshore wave conditions directly (in the time domain) based on the wind conditions and offshore wave conditions, a database containing information on the transformation of the offshore wave conditions to the near-shore areas was established through a series of steady state simulations.

This was done by calculating the wave conditions in the local areas as a function of
offshore wave and wind conditions and entering the results into the database. Based on this database, the wave time series at any grid point covered by the model setup can be synthesized based on the offshore wave and wind boundary time series, or, if the wind is blowing from land, the wind conditions only.
9.1.2

Extreme Conditions For the nearshore extreme wave conditions, the extreme values established for the offshore points for each direction (see section 8.6) was applied as input parameters to the nearshore models, and the corresponding extreme values in the local areas were determined directly.

9.2

The Numerical Model

-

MIKE 21 NSW

To describe the transformation of wave conditions from offshore to nearshore, DHI’s nearshore spectral wind-wave model (MIKE 21 NSW) was applied. MIKE 21 NSW is a spectral wind-wave model which describes the propagation, growth and decay of short-period waves in nearshore areas. The model includes the effects of refraction and shoaling due to varying depth, wave generation due to wind, and energy dissipation due to bottom friction and wave breaking.
The model is a stationary, directionally decoupled, parametric model. A parameterisation of the conservation equation in the frequency domain is performed

9-2

introducing the zero'th and the f i s t moment of the wave-action spectrum as dependent variables. The input parameters to the model are:

. A bathymetry data file
Offshore Wave Conditions (or only wind): Significant Wave Height, Hmo Mean Wave Period, T m Mean Wave Direction, MWD Directional Spreading

. Wind Speed . Wind Direction
Model Output is:

. In every grid point:
Significant Wave Height, HmO Mean Wave Period, T m Mean Wave Direction, MWD Directional Spreading

The model requires that the waves travel in directions which deviate less then 60" from that of the positive x-axis of the computational model grid. Therefore a model has to be set up for each wave direction to be investigated. A further requirement is that the grid cells are distorted with a finer resolution in the direction of the main wave direction.

9.3
9.3.1

Model setup
General
As the model requires that the waves travel in a direction which deviates less than 60" from that of the positive x-axis of the computational model grid, a model grid for each of the wave directions considered has been set up.

The wave analysis was carried out in 30" sectors, therefore a total of 12 grids was established for each of the five local areas.
The resolution of the models was 125 m in the x-axis (wave) direction and 750 m in the y-axis direction. The extension of each model (direction) was made large enough to ensure that the wave conditions in the areas of interest was not influenced by the boundaries of the wave models.

9-3

The numerical model utilises a cosine" directional spreading. From an inspection of the regional wave model results, it was found that a constant n value of 10 was a suitable value for all nearshore wave modelling. Given the wave parameters, the directional distribution of the wave energy is given by :

Here ndir is the number of discrete directions, E, =H,2/16 is the total energy of the discrete energy spectrum and the directional distribution function D is defined by

D(eJ =O

where fl is a normalisation factor,

9.3.2

Bathymetries
The model bathymetries are based on the same digitized bathymetrical data as the current/level models, see Section 6.1.2. For each of the five local areas, a mother bathymetry with a resolution of 250 m was produced, from which the 12 grids were generated. Fig. 9.1 shows as an example, the size and location of the 12 individual model grids for Yanbu.

As mentioned elsewhere in this report, it must be noted that the bathymetry information available was not of the very best quality, and that substantial changes in water depths occurred within very short distances for many areas. This in combination with many reef areas introduce some uncertainty into the modelling results. Procedures aiming at obtaining results on the conservative side have been introduced as appropriate and where possible. Note that no minimum water depth was imposed within the wave modelling study. However, it should be understood by the users that the resolution of the bathymetry will not reflect local changes which may occur within a single grid point of the model.

9-4

km

km

200

3 100

20

0

20

40

,

0

20

40

0

20

10

60

i

240
2100

Fig. 9.I

Size and locations for the 12 models for Yanbu.
a

9-5

The dimensions of the 60 (5 areas times 12 directions) models are listed in Table 9.1.

Table 9.1 Wave model grid dimensions. Model dimensions written in italic indicate waves coming from land, hence the only input parameter being wind speed and direction.

I
Dtion
(ON)

I
0
30

Jizan
xdim Ax= 125111 y-dim Ay= 750111

I
200 243 278 280 243

Jeddah
xdim Ax= 125111 ydim Ay= 75Om

I

Rabigh
xdim

I

Yanbu
xdim Ax=
125m

I

Duba
xdim Ax= ydim Ay= 750m

I
80
IO1

Ax=
125m 891

ydim Ay= 750111

ydim Ay=

750m

125m

1500
1351

906

108

102

1250 913 700

210 255

1120 501 361

1051 624

126 I37

1021 693

143 198

60

751

225
220 160

101

I I

90

120

I 1040 1 I 1450 I

I 494 I I 1101 I

131
151

I I

574 991

I I

181 161

I 1200 I I 1435 I

I I

601 721

I I

101 76

I I

I
9.4
9.4.1

330

I

1561

I

170

I

941

I

137

I

990

I

116

I 1565 I

150

I

505

I

81

1

Normal Wave Conditions
Simulations
Preparing a complete database based on a continuous 3-year time series, covering all combinations of H ,T m , MWD and wind speeds, would lead to an unrealistic s number of simulations, and would make the final database unwieldy. Instead it was decided to carry out basis simulations with typical wave conditions. For the directions where the wave direction is towards the coast, four simulations were carried out with mean wave periods of Tm = 4s, 6s,8s and 10s respectively.

9-6

As the wave period is the governing parameter for the wave transformation and as the wave height variations are relatively small (in time), it was decided to carry out all simulations with Hs = 1.0 and a wind speed of 6 m/s. A wind speed of 6 m/s was chosen as it corresponds to a wave height of about 1 m assuming a fully-developed sea with the relevant fetches. Including the wind in the simulations ensures the generation of the locally wind generated waves behind islands and reefs, where relevant.

The scaling of wave transformation parameters using linear methods improves with
the degree of discretization of the basic and important parameters. The scaling of wave periods is discretized using mean periods of 4, 6, 8 and 10 seconds. For the wave periods less than 4 seconds, preliminary testing showed that no significant change in mean period was predicted through the local models, and thus no scaling was performed for periods less than 4 seconds. Wave height is generally linearly scalable until bottom effects (breaking, shoaling, friction) begin to dominate the process. Using a single wave height to scale all waves between 0 and 2 meters at the boundary using a 1 meter solution has the implication that waves larger than 1 meter may not be affected by the bottom as much as they would normally be (conservative effect) and waves less than one meter will be affected by the bottom more than they would normally be (non-conservative effect). The effects are, however, quite small and the linear scaling in the height is assessed adequate for the large percentage of all the waves (70-80%) with heights less than 2 m at the offshore boundary. For significant wave heights greater than 2 meters, the scale factor is interpolated between the 1 meter solution and 50-year extreme solution, which greatly improves the results where wave-wave interaction would have the most significant impact on the results. It is true that for wave heights between 1.5-2.0 meters, the results may be somewhat non-conservative. However, considering the relatively short fetch lengths of the local models, and the discretization of the presented statistical tables of results, this should not affect the results in any noticeable way. Fig. 9.2 and 9.3 show examples of the 171 simulations carried out. The two examples are for Yanbu, wave direction = 300", and wave period = 4 and 10 seconds respectively.

9-7

(Gridsping 125 m)

Fig. 9.2

Wavefield, Yanbu. Hs,offshore = 1 . h Tm = 4s. Wave direction = 300",

(Gridspacing 125 m)

. .
0
200

400

600

800

1000

1200

1400

Fig. 9.3

Wave Field, Yanbu. Hs,offshore = 1 . h Tm = 10s. Wave direction = 300"

9-8

0

9.4.2

Transformation from Offshore to Nearshore

9.4.2.1 Data Preparation

The results from the simulations consist of one set of transfer solutions for each of the rotated bathymetry grids (125 m x 750 m), or 12 sets for each area. Each set of solutions contains the wave height and associated period and direction corresponding to the 4 basic wave periods and a significant wave height of 1 m at the boundary. Before entering all the results into the database, all results were transformedhterpolated into one grid representation of the considered area with a resolution of 500 m x 500 m grid. Then this 500 rn grid was transformed into the 1500 m reference grid (the same as for the current and water level data). This step was not done by simple interpolation, but by identifying the maximum wave height within each 1500 x 1500 rn grid cell (i.e. maximum significant wave height of the nine 500 m grid cells). The procedure was chosen to avoid unrealistically small values in the 500 m grid which could occur very close to land or reefs, resulting in too low values when merging the nine 500 m cells into one 1500 rn cell. For a given offshore condition on the boundary, the wave conditions within the local areas are determined by selecting the steady state transfers which are closest to the offshore wave period and wave direction. Then an interpolation is made between these results based on the offshore wave period and direction. This method is illustrated by an example in 9.4.2.2. The wave period and wave direction is transformed by interpolation (based on the offshore wave period and direction) between the results from simulations with adjacent wave periods and wave directions. The wave height is transformed by applying the scaling factor, which is equal to Hs,nearshore from the Hs.offshore=l.O m simulations on the actual offshore wave height. The wave height scaling factor is determined similarly as for the wave period and direction, i.e. by interpolation. The above describes the general method for the transformation. However some exceptions were defined. These exceptions are described in section 9.4.2.3.
9.4.2.2 Example of Transformation of Wave Height

Assume that the offshore wave conditions at a given time step are:

HS= 0.8m.
Tm = 5 . O ~ m d MWD = 262.5"

9-9

For the selected point in the local area, the wave height is now calculated based on the offshore wave period and wave direction. The wave height scaling factor, wave period and wave direction for the four simulations with adjacent offshore wave period and wave direction are found from the database. Wave Direction 240", Wave Direction 240", Wave Direction 270°, Wave Direction 270",

T m = 4 ~ T m = 6 ~ T m = 4 ~ Tm = 6 s

Then an interpolation is carried out to get the wave height scaling factor, the wave period and the wave direction nearshore for 262.5' and Tm = 5.0 s, as shown in Table 9.2.

Table 9.2 Example of offshore to nearshore tramformation
Wave Simulations Results From Simulations in Database
~~~

Scaling Factor Interpolation From Database based on Offshore Wave Period and Direction
Interpolation m on T (5.0 s) Interpolation on MWD (262.5 ")

Wave Direction

Mean Wave Period

Wave Height Scaling Factor

240" 240" 270" 270"

4 6 4 6

0.80
0.75 0.70 0.70

0.69

0.64

Final Wave Height, Hs = Hs.offshore . Scaling factor = 0.8 . 0.69 = The wave period and direction nearshore is found using a similar procedure.

Danrk Hydraulirk InstnutiDanish Hydrauliclnrtlu7e

9-10

e

9.4.2.3 Exceptions
Some exceptions to the transformation procedure had to be defined:

With offshore boundary waves:

H > 2.0 m: s
The scaling is done by interpolating between the 50 year results and the s results from H = 1.0 m

This was done because when Hs,offshore > > 1.0 the wave height might get overestimated where wave breaking occurs, as the reduction of a 1 m wave will be less than for a higher wave with the same period due to breaking.

Wave Height: Scaling factors as found for Tm = 4 s Wave Direction: Directions as found for Tm = 4 s Wave Period : The offshore (boundary) wave period.

Offshore Wave Direction: Coming from Land:
Wind direction also from land: Wave height scaled linearly with the wind speed from the simulations with wind speed = 6 d s . Wave period based on H : T m = 4 . s

m s

(SPM)

Wave directions 'as usual' (interpolate between the two adjacent 30" sets based on wind-direction). Wind direction towards land: The wave height is put into the Hs=0-0.5 m category. The direction and period will be as the offshore wave (from land). It should be noted that the exceptions to the general transfer procedure constitute a small percentage of all data.

9-11

9.5
9.5.1

Extreme Wave Conditions
Simulations
Based on the offshore wave extreme value analysis (for details see Section 8.7),the boundary wave and wind conditions for the nearshore simulations were established.
As an example, Table 9.3 lists the 50 year values applied as boundary data for

Jeddah.

Table 9.3 Boundary data for extreme wave simulations at Jeddah. Numbers in brackets indicate direction where only wind is applied (wind from land).

30
60

(1.54)

(4.3) -(2.7)

11.6 11.7
-

(0.58)

150 180 210 240 270 300

6.03 6.03 2.75 2.48 2.78 3.98

8.6 8.6 5.8

24.6 24.6 13.0 13.1 13.9

,

.

5.5
5.8 7.0

15.1

I

330

I

4.95

I

7.8

I

15.6

I

NSW simulations were performed for all of the above boundary conditions, and
these results were stored in each grid point. Before entering all the results into the database, all results were transformedhnterpolated into one grid representation of the considered area with a resolution of 500 m x 500 m grid. Figs. 9.4 and 9.5 present two examples of wave fields from the simulations from Jeddah.

9-12

Fig. 9.4

50 year simulation. Wave direction 180"
(Gridspucing 500 m)

120 110 100

90
h

E 80

0

e .a E+
Q

70
60

v

._ b 50
40 30
20

10
0 0
100 200

+
5
300
400

493

Fig. 9.5

50 year simulation. Wave direction 90" (windfrom land).

9-13

Then this 500 m grid was transformed into the 1500 m reference grid (the same as for the current and water level data). This step was not done by simple interpolation, but by identifying the maximum wave height within each 1500 x 1500 m grid cell (Le. maximum Hs of the nine 500 m grid cells). The procedure was chosen to avoid unrealistically small values in the 500 m grid which could occur very close to land or reefs, resulting in too low values when merging the nine 500 m cells into one 1500 m cell. The procedure has been very similar to the one used for the base 'case of the normal wave transfer to nearshore. Based on the results found for the 50-year conditions in the local area, the 1, 10 and 100-year extremes were determined by scaling of the 50-year results based on the offshore 1, 10 and 100-year values for each direction. This procedure was acceptable considering the modest variation of the boundary values.

9.6
9.6.1

Verificatiodcalibration
General
As for the current and water level study, very little calibratiodverification data was available. However, for wave measurements camed out in 1979 in the Yanbu area (TetraTech), a suitable period for the purpose of verifying the numerical models was chosen. The period from 1/5/79 to 1/11/79 was selected because this period contained the largest measured significant and maximum wave heights, and there were two distinct significant events occurring within the period. The selection of this period was utilized to illustrate the validity of the linear scaling procedures used for both "large" waves (> 2 m) and "small" waves (< 2 m), which are both covered during this period.

9.6.2

Measurements
Measurements were conducted at two positions;

.

Shib Green, which is part of the seamost reef about 15 Ian from the shoreline.

Al Nakla, which is close to the shoreline, behind the main reefs, at a water depth of approximately 24 m.
The two positions are marked on Fig. 9.6.

9-14

kin

150-140130120-7
3

110-

,E

100-

90-

80-

7060-

50-80

90

100

110

120

130

140

150

160

Fig. 9.6

Memuring Positions at Yanbu.

e

963 ..

Simulations
For the verification of the numerical models, the period 511 - 11/1 1979 was chosen. It is noted that both of the measurement locations are nearshore points, so that a direct verification of the regional wave model OWI3G was not possible. The output from OWI3G had to be transferred by the DHI NSW model to the locations of both measurement points. The verification was thus made on the combined numerical models. For the verification period, the wave conditions were simulated in the regional wave model to provide the near-shore modelldatabase with boundary data. Fig. 9.7 shows the offshore time series, Hs and Tp, for the offshore boundary point.

Danrk Hydraubk lnnrmt IDanah Hflravlic InititUte

9-15

10
8

i

I

I
I

I

6
4

2
0

i
I I

I

I
I

" I

I

I

I

I

I

I

0000 01/05 1979

0000
01/06

0000
01/07

0000 01/06

0000 01/09

0000 01/10

0000 01/11

Fig. 9.7

Timeseries o Hs and Tp at ogShore boundary point. Veri3cation period. f

9.6.3.1 Normal Conditions

To verify the near-shore wave approach, the wave conditions at the two measuring positions was calculated by transforming the offshore time series (Fig. 9.3) using the established transfer database as described in section 9.5. This means, that each timestep of the offshore timeseries was transformed based on the results in the database. T i is the same procedure as applied when working with the database to hs get the normal wave conditions at a position in the local area. In this case it is just the measuring positions which are chosen.
Figs. 9.8 and 9.9 show time series of measured and calculated wave height and wave period at Shib Green and AI Nakla, respectively.

9-16

3

-Hs - Doto Bose __ ......... Hs - Meozurements

2
E

....

-....- ... .

0 0000 01/05 1979

0000 01/06

0000 01/07

0o:oo
0 1/08

0000
01/09

0000 01/10

0000 01/11

-Tp -Data

Base

2

Fig. 9.8

Measured and calculated wave height and wave period at Shib Green.

9-17

-Hs - Doto Base
2.0
1.5
Hs

- Meosurernents

E 1.0
0.5

i
I

i
..I

0.0 0000 01/05 1979

1

0000 0 1/06

0000 01/07

0000
01/06

0000 0 1/09

0000
01/10

0000
01/11

8
6
104

-Tp - Doto Base . Tp - Measurements
, - i - '

2
0

0000
01/05 1979

0000
0 1/06

0000 0 1/07

0000
01/06

0000 01/09

0000 01/10

0000 01/11

Fig. 9.9

Measured and calculated wave height and wave period at A1 Nakla.

For both the wave height and wave period, a good agreement between the measured and the calculated data is found at the two positions. The phase difference may be due to inaccuracies in the timing of the hindcast model or in the measurements, which from the report appear to have been carried out at somewhat irregular time intervals.

9.6.3.2 Extremes
For the extremes, the wave conditions are treated as steady state, i.e. the wave height with associated period and direction are applied directly as boundary data to the nearshore models.

9-18

To verify the approach, a situation with a wave height, Hs = 2.5 m at the offshore boundary point was identified (01/05/79, 18:OO). The associated period, direction and wind speed was, TP MWD Wind speed
= 7.9 s, = 315” and
= 12 mls.

For the simulation, the 330” nearshore model was applied. Fig. 9.10 shows the calculated wave field at Yanbu with the boundary wavelwind conditions, given as

H s

= 2.5, Tp = 7.9 s (Tm= 6.6 s) , wave direction = 315” and wind speed = 12 m/s.

Fig. 9.10

Hs(m) in the Yanbu area for o,@hore wave conditions Hs, = 2.5 m, Tp = 7.9 s. wave direction = 315” and wind speed = 12 m/s.

At the two measuring positions, the wave height listed in Table 9.4 was found.

9-19

Table 9.4 Measured and Simulated Wave Heights for verification period.

llxL--kM
Ship Green

I

Measured Hs (m)

I

Simulated Hs (m)

I

The comparison between the numerical simulation (combination of hindcast model and NSW model) is very good.
9.6.4

Conclusion The verification process showed good agreement between measured and simulated
values. The results of the comparison have thus verified the procedures and methods applied in the present study, including the transfer from offshore to nearshore areas using the DHI MIKE 21 NSW model. However, it is emphasized that the verification is for a very limited dataset for the Yanbu area only. Considering the large areas of the Red Sea covered in the present study and the variability encountered, the verification is at a very minimum and can hardly be used as an overall justification of the results. Future measurements are recommended, which should be used to verify and update the results of the present study.

10-1

10

DATABASE

10.1 Introduction
This section gives a general description of some of the important technical considerations in using the design database. The design database for the Saudi Aramco operational areas of the Red Sea is based on the hindcast results of the numerical modelling program as has been described in the previous chapters of this report. A Graphical User Interface (GUI) software program developed by DHI provides the interface between the user and the processed data database.
The Database User Guide has also been written which describes the PC computer
system requirements, installation, and database use.

10.2 Mapping and Grids
The entire Red Sea map was digitized and interpolated to a fvted grid based on the Universal Transverse Mercator (UTM) projection zone, UTM-37. It was necessary for regional modelling purposes to relate the digitized map to a single UTM zone. UTM zone 37 was the logical choice since it covers the largest portion of the Red Sea. Since all local area maps were extracted from the entire Red Sea map, the local maps are also best represented by UTM-37. This poses a mapping distortion problem for the extreme northern and extreme southern portions of the Red Sea which are in zones 36 and 38 respectively. When viewing the map of the entire Red Sea, it will only be presented in either geographic (IongitudeJatitude) coordinates, or in UTM-37 coordinates. The user should note that the UTM coordinates for the large map are only valid for the area which lies within zone 37. However, the local area grid maps, even though based on zone 37, can be viewed in their proper zones. Jeddah, Rabigh and Yanbu are in zone 37, so they can be viewed correctly in either the Red Sea or the local maps. Duba and Jizan, however, are not presented correctly regarding UTM coordinates on the regional Red Sea map. When Duba and Jizan are viewed in the local area maps there is only a small distortion error when presenting them in zones 36 and 38, respectively. This is because the origin of each map is defined by the true geographic coordinates. Distortion errors increase moving away from the origin. When viewing the Duba local area using zone 36, at the farthest point away from the origin, position emors of less t a 50 m may be realized. Viewing the J h local area using zone 38, the hn errors are less than 150 m at the point farthest from the origin. These errors are considered negligible relative to the grid resolution (1500 m) which is presented.

Damk Hydraulsk InrtnutIDanirh Hydraulic IhstRuIe

10-2

10.3

Selection of Grid Points
Most of the results stored in the database are related to one of the map grid points. Although great care has been taken to accurately digitize the Red Sea bathymetry, there are certain cautions which should be considered when retrieving design data from a particular grid point. As has been stated in previous sections, there are numerous areas of uncertainty in the chart depths for various areas of the Red Sea. When selecting points, the user should not rely solely on the map position or solely on the water depth, but should instead first select the desired position, and then judge whether the depth is representative of what is expected for that position. If the depth is not what is expected, then the user could try to move a few grid positions around the desired location to see if a more representative depth can be located. In doing this, the user should also consider the relative position of the desired location to nearby land borders or reef formations where a change in position could introduce new uncertainties. This is especially true for wave data since the wave models do not include the effects of diffraction. With a grid resolution of 1500 m, it will be common that one grid point will seem too shallow, and the adjacent point too deep. The user will have to judge which point is more representative to the application at hand, -and make adjustments to the results as necessary.

C a t depths are interpolated to a fvred grid of 1500 m. Depths are not absolute hr
values for an exact position, but are more like an average depth for a 1500 m square surrounding the point. So the depths given should not be used to determine locations. The depths are more an indication of average conditions, and are representative of depths which were used in the numerical modelling.

10.4

Reliability of the Database Results
Although state of the art numerical models were used throughout the study program, the lack of available reliable measurements made it impossible to calibrate or verify most of the model results. Wherever uncertainties were a concern, conservative assumptions were attempted. However, these applications of conservatism should not be considered as factors of safety. The user shall apply whatever factors necessary as warranted by code, local law or common engineering judgement.

10.5 Discussion of Presented Results
In general, the user must first select a point on the map before retrieving information about a location. Depending on the type of data to be presented and the location selected, the results can be presented in various way. It should be noted that the results are not interpolated to positions between grid points.

10-3

Results from Regional Red Sea Map
When retrieving results from the large Red Sea map, the data presented will come from one of the 6 reference points, representing the area delineated by the regional dividing lines shown 0 1the map. So within a particular Red Sea region, the 1 results will always be the same no matter which point is selected. The user is warned that for retrieving statistics within a local operation area, first select the local map. If the user selects a point in the local area, but instead selects the location from the large Red Sea map, the data will come from the single Red Sea reference point and not from the local area database.

Results from Local Area Maps
When retrieving results from the local operational area maps, the data are presented differently depending on if it is wind, currents, levels or waves. If it is wave data, each grid point will have a unique set of results, except for the time series plots of the individual storms. The storm time series plots are representative of a single position at the offshore boundary of the area. If it is wind data, result$ are only presented for 2 positions, at the offshore boundary and at a position near the land boundary.

Current and level data are retrievable from each grid point within a local area, but it is stored in a way so that groups of grid points which would have similar results contain the same results file.
Certain other information can be retrieved without selecting a location, such as the animation of significant wave heights for the 30 storms, the summary storm list, and the miscellaneous data.

10.6 File Locations, Types and Formats
10.6.1 Data Directories

If the database is stored on hard disk, the files are stored in a directory structure as follows.

10-4

Main Directory
c:\redsea\

Map Directory c:\redsea\maps Executable Program Directory c:\redsea\bin Menu System Directory c:\redsea\menu

UNIRAS Graphics Directory c:\redsea\uNras
Temporary File Directory c:\redsea\temp Main Data Directory c:\redsea\data Data Sub-directones c :\redsea\data\redsea
\currents
\nOIllIal

(contains Red Sea regional specificfiles)
\extreme \levels \normal \extreme \wind \normal \extreme \waves \extrem e

(contains anim'on files)

c :\redsea\data\jizan \currents
\ n o d \extreme

\levels \normal \extreme \wind
\normal

c:\redsea\data\jeddah c :\redsea\data\rabigh c :\redsea\data\yanbu c :\redsea\data\duba

\etc.. ... \etc.. ... \e&. .... \e&. ...

10-5

D62 \419 \572 \710 \E77 \lo06 D95 \335 \630 \632 \681 \699 \759

\116
\944 \946

(containswave transform maps and wind and wave time series) (rtfers to OW grid #) (ref. pt. A) (ref. pt. E) (ref.pr. C) (ref.pt. D) (ref. p f . E) (ref. p f . F) (offshore bound. Jizan) (nearshore ref. at J i m ) (offshore bound. Jeddah) (nearshore ref. af Jeddah) (offshore bound. Rabigh) (nearshore ref. at Rabigh) (offshore bound. Yanbu) (nearshore ref. at Yanbu) (offshore bound. Duba) (nearshore ref. at Duba) (dir. din. factors and extreme valuesfor wind ond waves) (containst m series of norm waves) ie (refers to O W grid #) (ref.pt. A) (ref. pt. B) (ref. pt. C) (ref. pt. D ) (ref. pt. E) (ref. p f . F) (offshorebound. Jizan) (nearshore ref. at Jizan) (offdore bound. Jed&h) (nearshore ref. at Jeddah) (offshorebound. Rabigh) (nearshore ref. at Rabigh) (offshorebound. Yanbu) (nearshore ref. af Yanbu) (offshorebound. Duba) (nearshore ref. at Duba)

c:\redsea\data\exaeme c :\redsea\data\30stonns
D62 \4 1 E72 \710 \877 \lo06 D95 \335 \630 \632 \681 \699 \759 \776 \944 \946

10-6

-

10.6.2 File Types and Formats

There are numerous file types used to store the various database results. A general description follows.
Name
*.dt2 *.Et2 *.dt0 *.C@ *.dU General Descriotion Results stored in a 2-d grid Spec. file for .dt2 file Results stored in a I array d Spec. file for .dt0 file Wave transformation stored in a 2-d grid varied use *.Ut Video animation file *.vdo Tidal constituents *.con Tide reference levels *.lvl MIKE 21 program execution spec. files *.inp Time series of normal waves *.wav Time series of normal wind *.win Miscellaneous data misc.txt Storm summary list storms.txt Offshore boundary wind persistence P 1 O Nearshore boundary wind persistence PO2 Offshore wind statistics wind-off Nearshore wind statistics wind-nsh Extreme current and level values P* waveextrmt Red Sea and offshore boundary extreme wave values windextr.txt Red Sea and local area extreme wind values region-d.dt2 Map of extreme current directions region-s.dt2 Map of extreme current speeds Map of normal or extreme water levels region.dt2 Map of extreme normal currents region.dt2 Time series used for wave rose plotting wrose.dt0 Spec. file for time series for wave rose plotting wrose.ct0 Time series used for wind rose plotting rose.dt0 rose.ct0 Spec. file for time series for wind rose plotting

Format
binarv. binary; binary, binary, binary ASCII binary, ASCII ASCII ASCII, binary binary ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII ASCII binary, binary, binary, binary, binary, binary, binary. binary, MIKE MIKE MIKE MIKE 21 Tvoe2 21 <pe 2 21 Type 0 21 Type 0

MIKE21

MIKE21

MIKE 21 Type 2 MIKE 21 Type 2 MIKE 21 Type 2 MIKE 21 Type 2 MIKE 21 Type 0 MIKE 21 Type 0 MIKE 21 Type 0 MIKE 21 Type 0

11-1

11

RED SEA EXTENSION - NEW BATHYMETRIES
The work performed under the original contract of the Red Sea Hindcast Study was based on bathymetry data derived mostly from British Admiralty charts. Due to uncertainties in chart accuracy and the later discovery of Saudi Port Authority (SPA) charts which provided a finer scale resolution and were based on more recent surveys, an extension to the original contract was undertaken to improve confidence in design data in the Duba and Jizan local areas, where the highest uncertainty existed in the Admiralty charts.
This chapter describes the work done under the contract extension as regards new bathymetries, and presents comparisons of results where appreciable changes in results are found.

11.1 New Bathymetries
SPA charts of the Saudi Arabian coastline of the Red Sea were available at a scale of 1:150000. SPA charts 12 and 13 were used to digitize the local area in and around Duba. SPA charts 26 and 27 were used to digitize the local area in and around J i m . Supplemental data provided by SA was also used in the local areas where applicable. SA charts 8299 and 8299A were used in the proximity of the Jizan port. These SA charts were used in the previous phase and are described in Chapter 4. At the beginning of this phase, SA supplied charts and digital data of a bathymetric survey of the operational area near Duba. This data was also incorporated into the new bathymetry at Duba. The digitization methods used are the same as were used previously, and the new bathymetries have been updated in Chapter 4. For the purpose of visualizing differences between old and new digitized bathymetries, Figures 11.1 and 11.2 show plots of difference between new and old bathymetries for Juan and Duba respectively. The bathymetry difference plots indicate changes in water depths only and do not represent land differences. The grey coloured areas shown are the land areas produced from the new bathymetry. Land changes between old and new bathymetries are only very minor except for the inclusion of a few small islands in the southern portion of the Jizan area in the new bathymetry. Islands which appeared in the old bathymetries also still appear in the new bathymmetries. The major differences being the deeper area in the northern section of the Jizan area, and an overall shallowing of the reef/island formation in the southern area of the Duba local area. In general, the new Jizan bathymetry is deeper, especially in the northern approaches, and channels are resolved in greater detail.

11-2

New

-

Old (m)

Fig. 11.1

Difference between new and old bathymetry ar Jizan. (Gridspacing=ZSOm).

DanL Hydraulrrk Inslil~l /Danish Hydraulic lnllilUle

11-3

(Gridspacing 250 m)

300

280

260

240

220

200

180

E
0

160

._ F
v .-

140

v

5

120

100

80

60

40

20

0 0

20

40

60

80

100

120

Fig. 11.2

Difference between new and old bathymetry at Duba. (Gridrpacing =250m).

11-4

11.2 Effect of the new Bathymetries on the Hydrodynamic Conditions
The HD modelling was re-done for the Jizan and Duba areas using the new bathy' metries interpolated to a 15OOm grid spacing. To re-run the local area HD models, the regional Red Sea HD model also had to be re-run to provide new boundary conditions into the local areas. The regional model also incorporated the new bathymetry information into the 6ooOm regional bathymetry. The new regional bathymetry is presented in Chapter 7.
Generally, differences between old and new water level predictions were found to be negligible and are not presented here but are updated as necessary in Chapter 7.

To compare old and new results, difference plots of maximum current speed is shown in Figurs 11.3 and 1 . for JiZan and Duba respectively. These maximum 14 currents are derived from the maximum of all maximum current speeds from the 30 storms combined. Differences between old and new current speeds and directions are random and are reflected by differences in bathymetries. In general, where the bathymetry has become deeper, the current is reduced, and where the bathymetry has become shallower, the current has increased. An example of the increase in current speed can be seen in the southern part of the Duba area, where shallow reefs show increases in the maximum current speed up to 0.4d s . Conversely in J , m where the new bathymetry is generally deeper, the current speeds have been reduced over a very large area.

11-5

(Gridspacing 1500 rn)
75

70

65 60
55
50

45

25

20

Nev#

*bo"*

Old ( r n / s )
0.60
OB0
0.50

15

0.50-

10

5

- 0.ro - 0.30 - 0.10 0050.10 -0.35 - 0 0 ., -0.m - -0.05 -0.20 - -0.10 -4.10 - -0.20 -0 0 ' - -0.30 -0.50 - -3.10
0.10
020

00 '-

0.10

-0.60
Bib.

-

-0,SO
LO"*

-0.110

0

0

5

10

15

20

25

30

35

40

45

0
Fig. 11.3 Difference between new and old maximum current speeds at &an.

11-6

(Gridspocing 1500 m)

NevY

-

Old (rn/s)
0.40

*bo". 0.50 0.20

O.,O-0.05 -0.10
-0.20 -0.x
-0.LO
WO.

-

0.40 0.30
0.20 0.10

0.05-

- 0.05 - -0.05 - -0,to - -0.20 - -3.m

-0.AO La<

0

2

4

6

8

10

Fig. 11.4

Difference between new and old maximum current speeds at Duba.

iPglminmi.drmMcmi.*

Omsk Hydsulirk InrtitutlDanirh Hydraulic lnrtilule

11-7

0

11.3

Effect of the New Bathymetries on the Near-Shore Wave Conditions
General
The effects of the bathymetry changes on the near-shore wave conditions are presented, concentrating on the changes in the local areas i.e. the database values.

To illustrate the changes, the 50 year wave conditions calculated with the new versus old bathymetries are discussed.

11.3.1 DUBA

General
The main changes in the Duba area are caused by introducing smaller water depths in an area reaching from the An-Na'man Island and about 6 km NNW of the Island. According to the new charts, the area is a reef area with water depths close to zero, where the depths based on the first digitisation were between 5-10m depth. These changes cause the wave conditions to change, not only in the reef area, but also in the area in lee of the reefs. This lee area changes with the wave direction. Smaller changes are introduced due to changes in some scattered reefs about 20km NNW of the An-Na'man Island. Again it is in the area sheltered by the reefs which experience changes in the wave conditions, when comparing with the results obtained with the old bathymetries. Further, the depths close to the main land are generally increased in the new bathymetries resulting in higher waves.

To gain an understanding of the importance of the changes in wave climate due to the changes in bathymetry, it is instructive to first look at the wave height rose plot to see the predominant wave directions for the Duba area. The wave height rose plot for the normal wave conditions offshore of the Duba area is shown in Figure 11.5. The wave rose shows that most waves arrive from the directional sector centered around 300 degrees. This also corresponds to the direction of the extreme wave heights, as was previously shown in Table 8.1.
WindIWaves from land (0 " 120 ") For the direction with windlwaves from land (0-1200) the changes within the area of interest are small and concentrated to the shallow area NNW of An-Na'man Island itself, where the wave height is close to zero. Fig. 11.6 shows the wave s,) height difference plot (H%e,,, - H,, for the direction 120'.

-

11-8

N

m
Abra

2.0
1.5

-

- 2.5
2.0

2.5

0 0

bx

i.0-1.5 0.5 - 1 0 .

0.0

- 0.5

0.0

Fig. 11.5

Wave Rose plot for normal conditions offshore of Duba.

IW701-l7flt6e-l.dIWEMCht

11-9

(Gridspacing 1500 rn,

Hs,new

-

Hs,old (rn)

@ 1 3 @

1.5-

1.3/.I0.90.7 -

1: ,,
1.7 15

Ll
0.9

0.5 0.3

-

0.7
0.5

0.1-0.1-

0.3
0.1

-0.3
-0.5 -0.7 -0.9 ->.I

-0.1
-0,

- -0.5 - -0.7
-0.9
-1.1

-1.3 -1 5
-i.7 -1.9

- -,.5
-1.7 -2.9
I

-1,

. b B

-"

0

2

A

t

i

R

l

O

1990/01/01 0000'2

Fig. 11.6 HsneW - HsOM (m), 50 year wave conditions, Wave Direction 1200.

Waves from southerly directions (150-210)
In general, the wave height is decreased in the area as a result of the sheltering from the reef area NNW of the An-Na'man Island. The maximum reduction appears for 210°, where the reduction is up to 0.7 - 0.9m corresponding to a 4050% reduction. Fig. 11.7 shows the wave height difference plot for the direction 2100.

0
I997Ql-17nMbl.dr~EMCl.~

11-10

(Gridspocing 1500 m ,

301
25
Hs,new - Hs,old

2
Fig. 11.7

2

d

6

8

10

1990/01 /o 1 00002 1

Hsnew - Hsold (m), 50 year wave conditions, Wave Direction 2100.

Waves from westerly and northerly directions (240-330')
The general trend for the directions 240-330° is a decrease in the wave height in the lee of the reef area plus an increase of the wave height close to land in he northern comer of the bay. The latter is due to increased water depths in the new bathymetries. For the direction 300' the increase is about 0.5m corresponding to about 20%. Fig 11.7 shows the wave height difference plot for the direction 3000.

11-11
(Gridspacing 1500 rn,

30--

25 -

20 h

E

0

5: -

._ ._ 4
v

i

Hs,new
15.Pow

- Hs,old (rn)
1.9 1.9
1.7

i.7-

b

.
10-

1.51.31.1-

I.5 1.3
1.1

090.705-

0.9 0.1

0,J0.1-

05

-0.i -0.3 -0.5 -0.7
-0.9 -I./

-

03
01

- -05 - -9.7 0 -,., - --,.I9 - -i.5
-15--1., -1.7

- -0 I - -03

-,.9
8.b.

-

-1.7 -19
Lid

0

2

4

6

8

10

1990/Ol/Ol 000030

Fig. 11.8

Hs,,,

- Hsold (m), 50 year wave conditionr,

Wave Direction 300‘

0

11.3.2 JIZAN

General
The main changes in the wave conditions in the Jizan area are caused by increasing the water depths in the north-westem part of the area, especially in the ‘channels’ north of the islands, see Section 11.1.

In the southern part small islands and reefs have been better resolved, generally resulting in reduced wave heights due to the sheltering effect of the islands and reefs.
For reference, Figure 11.9 shows the normal condition wave height rose plot indicating the predominant wave directions in the Jizan area. The wave rose shows that most waves come from the directional sectors centered around 180 and 330 degrees. This also corresponds closely to the directions of the extreme wave heights, as was previously shown in Table 8.1.

Danrk Hydraulirk InilitullDanih Hydraulic In~lilule

11-12

N

rn

Ab"*

1.0

- 2.0 1.0-1.5
1.5

- 2.5 - l.0 - 0.5
00

2.5

0.5
0.0

0 Bebr

a

Fig. 11.9

Wave rose plot for the normal conditions offshore of Jizan.

I9?7Q1-17flb%l.drW€UC97Qhb

Qdmk HydrauLrk InP~lullDdnbh Hydraulic Insliiule

11-13

WindlWaves from land (0-1200)

In general the changes in the wave conditions are small for the directions with
windlwaves from land as the wind generated waves are relatively small and not very sensitive to water depth changes. For the 50 year simulations the maximum wave height in the local area is about 1.5m and the maximum changes are in the order of 0.2m. In Fig. 11.8 the wave height difference plot is shown for wave direction 90'
(GridsDocinq 1500 rn,

70

60

,-. E
50
0

z

j40
v

Hs,new

-

Hs.old (rn)

5 30
20

.t:

10

0 0
10

20

30

A0

1990/01/01 0 0 0 0 0 s

Fig. 11.10 Hs,,,

- HsOu (m).50 year wave conditions, Wave Direction 900.

Waves from southerly directions (150-210')

In the southern part of the local area a decrease in wave height is found as a result
of the better resolution of reefs and small islands in the southern area of Jizan, partly shelterkg the area north of these islands. In most of the area the decrease is small (less than 0.3m). In local areas an increase of up to 0.3m is observed. Fig. 11.11 presents the wave height difference plot for the direction 210'.

0

11-14

(Gridspacing 1500 m,

Hs,new

-

Hs,old
i.5

1
0

i.31,)-

I ;:; 1.5
L3
,.I

0,s-

o.,

::; ::; 1
0.1-

-

0.5

0.1
0.t

-0.1-

- -0.i - 0 5 - -0,J -0.7 - -0, -0s - -0.7 - > . I - -0s -1.5 - -1.) -0.1
-3,
-9.3

-1.7 - - I , ,

8.0.

-1.7 -1.9

1-1;

Fig. 11.11 Hs,,, - H, (m), 50 year wave conditions, Wave Direction 2100 s,

Waves from westerly and northerly directions (240 - 3000)

Due to the increased water depths in the northern part of the Jizan area the wave heights increase for waves from westerly and northerly directions.
The more the waves tend from north, the further south in the local area the effect of the locally increased water depths reaches. The maximum increase in wave height is found for direction 300°, where the wave height in the most northern part of the area is almost doubled from about 1.5 to 3.0m, see Fig. 11.12. Further it can be seen, that the area of increased wave height extends 50-6Okm SE in the local area.

11-15

The significant increase in wave height is due to the general deepening of the Jizan area over a large distance. Using the old bathymetry, much of the long period energy generated from the far-field offshore was dissipated due to interactions with the sea bottom. With the increased depths of the new bathymetry, the bottom dissipation has been significantly reduced, refraction effects have been reduced, and the longer period wave energy penetrates farther into the area. In the southern part of the area local decreases in the wave heights are found due to the sheltering effect of the small islands and reefs in the south.
(Gridspacing 1500 rn,

Hs.1new - H s d d (rn)
.bo"* 1.71.5I, .?.I-

1.9 1.5
1.7

1.5

>,I
1.1

0.90.705-

09 0.7

0,J- 0,s 0 . 1 - 03

-01
-03 -05 -0.7 -03

- -0 1 - -0,J

-

0.1

- -0.5
-

-a,>
-1.5

- -0.7 - -3.9
-2.3

-'.,--/.I

Fig. 11.12 Hs,,, - Hsord (m), 50 year wave conditions, Wave Direction 3000

Danrk HydrauliskInrtitutlDanirh Hydraulit lnilil~le

12-1

12

RED SEA EXTENSION DIFFRACTION STUDY

-

12.1 Introduction
The hindcast of waves in the Red Sea was based on numerical models which do not include the effects of diffraction. This is a general problem in that there exist various numerical wave models which can individually account for the common wave phenomena (Le. refraction, wind growth, reflection, diffraction, shoaling, breaking, bottom friction, seiche), but there exists no singular wave model which includes the effects of all these phenomena combined into one model which at the same time can cover a “large” area at a “fine” resolution in a “reasonable” amount of time. Due to the presence of islands and reefs in the local areas of the Red Sea, a study has been undertaken to asses the importance of diffraction as related to the Red Sea database. This study attempts to compare model results in a “small” subgrid area in the Jizan local area where we would expect to detect the effects of diffraction, and then to analyse the data and compare to the model results without diffraction, at both a f i e grid resolution and the same resolution as presented in the database.

12.2

Study Area
The study area at Habar Island was selected based on SA’S indicated area of interest, on the results given in the database, and to some extent, on the updated NSW results using the new bathymetry for Jizan. The database and the revised NSW results indicated that the waves approaching the area of interest are largest and have the longest period coming from 300 degrees. Inside the study area, however, the waves from 300 degrees have been reduced significantly by the sheltering from Habar Island. To model &is island and the sheltered area behind required setting up a larger model than was originally anticipated. The model area is approximately 6km wide x lOkm long using a grid spacing of 7.5m. The model grid is shown in Fig. 12.1.

12-2

(Gridspacing

7.5 rn)

0

200

400

600

800

1000

Fig. 12.1

Bathymetry of the dzj?action study area. (Gnakpacing. &=dy=7.5m).

12-3

12.3 Boundary Conditions
Boundary conditions-were extracted from the 50-year NSW results for wavei approaching from 300 degrees. The boundary conditions are shown in Table 12.1. Note that the boundary condition is an averaged condition extracted from the NSW results, and does not include post-processed factors of conservatism which are applied to the values in the final database.

.

.

-

Table 12.1

Subgrid boundary conditions

12.4

Wave Model
Due to the increased size of the model area and the water depths of the extended model area (larger than 5Om in the SW comer), it was not computationally practical to perform the modelling using the Boussinesq Wave (BW) model without making schematization assumptions necessary to guarantee numerical stability and model validity. The main problems are regarding minimum and maximum water depth limitations for the spectral conditions imposed at the boundaries versus computation time. To avoid making these schematizations, the Elliptical Mild Slope (EMS) model was investigated. The EMS model is computationally faster than the BW model, includes diffraction effects, but only solves for uni-directional monochromatic waves. Using the EMS model, result fields for 3 different spectral frequencies were combined by superposition of wave energies to better resemble an irregular wave field. However, after much analysis, and comparison to results of the NSW model and also to results of a small BW subarea, it was found that the EMS results were not providing results representative of a typical sea state condition expected in the Jizan area. Figure 12.2 shows the combined relative wave height field found from the EMS modelling. It can be seen that interference patterns between monochromatic wave trains are prevalent, and areas of wave amplification to the north of the island dominate the wave field and overshadow the effects of diffraction, making it very difficult to make a useful comparison between EMS and NSW results regarding diffraction effects. The high degree of interference and amplification was concluded to be due to the absence of directional spreading which m o t be included using the EMS model. With these findings it was decided to schematize the model bathymetry and boundary wave spectra i a n way that the BW model could be used. The outcome of this schematization is that all water depths deeper than 25m were made equal to 25m, all water depths shallower than 5m were made equal to Sm, directional deviation from the mean

12-4

wave direction was constrained to +/- 10 degrees, and that the boundary wave spectrum was cut-off for all periods shorter than 8.5 seconds. Generally speaking, the diffraction of shorter period waves influences a much smaller and more confined area nearer to the lee of the island since diffraction effects scale by the wavelength, and the shorter period waves have shorter wavelengths. It is our expectation and our experience that the inclusion of the full frequency spectrum and full directional distribution would produce results where the areas of amplification are even further reduced. The schematized bathymetry is shown in Figure 12.3. The time step used in the BW modelling was 0.45 seconds corresponding to a Courant number of 0.9.

.

12-5

(Gridspacing 7 5 rn) .

Rel. HrnO (rn)

0

200

400

600

800

1000

Fig. 12.2

Relative significant wave height using combined EMS results.

I59ldl~l7n6661.drllEMC9nIl.lb

DanrX Hydiduli$klnrtitul/Danish Hydraulk lnililule

12-6

(Gridspocing 7.5 rn)

800

700

600

'0 400

F

._ : : L & 300 ?
200

x

100

0 0 200
400
'

600

800

1000

Fig. 12.3

Schematized bathymetry.

12-7

To make valid comparisons with NSW results, it was also necessary to re-run the
NSW models using the same bathymetry modifications. Although differences do exist between identical NSW results using the original bathymetry and the modified bathymetry, the differences as regarding wave conditions behind the island are found to be negligible. The result fields for both NSW simulations are shown in Figures 12.4 and 12.5. The wave heights are normalized with the wave height at the upstream boundary.

.

12.5

Analysis
For the diffraction analysis both the BW and the NSW model has been run with the schematized bathymetry. The analysis is performed in a 3 step comparison.

Step 1: BW vs. Fine Grid NSW
The first analysis step is to compare the results of the BW model with an NSW simulation using an “identical” model set-up. This means that the following features of the NSW that were included in the study are suppressed: wind energy transfer, wave breaking, and bottom dissipation. This is useful in illustrating the actual differences with and without diffraction at the same scale of resolution. Figure 12.6 shows contours of steady state relative significant wave height fields for the BW model. This can be compared directly to the NSW result shown in Figure 12.5. The 2 wave fields compare very well. To visualize better the difference between results of the 2 models, line-section plots of relative wave height are presented for various sections through the model area where diffraction effects would be expected. Figure 12.7 indicates the location positions of the various line sections. Figures 12.8 and 12.9 show the line section plots. There are 2 main areas where differences are observed. At line section XA at about meter 6500, a sharp peaked shoal is observed. This shoal, though not of interest as regarding a study in diffraction, is mostly due to the schematization of the wave spectra, and would be reduced if higher frequency waves were included in the spectrum. There is also some smoothing in the NSW result due to the coarser grid resolution in the y-direction (dy=30m), which contributes to the local differences in the calculated wave heights.

12-8

0

1000

2000

3000

4000

5000

6000

7000

8000

Fig. 12.4

Relative signifcanr wave height using NSW fine grid model with no schemarizarion.

Surfoce Elevotion

0

1000

2000

3000

4000

5000

6000

7000

8000

0

Fig. 12.5

Relarive significant wave heighr using NSWfine grid model using schematized bathymetry.

IWIOI.17i7~l.drVEMCI.ab

Danrk Hydraubrk lnrulvtlDanirhHydraut lnrlilule

12-9

(Gridspocing 7.5 m)

Surface Elevation

0

200

400

600

800

1000

a

Fig. 12.6

Relative significant wave height using the EW model.

19914l~17fl6661.drkiEEMC9lQl.~

Danrk Hydraulirk 1nriiturIDanirh Hydraulic Inmuie

12-10

t

I

0
Lo

0

J
0

0

O

Y

O

0

Fig. 12.7

Location map of line seaions in model domain.

tm~t-i~n~t.drmucm~.~

12-11

........

xc

- BW

.

.--

0

1000

2000

3000

4000

5000

6000

7000

8000

[ml

........

xb

- BW

0

1000

2000

3000

4000

5000

6000

7000

8000

[ml

........
1.6

xa

- BW

1.2
0.8

0.4
0 .o

0

1000

2000

3000

4000

5000

6000

7000

8000

1990/01/01

000000

[ml

Fig 12.8

Line section comparisonr between the BW m d e l and the fine grid NSW m d e l for cross sections in the X-direction.

IWlQl-17n6bbl.&.WEMCl.h

DdmL HydraulsL InlitutlDantlh Hydraut lnititute

12-12

..,.....
16 .

yd

- BW

12 .
0.8 04 . 00 . 0

500

1000

1500

2000

2500

3000

[ml

... .....
16 .

yc

- BW

1.2
0.8 0.4 00 . 0

500

1000

1500

2000

2500

3000

Iml

........
16 . 12 .
0.8 04 .

yb

- BW

00 .

........
1.6

ya

- BW

-.

12 .
08 .
0.4

0.o 0

500

1000

1500

2000

2500

3000

1990/01/01 oo:oooo

[ml

Fig 12.9

Line secrion comparisons between the BW model and the fine grid NSW model for cross sections in h Y-direbion. e

12-13

The other important area of observable difference, and a diffraction effect, can be observed looking at line-section XC from between meter 3000 to 5000. In this relatively small area, the wave heights are increased by about 40% from the NSW results. However, in other sections, and especially in the direct shadow behind the island, the results compare, for all intents and purposes, identically.

Step 2: BW vs. Coarse Grid NSW
The second analysis step is to compare the results of the BW model with an NSW simulation using a 125m x 750m grid and using a similar setup as to what was used in the Hindcast study. This is useful in illustrating the differences which grid scale, wind-wave growth, bottom dissipation, wave shoaling and wave breaking contribute to the NSW model, which are absent from the BW diffraction model. Figure 12.10 shows a plot of the relative significant wave height using the coarse grid NSW model. Comparisons are made of the variation of wave height amplification factors along the same model line sections as shown in the Step 1 comparison and are shown in Figures 12.11 and 12.12. It is clear that the effects of wind generation behind the island contribute significantly to the overall wave field.

(Gridspocinq 125 rn)

Surface Elevation

0

10

20

30

40

50

60

69

0

Fig. 12.10 Relative signwcant wave height using the coarse grid NSW model.

12-14

.......,.
1.6 1.2

-

xc xc

- BW - NSW

125

.

08 .
0.4

0.0
0
1000 2000

3000

4000

5000

6000

7000

8000

[ml

.........
1.6

-

xb xb

- BW - NSW

125

1.2

08 .
0.4

0.0
0
1000 2000

3000

4000

5000

6000

7000

8000

[ml

.........
1.6 1.2

-

xo

xa

- BW - NSW

125

0.8
0.4

00 .
0
1000

2000

3000

4000

5000

6000

7000

8000

1990/01/01

000000

[ml

Fig. 12.11 Line section comparisons between the BW model and the coarse grid NSW model including wind and dissipation for cross sections in the X-direction.

12-15

.........
1.6 1.2
0.8
0.4

yd

- BW

.

00 . 00

.........
1.6 1.2

yc

- BW

08 .
0.4

-_-,
nn

0

500

1000

1500

2000

2500

3000

[ml

.........
1.6

yb

- BW

1.2
0.8

......... . .

0.4 00 . 0

.........
1.6

-.

yo

- BW

1.2

0.8
04 .

0.0
0

500

1000

1500

2000

2500

3000

1990/01/01 000000

[ml

Fig. 12.12 Line sem'on comparisons between the BW model and the coarse grid NSW model including wind generation and dissipationfor cross sections in the Y-direaion.

12-16

Step 3: BW vs. Database Results
The fml analysis step shows how results presented in the database may under-or-over-predict wave heights in the sheltered areas behind the island, when we compare the results in a 15OOm grid. The analysis technique used in the database was to interpolate the NSW results to a 50Om grid, then to select the maximum wave height of the 9-5OOm grid points within a single 1500m grid point, and then to report that value for the entire 1500 m grid area. Figures 12.13 and 12.14 shows the same line sections of relative wave height comparing the BW to the NSW results analyzed as if to be stored in the database. The figures also include the fine grid NSW and coarse grid NSW simulations results. It can be seen how the "database" results give conservative values to all of the fields shown compared to the BW model results.

.

..

12-17

.
c xc

.

1.6 1.2

----......... -

- NSW OB
- NSW - BW
125

xc xc xc

- NSW

08 .
0.4

00 .

0

1000

2000

3000

4000

5000

6000

7000

8000

1.6

----......... -

-

[ml
-

xb-NSWOE xb NSW 125 xb EW x b NSW

-

1.2

0.8
0.4

0.0
0

1000

2000

3000

4000

5000

6000

7000

8000

-__-_......
...
1.6 1.2
xo xa

[ml
NSW OB

- xa-NSW

- NSW KO - EW

125

0.8
0.4

0.0 1
0

I

1

1

1

1

1

1

1

I

1000

2000

3000

4000

5000

6000

7000

8000

1990/01/01

000000

[ml

Fig. 12.13 Line section comparisons behveen the BW model and the coarse grid NSW model results prepared similar& LL( done for the &tabme for cross-sections in the Xdireaion.

-_-_.........
1.6 1.2
0.8 0.4

12-18

yd-NSWC05 yd - NSW 125
yd

- BW

nn

I._

,

0

500

1000

1500

2000

2500

3000

yc - NSW ---_- - NSW DB yc 125 ......... yc - BW

-

tml

1.6 1.2
0.8
0.4 nn

_ . I

,

0

500

1000

1500

2000

2500

5000

-_-_.........
1.6 1.2
0.8

[ml

yb-NSWOB yb NSW 125 yb EW

-

0.4

"."

nn

,
0
500
SW ----- y a --NNSW W125 yo

I

1000

1500

2000

2500

3000

[ml

-

.........
1.6 1.2
0.8
0.4

yo

- EW

0.0 0 500

1000

1500

2000

2500

3000

1990/01/01 000000

[ml

Fig. 12.14 Line section comparisons between the BW model and the coarse grid NSW model results prepared similarly as done for the &abase for cross sections in the Ydirection.

12-19

0

12.6 Discasion and Conclusions
In the Step 1 comparisons it is evident from the line sections that in the immediate shadow zone behind the island there is little difference between the model results, whether diffraction is included or not. The small area where diffraction can be observed does not extend more than about 1 lan behind the island, or less than one database grid point. This point helps to explain how the method of analysis and presentation used in the database provides a level of conservatism which overshadows the relatively small scale effects of diffraction. The reason for having the small effect of diffraction as compared to the results of a uni-directional, single frequency wave is the influence of irregular waves and especially the directional distribution of these on the diffraction pattern which is smoothed out.

.

In Step 2 the most interesting point is to see the influence of wind to the NSW results.
In Step 3, the comparison illustrates the degree of conservatism which exists in the database where there is a rapid variation in wave conditions in a relatively small area. Considering the inherent limitations of a rather small and local investigation, it may not be prudent to draw too wide from conclusions in a general sense. However, the results do not warrant or justify any increase to the database values to include effects of diffraction. Conversely, if design values are sought in the immediate lee of an island area at a fine scale, it may even be possible to reduce the database design values within a very fine subgrid area due to the built-in conservatism. However, this would have to be evaluated on a case by case basis.

Danrk Hydrauiirk InrtltutlDanahHydraulic lnrliwte

13-1

13

REFERENCES
Hodeidah Fishery Harbour. Fishery Development Project at Hodeidah. Prepared for the Ministry of Agriculture of Yemen by Danish Hydraulic Institute, July 1980. (DHI-ref 80-675) Jeddah 5, Paver and Desalination Plant, Recirculation Study. Report submitted to Fichtner Consulting Engineers on behalf of Saline Water Conversion Corporation, Saudi Arabia, by Danish Hydraulic Institute. February 1993. (DHI-ref. 6979) Yanbu Beach Resort. Development Plan and Urban Design Study for Yanbu Industrial Complex. Contract GST E-4010, part B-1. Prepared by Danish Hydraulic Institute. June 1979. @HI-ref 78-956) Yanbu Wave Study. Preliminary Design Wave Conditions for the Projected Harbour Facilities at the Domestic Refinery, Yanbu. Final Report. Prepared for the General Petroleum and Mineral Organisation (PETROMIN) by Danish Hydraulic Institute. September 1981. (DHI-ref- 81-772) Yanbu Marine Study. Field Status Report. Prepared by Danish Hydraulic Institute October 1981. (DHI-ref 81-771) Yanbu Marine Study. Field Survey Status. Prepared by Danish Hydraulic Institute. June 1981. (DHI-ref 81-771) Yanbu Marine Study. Field Measurements, Data Analysis and Conclusions. Final Report. Prepared by Danish Hydraulic Institute. June 1981. (DHI-ref 81-771) Yanbu Sediment Plume Study. Report submitted to Aramco Overseas Company by Danish Hydraulic Institute, May 1990. (DHI-ref. 6226)

Ill

.

.-

121

131

141

151

161 171 I81

0

PROVIDED BY SAUDI W C O __________________________________
Climatic Study of the Red Sea South and Gulf o Aden. Near Coastal Zone, Report f Prepared for the Commander, Naval Oceanography Command by the Naval Oceanography command Detachment, Asheville, N.C. USA, September 1982.
110i 1111
El- Sabh, M.I. and T.S.Murty, Age of Tides in the Red Sea and G l of Aden. uf Proc. Symp. Coral Reef Environ. Red Sea, Jeddah, p. 7, January 1984.

Final Report for Meteorological Monitoring Nehvork at Madinat Yanbu AlSinaiyah. Report prepared for the Kingdom of Saudi Arabia by the Ecology and Environment of Saudi Arabia Co., Ltd., April 1982.

13-2

1121

Koch, S.P. and H.M. Shabib, Design Criteria for Yanbu Expansion Facility. Prepared for Aramco in conjunction with Exxon Production Research Company, Offshore Division, EPR.9PS.89, February 1989.

I131

Marine Site Investigations, Yanbu Industrial Complex, Final Report. Report prepared for the Royal Commission for Yanbu and Jubail Directorate General for Yanbu by Tetra Tech. TC-814, July 1977. Meteorological and Oceanographic Data Report, Yanbu Industrial Complex, Interim Report # 2. Report prepared for the Royal Commission for Yanbu and Jubail Directorate General for Yanbu by Tetra Tech. TC-814, January 1978.
Murty, T.S. and M.I. El- Sabh, , Weather Systems, Storm Surges and Sea State in the Red Sea and the Guy of Aden. Proc. Symp. Coral Reef Environ. Red Sea, Jeddah, pp. 8-38, January 1984.

.

I141

1151

0

1161

Site Investigations for the Yanbu Industrial Complex, July to December, 1977. Report prepared for the Royal Commission for Yanbu and Jubail Directorate General for Yanbu by Tetra Tech. TC-814, January 1978. Site Investigations for the Yanbu Industrial Complex, Final Report. Report prepared for the Royal Commission for Yanbu and Jubail Directorate General for Yanbu prepared by Tetra Tech. TC-814, January 1978. Volume N - Meteorological & Oceanographic Data Collection, Yanbu Industrial Complex, Kingdom of Saudi Arabia. Final Report prepared for Royal Commission for Jubail and Yanbu Directorate General for Yanbu Region. Contract No. GST E 4009. August 1980.

I171

I181

1201 I211 I221 I231

Pressure and Wind Statistics for Gizan. Jeddah and Yanbu in Excel Spreadsheets. Delivered to DHI by SA. Source unknown.
Tabulated Tide Tables for 1995 at J , m Yanbu and Jiddah. Delivered to DHI by SA. Source unknown.

Wave Study Report. Delivered to DHI by SA. Saudi-Consult by Dames and Moore). Yanbu Airport Wind Scaner Diagram. unknown.

Source unknown (possibly from Source

Delivered to DHI by SA.

13-3

1241

Admirally Tide Tables and Tidal Stream Tables, Vol. 2 , published by the Hydrographer of the Navy, 1990. Der Norske Veritas, Rules for Classification of Fixed Offshore Installations, Det Norske Classification N S , Norway, 1989.
Edwards, F. J. Climate and Oceanography. In the Red Sea (Edwards, A.J. & Head, S.M., eds). pp 45-68. 1987. Morcos, S . A. Physical and Chemical Oceanography of the Red Sea. Oceanography and Marine Biology. Vol 8. Ed. M. Barnes. George Allen and Unwin Ltd., London. 1970. Pugh, David T. Tides, Levels and Mean Sea Level. John Wiley & Sons.
.

1251

I261

I27 I

I281 1291 1301 I311

Red Sea and G l of Aden Pilot. Hydrographer of the Navy, London, 1967. uf Shore Protection Manual, Vol. 1, CERC, U.S. Army Corps of Engineers, 1984.
Sultan, S.A.R., F. Ahmad and A. El-Hassan. Seasonal Variations of the Sea Level in the Central Part of the Red Sea, Estuarine, Coastal and Shelf Science, 40, pp. 18, 1995. U.S. Navy Weather Service Command, "Summary of Synoptic Meteorological Observations, East Afncan and Selected Islands Coastal Marine Areas Vol. 2", 1979.

1321

14-1

14

NOTATION
H S T
d S

%lo

mO ml m2
Tm

T z
%ax

H S
n E

D

P
MWD
Ax AY

c

C

g

h At RMS
2 V

‘tide

7
M2 01 s2 K1

significant wave height (m) peak spectral wave period (s) wind speed (m/s) significant wave height (4 * sqrt (mO)) zero’& spectral moment (m2) first spectral moment (m2/s) second spectral moment (m2/s2) mean wave period (s) mean upcrossing wave period (s) maximum wave height (m) significant wave height (m) directional spreading index wave energy directional distribution function directional distribution normalising factor mean wave direction (ddeg) grid spacing in x-direction grid spacing in y-direction Courant number -wave celerity ( d s ) gravitational constant (m/s2) water depth (m) time step (sec) root mean square distance from still water leve, positive upwards (m) total current speed (m/s) current speed, tidla component (m/s) current speed, wind component (m/s) reference depth for wind generated current = 50 m principal lunar semi-diurnal tidal constituent principal lunar diurnal tidal constituent principal solar, semi-diurnal tidal constituent luni-solar declinational diurnal tidal constituent

a

APPENDIX A
Drawings

wind Speed Area : Position : Eeight : Duration :

....................................................................................
I Wind Speed I (m/r)

Directional Distribution JIW 1870753, 1035093N 10 (m) 10 min. Average

-

-

Off Shore

I
I

Direction (Coming from, relative t o north) 0 30 60 90 120 150 180 210 240

270

300

I 330 I Dmni I

I 22.0
I

I 20.0 I I
I

I
I I

I
I

I

18.0 16.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0 1.1 -

24.0 I 22.0 20.0 18.0 16.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0

I I
I
O+

O+

0.1 0.4 . . o+
~

o+ o+
o+
0.4 0.8 0.2 0.7 1.2 0.3
~

1.7 1.7 0.2

0.5 0.7 0.2

0.1

0.1 0.2 0.5 1.7 2.8 2.1 0.2

0.2 1.2 3.4 6.7 9.6 6.8 2.0 0.1

o+
0.1 0.4
1.0

o+ o+ o+
0.1 0.3

o+
0.1 0.2 0.7 2.5 7.1 7.1 2.4 0.2

o+
0.1 0.2

o+ o+
0.3 2.0 1.8

2.3 1.3

1.4
1.1

1.3
1.4 0.1

o+

0.1

o+

0.3 1.3 5.2 7.7 3.0 0.2

0.3 1.5 4.7 11.8 26.0 34.5 19.5 1.7

I I I
I I I

I
I

I

calm : 0 . 0

wind Speed Area : Position : eeight : Duration :

Directional Exceedance JIM

-

-

Off Shore

1870753, 1835093N
10 (m)

10 min. Average

I I

I > 18.0 I > 16.0
14.0 I 12.0 I 10.0 I 8.0 I 6.0 I > -4.0 I > 2.0

>

20.0
O+

O+

> > > > >

O+
0.1 0.5 2.2 3.9 4.0
O+
O+

o+
O+

I

>

1.1

0.5 1.2 1.4

0.4 1.2 1.3

0.2 0.9 2.0 2.3

0.3 0.1 1.4 0.3 4.8 0.8 11.5 2.5 21.1 5.327.9 7.4 29.8 7.6 29.9

o+
0.1 0.5 1.6 3.8 5.1 5.1

o+
0.1

0.1 0.3

o+
0.1 0.4 1.8 2.9 3.0

o+
0.1 0.2 1.6 3.0 3.1

0.3 0.9 1.6 3.5 0.4 6.8 10.6 2.3-,14.517.7 4.1 17.5 20.1 4.2 17.8 20.3
0.1

o+

0.4 1.9 6.5 18.3 44.3 78.8 98.3

100.0 I

Calm : 0 . 0

CGmt:

SAUDI ARAMCO
RED SEA HINDCAST STUDY
Orwing m.

i

Project:

1

Danish Hydraulic

lnttltute
1996

0

fik

O d t c Wed Apr

Directionu Distribution of Winds Frequency of Occurence ond Exceedance at offshore boundory of Jizan

%de:

Init:

drk

6.1

wind Speed kea : Position : eeight : Duration :

....................................................................................
Wind Speed

Directional Distribution JBDDAE 469766E. 236402711 10 (m) 10 min. Average

-

-

Off Shore

(mls)
22.0 20.0 18.0 16.0 14.0 12.0 10.0 8.0 6.0

Direction (Coming from, r e l a t i v e t north) a 0 30 60 90 1 2 0 1 5 0 1 8 0 210 240

-

...................................................................
I I
O+

270

300

330 I Omni I

I

4.0
2.0 1.1

-

24.0 22.0 20.0 18.0 16.0 14.0 12.0 10.0 8.0 6.0
4.0

I I I
I I

0.1 0.3 1.2 1.8 2.5 1.7

o+

O+

o+ o+
0.1 0.3 0.8
1.1 0.2 0.1

o+
0.1 1.3 4.5 11.7 18.8 13.7 3.1 0.1

I

I
I I

o+
0.2 1.5 5.2 15.5 30.6 33.2 12.4 1.3

I
I

o+
0.1

o+
0.1

O+

0.2 0.7 0.8

0.4
0.8 0.2

0.1 0.2
0.4 0.2

2.0

o+

0.1

0.3 0.5 0.6 0.1

0.1 0.4 1.2 1.7 0.7 0.1

o+
0.1 0.5 1.0
0.5
O+

0.1
0.9 0.6

0.1 0.8 2.1
0.9

0.1

o+

o+

0.2 1.6 6.0 0.3 2.1 0.2

I
I I

I
1

I
I

I

Calm : 0 . 1

I I I I

22.0 20.0 > 18.0 > 16.0 I > 14.0 I > 12.0 I > 14.0 I > 8.0 I > 6.0 I > 4.0 I > 2.0 I > 1.1

> >

I I
I I I

o+
0.1 0.5 1.7 3.5 6.0 7.6 7.7

I
I I I

o+ o+
0.1 0.3

O+ 0.1

0.1
0.4

o+
0.1 0.6
O+

o+
0.1 0.5 1.3 1.5
0.1 0.3 0.7 0.8

O+

0.5

0.3

1.7

I
I I

1.0
1.8 1.9

0.8 1.5 1.5

3.4
4.1

4.2

1.1 2.3 2.5 2.5

0.2

1.6 2.1 2.2

1.1
1.7 1.7

0.2 0.1 1.7 0.3 6.9 O.l-.l.8 122.4 0.8 7 . 8 I 53.0 2 . 9 1 6 . 1 50.2 1 8 6 . 3 3.9 18.2 53.3 I 98.7 3.910.553.4 199.9

o+

o+
0.2 1.5 5.9 17.7 36.5

I I I
I

o+

I I I
I

I I I

Calm : 0 . 1

ot offshore b o d m y of Jeddah

Wind Speed Area : Position : Eeight : Duration :

-

Directional Distribution

-

Off shore

RAEIQE

464490E. 24843098 10 (m) 10 min. Average

I Wind Speed

I (m/s) I 22.0 I 20.0 I 18.0
I 16.0 I 14.0 I 12.0 I 10.0 I 8.0 I 6.0 I 4.0 I 2.0 I 1.1

1_____________1-------------------------------------------------------------------.

I Direction (Coming from, relative to north) I 0 30 60 90 120 150 180 210 240

270

300

330 I Omni

-

-

24.0 22.0 20.0 18.0 16.0 14.0 12.0 10.0

I I I
I

I
I I I I

I

o+
O+
O+ O+

o+

o+

I
I

I

8.0 I 6.0 I 4.0 I 2.0 I

0.1 0.7 2.3 3.7 1.6 0.1

o+
0.2 1.1

o+
0.1 0.3 0.4 0.1
O+ 0.4

o+
0.1
0.1 0.7 0.5
0.4

o+
0.2 0.9
1.0

O+

0.7

0.8 0.1

0.7 0.1

o+

0.8 0.1

0.4 0.1

0.1 0.4 1.1 0.4

0.1
0.4 1.5
0.4

0.3 1.7 4.2 1.1

o+

o+

0.1 0.4 1.3 2.2 6.2 6.9 13.1 13.3 9.6 11.8 2.0 2.0 0.2 o+

I
I

0.6 3.6 I 14.5 I 32.7 36.0

11.7
0.8

calm : 0.1

Wind Speed : Position : Eeight : Duration :
*ea

-

Directional Bxceedance

- Off Shore

RABIGE
464490B, 2484309N 10 (m) 10 min. Average

....................................................................................
I Wind Speed I Direction (coming from. relative to north) I I (m/S) I 0 30 60 90 120 150 180 210 240 270 300 330 I Omni I 1_____________1_________________________-------------------------------------------l I > 22.0 I I I

I

I > 18.0 I > 16.0

>

20.0

I

I I
I

> 14.0 I > 12.0 I > 1o.b I > 8.0
I

I > I > I > I >

6.0
4.0

I I I I

o+
0.1
0.0

O+ O+

o+
0.2 1.2 2.0 2.1

o+
0.1 0.4 0.7 0.9
O+

o+
0.4 1.1 1.2 0.1 0.8 1.3 1.3 0.1 0.5 1.2 2.0 2.0

o+
0.2 1.1 2.1 2.5 2.5 0.1 0.4 1.5 1.8 1.8

O+

3.1
6.0

I
I I

2.0 1.1

8.4 8.5

0.1 0.5 2.1 2.5 2.5

0.3 1.9 6.1 7.2 7.3

I I I o+ o+ I 0.2 0.4 I i : 4 2.6 I 7.6 9.5 20.7 22.8 30.334.6 32.3 37.4 32.5 37.4

I

I

o+
0.6 4.2 18.8 51.5 87.5 99.1

I I

I I
I

I I I

100.0 I

Calm : 0.1

0

Frequency of Occurence

and Exceedonce

Wind 8peed : Area Position : Eeight : Duration :

YANBU

-

Directional Distribution

- off shore

376318E. 2604611N

....................................................................................
Wind speed

10 (m) 10 d n . Average

____________,___________________________----------------------------------------(de)
270 300 330 I Omni I

I Direction (Coming from, relative to north) I 0 30 60 90 120 150 180 210 240
I

I

22.0 20.0 18.0 16.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0 1.1

-

- 24.0

22.0 I 20.0 I 10.0 I 16.0 I 14.0 I 12.0 I 10.0 I
8.0

I I I
I I

O+

0.1
0.8

0.1
0.8

I

6.0 I 4.0 I

2.0 I

0.1 0.6 1.3 2.5 1.3 0.1

O+

O+ O+

O+ O+

0.1
0.7 0.3

o+
0.3 0.8 0.5

O+ O+

0.1 0.4 0.7
O+

0.1 0+ 0.3 0.3 0.5 0.6
O+

0.1 0 . a
0.6 0.7

0.1 0.5 0.4
O+

0.2

0.9 0.5

0.2

0.1

o+

o+

0.1 3.6 3.2 0.8 9.0 8.6 2.5 12.3'14.5 3.2 8.5 10.4 0.7 2.0 2.2 o+ 0.1 O+

0.2 I 1.6 I 6.9 I 19.3 I

o+

31.9 29.1 10.4 0.6

\ I I I

Wind Speed Area : Position : Eeight : Duration :

YANBU

-

Directional Bxceedance

-

off Shore

....................................................................................
I Wind speed I (m/s)
I 330 I Omni I )_____________I_________________________------------------------------------------I I > 22.0 I I I I > 20.0 I I I I > 18.0 I I I I > 16.0 I o+ I o+ I I 0.1 0.1 I 0.2 I I > 14.0 I > 12.0 I 0.9 0.9 I 1.8 I I > 10.0 I 0.1 O+ O+ O+ 0 . 1 4.4 4.1 I 8.7 I 0.9 13.5 12.7 I 20.0 I I > 8.0 I 0.7 O+ 0.1 o+ 0.1 o+ o+ I > 6.0 I 1.9 0.1 0.2 o+ 0 1 0.5 0.3 0.1 0.2 3.4 25.7 27.2 I 59.9 I . I > 4.0 I 4.5 0.6 0.4 0.3 0.7 1.2 1.2 0.6 1.1 6.5 34.2 37.6 I 89.0 I I > 2.0 I 5.8 1.3 0 . 9 0 . 9 1.4 1.5 1.6 1.0 1.6 7.2 36.2 39.9 I 99.3 I I > 1.1 I 5.9 1.3 0 . 9 1.1 1.5 1.6 1.7 1.1 1.6 7.3 36.3 39.9 I 100.0 I

3763183, 2604611N 10 (m) 10 rain. Average

I

I Direction (Coming from, relative to north) 0 30 60 90 120 150 180 210 240

270

300

Calm : 0.0

Frequency of O c c u r m e am3 Exceedmce

o offshore boundary of Yanbu t

Wind speed - Directional Distribution 9yea : DuaA POsition : 758868B, 301528313

-

Off Shore

20.0 18.0 16.0 14.0 12.0 10.0
8.0

I - 22.0 I I - 20.0 I I - 18.0 II - 16.0 I o+ 0.2 0.2 I - 112.0 I 0.1 o+ 4.0 0.1 1.0 1.2 I o+ o+ 0.5 3.9 4.6 I 0.1 0.2 2.1 12.1 14.9 I - 10.0 I 0.4 o+ o+ 29.6 I 6.0 - 8.0 I 1.3 0.1 0.1 o+ 0.1 0.1 0.1 0.2 0.5 6.2 21.1 33.2 I 2.4 0 . 4 0.2 0.1 0.3 0.4 0.3 0.2 0.5 2.1 8.118.4 4.0 1.6 2.5 - 4.0 2.3 0.8 0.3 0.2 0.3 0.3 0.2 0.4 0.6 0.1 0.1 4.8 14.4 II 1.8 0.4 0.1 0.1 o+ o+ o+ 0.1 0.1 0.4 1.1 - 2.0 ________________________________________--------------O+ O+

6.0

2.0

0.4

Calm : 0.1

Wind speed : Area Position : Eeight : Duration :

....................................................................................
I Wind saed

Directional Bxceedlnce DUBA 758868B. 3015283N 10 (m) 10 min. Average 60
-90

-

-

off shore

I Direction (Coming from. relative to north)
I I O

I
270

I (mls)

.____

30

120

150

180

210

240

300

330 I Omni

> I >
I I I I I

22.0
20.0

I

>
> > >

18.0 16.0

I I
I

I

I > 14.0 I > 12.0 I I
I I

I I
I I

o+
0.1 0.5 1.8 4.1 6.4 6.9
O+
O+

I I
0.2 1.2 5.1 17.2

> > >
>

10.0 8.0 6.0 4.0 2.0 1.1

I I
I

I

0.1 0.2 0.6 1.4 1.5

O+

O+

0.1

o+
0.1 0.3 0.3

0.3 0.7 1.0

0.3 0.5 0.6

0.1 0.4 0.7 0.8

0.1
0.4

0.7 0.7

0.1 0.2 0.6 0.7

0.2 o+ o+ 0.7 0.1 0.2-'2.8 0.3 0.7 8.9 0.8 2.8 17.0 1.4 4.3 19.5 1.5 4.5 19.6

I I
I

38.4 56.7 61.5 61.9

I I I
I

0.2 1.4 6.0 20.9 50.5 83.7 98.1

I 99.9

Calm : 0.1

I

w
hie&

Danish Hydraulic
F l k

Institute
wed * p

RED SEA HINDCAST STUDY
Lkainr) m

Diectional Distributicn of Winds Frequency of Ckcllrence and Exceedaxe

I
SCde:

w . drk t

at offshore botmdary of Dub0

6.5

I

Wind speed Area : maition : Iieight : Duration :

....................................................................................
Wind Speed

Directional DistribYtion REDSEA Reference m i n t A (41.538 15.93N) 10 (m) 10 min. A m r a g e

-

(de)
22.0 20.0 18.0 16.0 14.0 12.0 10.0 8.0 6.0

Direction (Coming from, relative to north) 0 30 60 90 120 150 180 210 240

0.2 1.0 4 . 0 - 6.0 1.3 2.0 - 4 . 0 0.3 1.1 - 2.0 ______-------Calm : 0.1 Wind speed -ea : Position : Ksight : Duration :

-

________________________________________--------------24.0 22.0 20.0 18.0 16.0 14.0 12.0 10.0 8.0

270

300

I 330 I Omni I I I I

I

I

I
0.1
0.4

I

o+
0.1

o+

1.8
4.3

o+
0.2
0.6

o+

o+
0.2
0.4

0.1

0.1 0.3 0.5 0.1

0.1 0.6 1.0

0.3

0.2 0.7 1.5 2.5 2.4 0.1

o+
0.1 0.5 1.4 1.3 0.1
O+

o+
0.2 1.5 1.3 0.1 0.1 0.6 2.1 1.6 0.1

7.7 9.7 7.4 2.0 0.1

1.4 2.5 1.2 0.1

0.1 0.3 1.2 5.0 7.1 3.1 0.2

0.1 I o+ 0.4 I 0.1 I 2.0 I 0.5 5.5 12.7 2.1 25.1 5.9 33.5 6.6 19.0 3.1 1.7 0.2

I I

----_

-

Directional Exfeedance

RBDSEA
Reference m i n t A (41.53E 15.93N) 10 (m) 10 ndn. A v e r a m

0.1
0.4

o+
I > 10.0
I I I

o+
o+
0.2 0.0 2.2 4.8 5.9 6.0
O+

o+
0.1

>
>

I >

8.0 6.0
d.0

I I
I

> I >

I

2.0 1.1

I
I

0.1 0.3 1.3 2.6 2.9

O+

0.1
0.4

0.3 0.7 0.8

0.9 1.0

0.1 0.7 1.7 2.0

0.2 0.9 2.3 4.8 7.3 7.3

2.2 6.5 14.2 23.9 31.3 33.3 33.4

o+
O+

0.4

0.1 0.6 2.0 3.3 3.3

0.3 1.8

3.0
3.1

0.1 1.6 0.7 6.6 2.9 i3.7 4 . 4 16.0 4.5 17.0

0.1 0.5 0.1 I 2.5 8.0 I 0.7 20.7 I 2.8 45.8 I 8.8 79.3 I 15.4 98.3 I 18.5 100.0 I 18.7

I

o+

I

Calm : 0.1

UUlC

SAUDI ARAMCO
RED SEA HINDCAST STUDY
Orwing m

FT*v

Danlsh Hydraulic
File:

lnstltuto
1996

is w x z

D a t e Wed Apr
bit:

Directional Distribution of Winds Frequency of Occurace and Exceedance ot reference point A

SEdC

drk

6.6

wind Speed : Area Position : Height : Duration :

Directional Distribution REDSEA Reference Point B (39.88B 18.lSN) 10 (m) 10 mln. A v e r a g e

-

- 24.0 I I - 22.0 I I - 20.0 I I - 18.0 O+ o+ I o+ I o+ 0.1 I 14.0 - 16.0 o+ o+ 0.1 I 0.3 I O+ I 12.0 - 14.0 0.5 O+ O+ O+ 0.4 0.6 1 1.5 I I 10.0 - 12.0 0.1 O+ 0.2 1.4 o+ 1.0 2.9 I 5.7 I 0.6 3.1 1 8 . 0 - 10.0 0.4 0.1 o+ O+ 1 O+ 0.2 2.1 6.4 I 13.0 I I 6.0 - 8 . 0 1.3 0.3 0.3 0.4 1.8 5.3 0.3 0.1 O+ 0.4 3.7 12.0 I 25.9 I 2.5 0.6 0 . 8 1.0 3.2 5.8 1.0 0.3 0.2 0.7 4.3 12.0 I 32.5 I I 4.0 - 6.0 2.3 2.4 0.9 0.6 0.5 1.2 2.5 4.3 I 19.5 I I 2.0 - 4.0 1.8 0 . 8 0.9 1.1 I 1.1 - 2.0 I 0.1 0.1 0.1 0.2 0.1 0.2 0.1 o+ 0.1 0.1 0.2 0.2 I 1.5 I ....................................................................................
22.0

i

I 20.0 I 18.0 I 16.0

Calm : 0.0 Wind speed Area : Position : : Eeight Duration : Directional Exceedance REDSEA Reference Point B (39.883 18.15N) 10 (m) 10 mLn. A v e r a g e

-

I Wind Speed I (mls)

I Direction (coming from. relative to north)
I I
O+

0

30

60

90

120

150

180

210

240

270

300

330 I Omni I

I

> > I > 1 > I > I >
I
I

I > I >
I

22.0 20.0 18.0 16.0 14.0 12.0 10.0
8.0

O+

O+

o+
0.1

o+ o+ o+
0.4 1.4 2.5 2.7

> 6.0

I I I

> > >

4.0 2.0 1.1

0 . 5 0 . 2 0 + 1.8 0.4 0.3 4.3 1.1 1.1

I

6.0 6.2

1.9 2.0

2.1 2.2

0.1 0.1 0.6 0.3 2.0 0.95.10.1 2.6 10.4 0 . 4 5 . 8 16.3 1.4 8.2 18.7 2.4 8.3 18.9 2.5

o+

.

o+
0.1 0.3 0.9 0.9

0.1 0.5 1.1 1.1

0.1 0.4 0.6 0.1 1.4 3.5 0.33.59.9 0.7-.7.2 21.9 1.4 11.5 33.9 2.6 14.0 38.2 2.7 14.2 38.3

o+ o+

o+

0.3 1.8 7.5 20.5 46.4 78.9 I 98.5 I 100.0 I

Frequency of Ocnrence md Exceedonce

Wind spesd : Area Position : : Eeight Duration :

-

Directional Distribution

RED9F.A

Referen- m i n t C (38.33E 20.43N) 10 (m) 10 min. Average Direction (Coming from, relative to north) 0 30 60 90 120 150 180 210 240 330 I (rmni

I Wind Speed I (m/e) I
I 22.0 I 20.0 I 10.0 I 16.0 I 14.0 I 12.0 I 10.0 I 8.0 I 6.0 I 4.0 I 2.0 I 1.1

__________ __ .
-

270

300

- 22.0 - 20.0 - 18.0
16.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0
O+

24.0

I I

o+
O+

I
I

o+

__ __ --

0.1 0.6 1.5 2.6 2.8 1.9 0.1

o+
o+
0.4

0.1 0.4 0.1 0.6 0.7 0.1 0.1 0.5 0.5 0.1 0.1 0.6 1.4 1.0 0.1

o+
0.1 0.4 1.1 0.7
O+

o+
O+

o+
0.1 0.4

0.8 2.4

0.9
1.0 0.1

3.2
1.4 0.2

o+

0.2 0.4 0.3 0.1

0.2 0.4 0.4 0.1

0.8 0.7 0.1

0.3 o+ 1.6 0.1 4.6 0.8 11.3 2.818.8 4.5 15.9 1.9 4.1 0.1 0.1
O+

0.4 1.8 I 5.8 I 14.8 128.9 I 32.6 I 14.6 I 1.2

I
I

Calm : 0.1 wind Speed Area : Position : Eeight : Duration :

-

Directional Bxcoedlllae

REDSEA

Reference m i n t C (38.33E 20.43N) 10 (m) 10 min. Average north) 210 240

I Direction (Coming from, relative to I (m/s) I 0 30 60 90 120 150 180 1_____________1_________________________-------------------------------------------1 I > 22.0 I I > 20.0 I I > 18.0 I I > 16.0 I I > 14.0 I o+ O+ I > 12.0 I 0.1 0.1 I > 10.0 I 0.7 O+ 0.4 I > 8.0 I 2.2 0.1 0.1 1.2 0.1 I > 6.0 I 4.8 0.4 0.1 0.1 0.7 3.6 0.5
I Wind I I

I
270
300 330 I Omni I

I I

o+ o+
O+

I I

o+
0.1 0.2 0.7 1.0 1.1

o+

o+ o+
0.2

>

I ....................................................................................
Calm : 0.1

I >

>

4.0 2.0 1.1

I I

I

7.6 9.5 9.5

1.4 2.4 2.5

0.7 1.4 1.5

0.7 1.2 1.3

2.1 3.1 3.3

6.8 8.1 8.3

1.6 2.3 2.4

0.2 0.6 1.0 1.0

0.1 0:9 0.5 3.7 1.3 8.3 2.0 10.1 2.1 10.2

0.3 1.9 6.5 17.9 36.6 52.5 56.6 56.8

I
I I I

I I
I

I 0.4 I 2.2 I 7.9 I 22.7 I 51.6 I 84.2 I 98.8 I 100.0 I

o+

I I I

ciat:
project:

SAUDI ARAMCO
RED SEA HINDCAST STUDY
Dimity m.
I (

Danish
Fk
SCde:

Hydraulic

Institute

o a k Wed Apr
drk

,

1996

Directional Distribution of Winds Frequency of Occtrence a d Exceedonce at reference Doint C

6.8

I

Wind Speed

- Directional Distribution

Area : REDSEA Position : Referenfe m i n t D (37.3% 23.03N) Eeight : 10 (m)
Duration : 10 min. Average

....................................................................................
I wind
I
(m/S)

Speed

I Direction (Coming f r o m , relative t o north) I 0 30 60 90 120 150 180 210 240 I
I I

270

300

330 I Omni

I 22.0 I 20.0 I 18.0

- 22.0 - 20.0 I 16.0 - 16.0 I 14.0 - 16.0 I 12.0 - 14.0
I 10.0 I 8.0

-

24.0 I

o+
o+

I I
I I

o+
o+
0.7 3.1 9.5 20.2 30.3 25.5 10.2 0.6

I
I

I
I

6.0 4.0 2.0 1.1

-

I I
I I I

o+

o+
0.2 0.6 1.0 1.4 0.8

12.0 10.0 8.0 6.0 4.0 2.0

o+
0.1 0.1 0.3 0.5 0.1
O+

o+
O+

o+
0.1 0.2 0.8 1.3

o+
o+
0.1 0.3 0.7

o+
O+ O+

o+
0.1 0.1 0.5 0.4 1.4 0.3 0.7

I I I

o+

0.2 0.2 0.3 0.1

0.1 0.3 0.7

o+

0.1 0.6 0.7 0.1

0.8

0.5

o+

o+

0.1 0.4 0.4 0.1

o+

0.3 1.0 2.2 5.2 5.8 1.9 0.1

0.7 2.7 8.1 16.7 21.8 12.8 2.7

I I
I I I

I
I

I
I
I

o+

I I
I

Calm : 0 . 0 Wind speed Area : Position : Ecight : Duration : Directional Exceedance REDSEA Reference m i n t D (37.338 23.0311) 10 (m) 10 min. Average

-

....................................................................................
I Wind Speed I Direction (Coming from. relative t o north) I (mls) I 0 30 60 90 120 150 180 210 240 1_____________1_________________________-------------------------------------------1 I > 22.0 I I > 20.0 I
I I
270 300
330 I Omni

I

I

I
o+ o+
I

> I >

I
I

> I > I > 8.a I > 6.0
I
I

>

18.0 16.0 14.0 12.0 10.0

I

I I I I
I

o+ o+
O+ 0.1

I

I I
I

o+
0.3 0.9 1.9 3.3 4.2 4.2

o+
0.1 0.3 0.6

o+
0.2 0.4 0.7 0.1 0.3 1.0 1.0

o+
O+

o+
0.1 0.3 1.1 2.4 3.2 3.2

o+ o+
0.1 0.4 1.1 1.6 1.6

o+
0.1 0.1 0.5 0.9 0.9
O+

I >

> >

4.0 2.0

I I
I

1.1

I

1.0 1.1

0.8

0.2 0.7 1.5 1.6

0.1

0.5 0.8
0.8

0.6 2.1 2.7 2.8

0.7 3.4 1.3 11.5 3-528.2 8.750.1 14.562.8 16.4 65.5 16.4 65.5
O+ 0.3

I I 0.7 I

I 3.6 I I 13.3 I
133.4 I 163.7 I 189.1 I I 99.4 I I 100.0 I

Calm : 0 . 0

0

Frequency of Occurence ond Exceedonce at reference mint D

wind Speed Area : Position : Eeight : Duration :

-

Directional Distribution

REDSEA

Reference mint E (35.48E 26.00N) 1 0 (m) 10 min. A v e r a g e

I Wind Speed

l_____________I___________
I 22.0 I 20.0

I (m/s)

I D i r e a t i o n (Coming fmm, relative t o n o r t h ) 60 90 1 2 0 1 5 0 1 0 0 210 240 I 0 30

270

300

I 18.0
I 16.0
I 14.0

I 12.0 I 10.0
I
I
8.0 6.0

-

I

I
I

4.0 2.0 1.1 -

- 12.0 - 10.0 -

I I I I 16.0 I 14.0 I
24.0 22.0 20.0 18.0

0.1

O+

I
I I

8.0 6.0

I

4.0 I 2.0 I

0.1 0.5 1.0 2.0 2.1 0.2

o+

o+
O+

o+
O+

o+
0.1 0.4 1.7
0.8

o+
0.2 0.4 0.5 0.2

0.1 0.1 0.5 0.2

O+

0.1

0.3
0.2

0.1 0.1 0.7

0.3 0.1

0.3

0.1 0.4 0.1

0.1 0.1 0.3 0.2

0.1 0.4 0.3

0.3 0.9 2.8 6.4 6.9 1.7

o+

o+

o+

o+

o+

o+

330 I Omni I I I I I I I I 0.1 I 0.1 I 0.8 I 0.9 I 2.6 I 2.9 I 8.4 I 9 . 5 I 16.8 I 20.5 I 21.1 1 2 9 . 6 I 1 2 . 8 I 26.0 I 2.6 I 9.4 I 0.2 I 0 . 9 I

I

calm

: 0.0

wind Speed Area : Position : Eeight : Duration :

-

D i r e c t i o n a l Exc-ce

REDSEA
R e f e r e n c e m i n t E (35.48E 26.00N) 10 (m)

1 0 min. A v e r a g e

I Wind

Speed

I (m/s) I > 22.0 I > 20.0

I D i r e c t i o n (Ccming from, relati- t o n o r t h ) 60 90 1 2 0 1 5 0 1 0 0 210 240 I 0 30
I I

270

300

330 I Omni I

I

I > 10.0
I > 16.0 I > 14.0 I > 12.0 I > 10.0

I I
0.1 0.4 1.3 4.1 10.5 17.4 19.1 19.1 0.1 0.9 3.5 11.9 28.7 49.8 62.6 65.2 65.4

I I

I
I I

I
I I

I
0.1 1.0 3.9 13.4 33.9 63.6 09.6 99.1 100.0

I

I I
I

I >
I

I I
I

o+
0.1
0.7

I

o+

I > 4.0 I > 2.0 I > 1.1

>

8.0 6.0

o+
0.3
0.7 1.2
1.3

o+
0.1 0.2 0.7 0.9

o+
0.1
~

o+
~

o+
0.1 0.5 2.1 2.9 2.9

0.1

0.1 -

I I I I

1.7 3.7 5.8
6.0

o+
0.1 0.4 0.5

I
I

0.3 0.4
0.5

0.3 0.9 1.2 1.2

0.1 0.6

0.7 0.7

0.2 0.5 0.7 0.7

0.2
0.6

0.9 0.9

I I
I

I I I I I
I

Mmt:
Project:

SAUDI ARAMCO
RED SEA HINDCAST STUDY

s
W

1L

Danlsh Hydraulic

Institute

x

Directional Distribution of Winds Frequency of Occwence and Exceedmce at reference point E

Wind Speed : Area Position : Eeight ; Duration :

-

Directional Distribution

REDSEA

Reference Point P (34.66B 28.668) 10 (m) 10 mln. Average

I 14.0 I 12.0
I 10.0 I 8.0 I 6.0 I 4.0 I 2.0 I 1.1

- _-__ __ __ ___-_
Calm : 0.8 Wind Speed
M a

-

16.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0

0.1 0.5 1.7 4.6 5.7 4.8 3.1

o+
o+ o+
0.2 0.2 1.2 1.0 0.1 0.4 0.3
O+

o+ o+
O+
O+

o+
0.7

o+
0.1 0.2

o+
0.1 0.1 0.1 0.1 0.3

O+

o+

o+

0.1 0.3 0.2

0.1 0.4 0.7

0.1 0.2 0.6 3.1 0.4

o+ 0.2 0.1 1.4 0.7 5.9 1.7 13.7 3.910.5 7.9 11.0 0.8 1.3

0.3 2.0 8.4 20.5 29.5 30.3 8.1

_____

-

Directional B x c e e d ; m c e

: REDSBR

Position : Reference Point F (34.66B 28.6613) Eeight : 10 (m) Duration : 10 mln. Average

I

_--_-I
I

omni I
I I

I > 14.0 I > 12.0 I > 10.0
I I I

I

> I > I >

> >

I I
I I

9.0

6.0 4.0 2.0 1.1

I I
I

._

0.1 0.5 2.2 6.8 12.4 17.3 20.4

o+ o+
0.1 0.3 0.5 1.7 2.7
O+ O+

o+
01 .
0.6
O+

o+
0.2 0.2 0.2 0.1 0.2 0.5

0.1

0.8

0.7 0.7

0.3 0.4

0.1 0.1 0.4 0.7

O+

0.2 0.6 1.4

0.2 O+ 0.1 1.6 0.0 7.4 0.3 2.5 21.1 0.9 6.539.6 4.0 14.4 50.6 4.315.2 51.9

o+

o+

o+

I

o+

I 0.3 I 2.3 I 10.7
I 31.2
160.8

I 91.1 199.2

Calm : 0.8

0

Frequency of Occurence md Exceedme

at reference point F

-.- ,
0000 02/04 1990
0000

02/06

0000 02/08

0000 02/10

00:00

02/12

00:00 0 2 / 14

0000 02/16

0000 02/18

Yanbu d Bohr WL. regional model

0.2

0

O.l 0.0

-0.1

-0.2
-0.3
0000 02/04 1990
0000 02/06

0000 02/08

0000 02/10

0000 02/12

0000 02/14

0000 02/ 16

0000 02/18

Madinat Yanbu WL. regiond model

0.2
0.1

0

0.0

-0.1

-0.2
I I 1 1

1

1

1

0000

0000

0000

02/04 1990

02/06

02/08

0000 02/10

0000 02/12

0000 02/14

0000 02/ 16

0000 02/ 18

0

Fri Oec 8 1995 Comprnison of pred.

and cakd WLs

ot Ymbu d Bahr md Madinat Yonbu

___

Rabigh WL, regiord mcdd

aj5ie

0

?‘?‘tg i2

0’5 0.25

0.0

-0.25

.

. _

-0.5 0000 02/04 1990

0000 02/06

0000 02/08

0000 02/10

0000 02/12

0000 02/14

0000 02/16

0000 02/18

_______ WL. regional model Jeddoh
0.5 0.25

0.0
-0.25

-0.5
0000 02/04

0000 02/06

0000 02/08

0000
02/10

0000 02/12

0000 02/14

0000 02/ 16

00:00 02/ 18

1990

________
0.5

Jizon WL, regional model

0
E

0.25
0.0

-0.25 -0.5 0000 02/04 1990

0000 02/06

0000 02/08

00:00 02/10

00:00 02/ 12

0000 02/14

0000 02/16

0000 02/18

I
SCdc:

Cwnporison of pred. m d calcd. WLs
hit:

4

at Rabigh, Jeddah. and Jizm tidd st.

48

53

58

63

68

73

78

0.5 0.1

03
0.2

0.1 0.0
-0.1

0000
01/25 1077

OD00 01/26

00W 01/27

0b00 01/28

0000
01/20

0000 01/30

I
I

umt:
Froject:

SAUDI ARAMCO
W

Red Sea Hindcast Study
Draing m.

3

F i k
SCdC:

Date

T Dec 12 895 w

7701 Storm:
R&i$

axrent field (example)

1:57M)OO

cit:

7 ..3

mJ

od cdcldated water level series

(Grid spacing 1500 m)

0.4

03 .
0.2
0.1 0.0 -0.1

-0.2

-0.3
-0.4

0000
03/01

0000

0000

WOO
03/04

0000

03/02

03/03

03/05

00.00 03/06

0000 03/07

0000

03/08

1877

CCat:

SAUDI ARAMCO
Red Sea Hindcast Study
b a i n g m.

ij
W

R+Ct

Danlsh Hydraullc Institute
Fk i
Ode:

ii

Tue Dec 12 '895

7703
Robigh arrent field ( e x q l e ) and calculated water level series

SEdc:

1570000

Ht:

mjl

7.4

140-

135-

130-

125-

120.

0

E 115.

5: .- 1 1 0 e 0
0

v

E:

5 105
v

100

95

90

85

--++
80
i

0.2 rn/s

54 ' . 58

63

68

73

78

82

0.4

03

02
0.1

0.0

w.00
12/09 1978

00.00
12/10

OOW
12/11

WOO
12/12

00.00
12/13

aat:
Rq'st:

SAUDI ARAMCO
Red Sea Hindcast Study
( mn m. h i g

s

Danlrh Hydraulic Institute

1
7.5

0

F l k

Ode
bit:

Sue Dec V 895

7812 Stcfrn:

Robigh w r e n t field (example)

.

1:570000

rnjl

and calcdated water level series

(Grid spacing 1500 rn)

--3 0.2 rn/s

0.16

0.2

0.12

0.08
0.04

00 .
-0.04

-0.08
-0.12 -0.16

DOM
05/15
1979

moo
05/ 16

W9D

WM
05/18

DM O
05/19

05/ 17

I
111-

I
,110,1111-

SAUDI ARAMCO
7905 Stwm:
Drmirg m.

.N

note

Tue Dee 12 1995

%de:

1:57M)Oo

kit:

R&i$

rnll

arrent fie@ (example) md cdnloted water level series

7.6

(Grid sDacina 1500 rn)

140-

135-

130-

125-

120.

h

E 115.
0 0

Y

._ 110, F
U .5 105

8

v

100

95

90

85

8R

0.2 0.1 0.0 -0.1 -02

-03 -0.1
0000 03/01

0000
03/02

mw
03/03

WW
03/M

oooo
03/05

tee0

(Grid spocing 1500 m)

140-

135.

130.
.

. -

125.

120.

/.
0 0

E 115.

'D
0 0

.FllO.
a
(0

0

v

& 105

100

95

90

85

84
0.:
0.2 0.1

0.2 m / s
ODW

0.C
-0.1

-0.2

w:W M 1 / 1
1880

04/ 12

0O:W 04/ 1 3

oow
M/l4

OLIO0

04/15

W:W 01/16

amt.
Project:

SAUDI ARAMCO
Red Sea Hindcast Study
b a i n g m.

I;
W

Danlsh Hydraulic Institute
File:

Dot=

Tue Dsc 12 1995

B W 4 Storm:
Rabi@ w r e n t field (example)

%de:

1:570000

tit.

m cdcldoted woter level series d

7.0

(Grid spacing 1500 m)

140-

.

. -

120.

100,

95

90

85

6.08 0.04
0.0 -0.04 -0.08 -0.12 -0.16 WOO
06/04 1980

0000
06/05

0000
06/06

OMJO
06/07

W:O0 06/06

0000 06/09

aal:
P,OjCCt:

SAUDI ARAMCO
Red Sea Hindcast Study
Draring m.

F I

Danish Hydraulic

Inatltute
12 1995

!
7.9

0

Fk i

DatC Tua

8006 Storm:
Rabi@ arrent field (exanple)

l570000

Init: mjl

md cakrdated water level series

om 1
90/01

ww

SO/Ol O W

m/oi
0300

CO/Ol

zo/o1
W O 91'021'0800-

cw o

tow
00

too
800

.-

21p

s/w 2'0

be
SQ

c -

06

S6

001

SOL

y n

01 L

1 2'
!s
0
v

SI1 3

.oz 1
.SZL

.or1
-SEL

-0PL

(Grid spacing 1500 rn)

0.3
0.2 0.1 0.0 -0.1 -0.2 -0.3 4.4

woo
02/09 1981

0000
02/10

0000
02/11

0000
02/12

0000
02/ 13

Cbmt

SAUDI ARAMCO
Red Sea Hindcast Study

G

P+Ct

e f

Tue Dec 12 895
SCde:

k570000

48

53

58

63

68

73

7a

0.5
0.4
0.3

0.2 0.1

0.0

ow0
12/20 1981

o(tw
12/21

om o
12/22

0a00
12/23

Wm
12/24

0

0

12/25

(Tat
PrOjCC':

SAUDI ARAMCO
Red Sea Hindcast Study
Oraing m.

ij

Danish Hydraulic Institute
File:

a
W

ode

Tus Lkc 12 1995 Rabish current fieM (exorple)
mjl

1:570000

and cdcldoted water level series

7.12

(Grid swcinq 1500 m)

140

135

13C

125

12c

E 11:
0 0
IC
7

.Fl1(
8

: :
2
v

&

lo!

lo(

9 :

9 (

8 :

81

0.4

03 .
0.2 0.1

00 .
-0.1 -0.2

-0.3
-0.4

0000
03/24 1982

m00 03/25

0w0

0000
03/27

03/26

0000 03/28

0000
03/29

0000
03/30

SAUDI ARAMCO

Rabigh current field (example)

and calcdated water level series

(Grid spacing 1500 m) 140.

135.

130.

125.

120

0 0

E 115

‘0
._ FllO
Q

i
105 100

v

&

95

90

85

80

0.2 m/s
0000
02/24
0000 02/25
0000 02/26

0.5
0.4

03 .
0.2 0.1

00 .
-0.1 -0.2

-. 03
0000
02/23 1983

0000 02/21

Cht:

SAUDI ARAMCO
Red Sea Hindcast Study
Dro*ing m.

&
W

PTOjcEt:

Danish
Fik

Hydraulic
DatC

institute
Tue Dec 12 1995

x x
7.14

8302 Storm:
Rabigh current field (exomple)
and cdcldated water level series

1:570000

Init:

mjl

(Grid mocina 1500 rn)

.

..

0.5

0.4

0.3
0.2
0.1

00 .
OOW 02/20 1984

ODW 02/21

WW
02/22

ow0
02/23

wao
02/24

ow0
02/25

oom
02/26

a

Rabicfi arrent field ( e x q l e )

d cdcldated water level series

140-

135.

130-

125-

120.
h

E 115.
0 0

Y
Ln

._ FllO
8 a
U ., +

105

v

100

95

90

85

80

0.2 m/s
OGOD 03/25 0000 03/26 De00 03/27 W:00 03/28 0000 03/29

0.1 0.3 0.2 0.1 0.0 -0.1 0000 03/21 1984

mal:
ROjeCt

SAUDI ARAMCO
Red Sea Hindcast Study
LWaing m.

s

Tue Dec 12 1995
scde:

8403 Storm:
Rabigh arrent field (exp'IpIB) and cdcdoted water level series

1:570000

7.16

(Grid spacing 1500 rn)

aa .a %
U"

03

0.2 0.1

0.0

moo
11/30 1985

1

I

I

I

I

M)W
12/01

o m
12/02

WW
12/03

moo
12/04

I
0000
12/05

amt:

SAUDI ARAMCO
Red Sea Hindcast Study
Lhwirr) m

K 4
W

frqcst:

Danish Hydraulic Institute
Fk i
Ode:

x I
7.17

Tue Dec

12 B95

85n
Rabigh wrent field (exanple)

%de:

1570000

tit:

mjl

and cdcdated water level series

140

135

130

125

120
h

E 115
0 0

‘D
._ l O Fl
U ._
v

i
100

b 105

95

90

as
ao
48 53

sa
I

//
0.2 rn/s 63 68

73

78

a2
-.

0.4

,

Sherm Robigh cdc WL
I
N

I

0.3
0.2 0.1

WOO 12/16 1S f f i

o[too
12/17

oom
12/18

WO O 12/19

Wm
12/20

Rabigh arrent field (exp-nple)

and cdcldoted water level series

(Grid spocing 1500 m)

140-

135, d
I/

d
13d:

6
Y

,

125-

120.

0 0

E 115.

2 .-f l l O . Z
? m i n 6 105.
100.

v

95.

90.

85.
----3

8e:

0.2 m/s

0.12 0.08

0.04
0.0

-O.M
4.08 -0.12 4.16

WI CD
06/ 12 1986

DaW
06/13

DDW
06/14

O w 0 06/15

0000 06/ 16

aat:
ROjat:

SAUDI ARAMCO
W

Danlsh Hydraullc
fik
SSdc:

Insiltuta
Wed Des

Red Sea Hindcost Study
orawirq I . n

x I
7.19

DOtC

l T395 3

8606 Rabi@ cwrent field (example)

1:570000

Nt: mjl

ad cdcubted water level series

140.

135.

130

125

120

E
0
0

1 5

Y .e
rn
v

IO

!
105
100

95

90

85

0 C

0.2 rn/s 48 53

58

63

68

73

78

82

0.4

0.3

0.2
0.1

0.0

oom
02/06 1987

I

I mw
02/07

I
ow0
02/08

I
ww
02/09

I
Ow0
02/ 10

ow0
02/11

amt:
project:

SAUDI ARAMCO
Red Sea Hindcast Study
&wing m.

ij w

Danlrh Hydraulic Institute
File:

5

Dote

Wed Dtc 1 1995 3

8702
Rabigh arrent field (example)
and calculated water level series

SCdc:

1:570000

kt i:

mjl

7.20

(Grid w i n q 1500 m)

140.

135.

130.

125,

120

0

E 115

() I

5:

._ FllO

8 a
2
v

b 105

100

95

90

85

80

0.2 m/s
mo o
04/ 17

0.5 0.4 0.3 0.2 0.1 0.0 -0.1 0090 04/16 1987

WOO
04/10

0000
04/19

ow0
04/20

0000 06/21

Rabigh current fiekl (example)

old cokdated water levd series

(Grid spacing 1500 rn)

140-

135.

130.

125.

120
h

E 115
0
0

Y

._ PllO
% U ._
v

8

& 105
100

95

90

85

0.2 m/s

0.4

0.3 0.2
0.1 0.0 -0.2

-. .

oow
01/11 1988

I

I ww
01/12

I
00 01/13

I
OOW
01/14

I
0 : Ow
01/15

WW
01/16

(Talk

SAUDI ARAMCO
Red Sea Hindcast Study
a"iq m.

i3

FYqst:

Danlsh
Fk i
SCdS

Hydraulic
oats:

Institute
Wed Dec

~!2 x
1.22

U 1995

" 1 Storm: 0
Rob;@ u r m t field ( e x q l e )

1:570000

rnll

a !cdkdded water x

level series

(Grid spacing 1500 rn)

140-

15 3:

,
#

#

139

125.

120.
h

E 115.
0 0

'0

.FllO.

: :

: :
v

71 .& 105

100

95

90

85

84

0.2 rn/s
Sherrn R&i@ cdc WL
I

. 0.4

".0.2
0.1
0.0

-0.1

oom
05/07 1000

OCzOD 05/00

Ow0 05/W

WOO
DS/ID

OD00
05/11

DOC0 05/12

I
%dC:

t570000

I

Rabish arrent field ( e x q l e )
hit: mjl

m d cdcldated water level series

(&id spacing 1500 rn)

48

53

58

63

68

73

78

05 .
0.4

03 . 02 .
01 .

00 .
WDO 11/29 1989
O[tW

11/30

ODW 12/01

om0
12/02

0M)O
12/03

woo
12/M

WilO 12/05

amt
FTOjrsk

SAUDI ARAMCO
Red Sea Hindcast Study
Druaiq m.

&
W

Danish Hydraulic
F i k
OOtC

Institute
Wed Lhc 0 1995

i$
7.24

scdc: 1:570000

H: t

storm R d i g h crrrent field (exunple)
and calclkted water level series

mll

(Grid smcino 1500 r d

140

135
#

Y
Y 11

130

125

120

, . E 51 0

115

.F?

x: E

11c

v

0 .& 105

1oc

95

9 C

8:

8 s

0.2 m/s
MOO
01/21

0.2

0.1 0.0 -0.1 -0.2 -0.3 -0.4

moo
01/23 1991

WOO
01/25

ODW
01/26

WOO
01/27

w:W
01/28

I
Danish Hydraulic

umt:
W

Institute

Red Sea Hindcast Study
Rabigh a x r a t field (exaTple)

x 1

ssdc:

1:570000

I

I

hit:

3

md ccdclhted water level series

1

1.25

I

(Grid s m i n a 1500 rn)
140.

135,

130.

125

120

h

E 115
0 0

Y

._ FllO
V ,.+

k

105

v

1oc

95

9c

85

EL
0.4

e12

0.3
0.2 0.1 0.0 -0.1

-0.2
OODO 03/21 1991

woo
03/22

OODO
03/23

o(tw
03/24

oom
03/25

wm
03/26

0
and cdcubted water level series

(Grid $pacing 1500 rn)

e

140-

135-

130-

. 125-

120.

._ PllO

,+ 105
v

U ._

Ii

100

95

90

85

e

80

---+m/s 02 .
- Sherrn R&i@ ,
colc W l
I

0.4

;

:
0.1

00 .

moo

11/28 1991

o(tW 11/29

mw
11/30

OOW 12/01

cimt
Rojed:

SAUDI ARAMCO
Red Sea Hindcast Study

c

N

w
&awing rn

Danish Hydraulic Institute
F k

5

Th, Dm

n95

I
SC&

l570000

1 ~ '4

gin storm R&i@ curent field ( e x q k ) and cdcldded water level series

1

1.21

I

(&id spacing 1500 m)
140.

I

135: ui
J;

rr

13g
d
#

rc

125

120

h

E 115
0
0

Y
w

.- l l O F

! i

s? 105
1oo

2

95

go

85

s
0.

f 0.2 m/s

0. 0. -0. -0. -0. -0. -0.

0o.m
01/01 1992

moo
01/02

0m0
01/03

mo o
01/04

MK)O
01/05

0w0
01/06

amt:

SAUDI ARAMCO

w

(Grid spacing 1500 m)

130

120

0 0

E 115

2

0.5
0.4 0.2 0.1 0.0 -0.1 -0.2 -0.3 -0.4

0.3

0000
02/22 1892

oeoo
02/23

OOW
02/24

WW
02/25

oom
02/26

00W
02/27

0000 02/28

( a: T t

SAUDI ARAMCO
Red Sea Hindcast Study
&awing m.

Fi
w

F l k

ROj&

Danish Hydraulic Instliute

x x
7.29

,,.,, ,
mll

w5

9202 Stwrn:

SCde:

1570000

tit:

Robigb arrent field (exmnple)

and calculated w o k levd series

(Grid spocirq 1500 rn)

03 . 02
0.1

00 .
-0.1

-02 -0.3 0000
02/01 1993

0000
02/02

0000
02/03

0000
02/04

0000
02/05

0000 02/06

Clat.
Project:

SAUDI ARAMCO
Red Sea Hindcast Study
arming m.

F 4

Danish
Fii

Hydraulic
D*

Institute
Thu Dec 4 1995 '

# x
7.30

9302
R&i$

SCdc:

1570000

tit:

arrent field (example)

rnl

and cdcldated water level series

140

135

130

125

120
h

E 115
0 0

2
._ FllO
V ._ 105 & ,

i

./

100

95

9c

85

80
I

0 2 m/s . 53 58 63 68 73 78 82
-.

0.5 0.4 0.3 0.2 0.1 0OW 11/20 1993 W O O 11/21

000 11/22

MIW 11/23

W90 11/24

000 11/25

clent:
Rq'cd:

SAUDI ARAMCO
Red Sea Hindcast Study
&.=win) rn

K
Y

Danish Hydraulle
Fk i
SCdC:
Date

lnstltute
Thu D e c W 1995

1z I

1:570000

mjl

93" R&i@ crrrmt field (exonple) m cdcrdated water level series d

7.31

140-

135.

130.

125

120

E 115
0 0

Y
0
0

._ l l O F

: :
v

&

U

105

1 oc

95

9c

85

8 C
48

0.2 m/s

53

58

63

68

73

78

0.6 0.4

0.2

"" ,
0000
01/12 1094

0000 01/13

00.00 01/14

0000 01/15

ODD0 01/16

00.00 01/17

mal:
WqKt

SAUDI ARAMCO
Red Sea Hindcast Study
Orming m.

Danish
fik
SC*

Hydraulic

Institute
1995

6 I J x x

Dote: ThU

9401 Storm:
Robia current field (example) m d calculated water Ievd series

l.570000

h'k

rnjl

I

7.32

I

fGridswcino 1500 m)

.

Fik

O b a a r r e n t field (example)

C66L

so/zo
0000

so/zo
0o:oo

tO/ZO

0o:oo

CO/ZO WGO

zo/zo
0o:oo

lO/ZO

OWU

LO1' 0

00 L'O

M Yo3 1400 lo n W A

-EO .

L O

s / u Z'O
c

SL L
DE L

SQ L
06 L 56 L

ooz

soz
012

9
I) I

SLZ 5.
0

7 J

ul

ozz
szz
OfZ

g
0

Sfz

o*z
SPZ

osz
ssz
.09z

(Grid soacina 1500 m)

60

55

50

45
h

E

.P
0

-

40

; 35
3c

2

v

c b

25

2c

V

1E

1(

0.3
0.2 0.1

00 .
-0.1

-0.2 -0.3
MMO 02/01 1993
O(t00 02/02

MMO

mo o
02/04

0000
02/05

02/03

O w 0 02/06

0
ond calculated water level series

(Gridspacing 1500 rn)

.

. -

.~
06 . 0.4

Jizan calculated WL

0.2
0.0

-0.2 -0.1
-0.6 00730 02/01 1993

00w
02/02

0000

0000

0000

0000

02/03

02/04

02/05

02/06

Drawing m.

and colcLdated water level series

7.36

.

.

significant W v Eeight ae Directional Distribution Area : REDSeA Position : Reference Point A (41.53B 15.93N)

-

.....................................................................................
I ES (m)
I

1_____________1_________________________--------------------------------------------l I 5.5 6.0 I I 5.0 - 5.5 I I 4.5 - 5.0 I I 4.0 - 4.5 I I 3.5 - 4 . 0 I O+ I 3.0 - 3.5 I 0.3

I Direction (Waves coming from, relative to north) I 0 30 60 90 120 150 180 210 240 270

300

330 I Omni

I I
I I

-

I

I I
I I

I I

I I I I

2.5 2.0

-

O+
0.1 0.2 o+ 0.7 0.2 5.2 1.8 20.8 1.1 4.3 0.1 0.1

o+ I 0.3 I
1.7 4.8 9.1 24.5 49.3 10.3
0.0

1.5

I I I

1.0 0.5 0.1 Calm

-

-

3.0 2.5 2.0 1.5 1.0

I

I I I

o+
o+
O+ O+ O+ O+ O+

I 0.5 I I

O+

o+

o+

1.5 4.3 7.4 O+ 14.0 0.3 12.6 1.1

o+

I

0.1
0.4

0.5
3.3 6.6 1.3

1.5 4.8 1.3

0.1 1.1 0.6

0.2 1.1 0.5

I I I I
I I

Significant wave Eeight Directional Bxceedance : REDSEA Area Position : Reference m i n t A i41.53B 15.93N)

-

I I I I I

I > I >

> > >
> >

3.0

2.5
2.0 1.5 1.0 0.5 u.1 Calm

I I I

I I
I I

o+ o+
0.1 0.1

O+ 0.1

O+ O+ O+

O+ O+ O+

0.3 1.9 6.2 13.6 O+ 27.6 0.3 40.3 0.341.4

O+

o+
0.2 0.6 2.1
0.1 1.2 1.8 0.2 1.3 1.8

6.9 8.2

0.1 0.1 0.3 0.2 O+ 1.0 0.6 0.2 6.2 3.9 2.0 27.1 10.5 3.191.411.8

I

I

2.1 6.8 15.9 40.4 09.7 100.0 0.0

MUlt
Prqea:

SAUDI ARAMCO
RED SEA HINDCAST STUDY
Drmirq m.

Fi

Directional Distribution of Waves
I
%de:

Init:

Frequency of Occurence and Exceedonce drk at reference point A

8.1

Area

Significant wave eeight Directional Distrfblltion : REDSEA position : Reference Point E (39.88B 18.lSN)

-

.....................................................................................
-

I 8s (m) I Direction (waves coming from, relative t o north) I I I 0 30 60 90 120 150 180 210 240 210 300 330 I Omni I 1_____________1_________________________----------------------------------------I 5.5 6.0 I I I 5.5 I I I I 5.0 I 4.5 5.0 I I I I I I 4.0 4.5 I 4.0 I O+ I 3.5 I 3.0 3.5 I 0.1 I 2.5 3.0 I 0.3 I 2.0 2.5 I 0.1 0.1 1 . 4 1 1.5 2.0 1 0.3 o+ o+ o+ 0.3 3.6 o+ 0.1 7.2 I 12.1 I I 1.0 1.5 I 1.2 0.3 0.2 0.3 1.1 6.4 0 . 1 0.3 2.4 18.6 I 30.8 I I 0.5 1.0 I 2.5 0 . 8 0.6 0.9 4.0 9.7 0.2 0.1 0.2 0.3 3.1 24.1 I 46.6 I 0.5 I 0.2 0.1 0.1 0.1 0.2 0.3 o+ o+ 0.7 2.5 I 4.3 I I 0.1 I Calm I I 0.0 I

-

Significant Wave Height Directional Exceedance Area : RF.DSE-4 Position : Reference Point B (39.88E 18.15N)

-

.....................................................................................
.-

I 8s (m) I Direction (waves coming from, relative t o north) I I 0 30 60 90 120 150 180 210 240 270 300 330 I Omni 1_____________1_________________________-------------------~------------------------, I > 5.5 I I I > 5.0 I I I > 4.5 I I I > 4.0 I I I O+ o+ I 0.1 I > 3.5 I > 3.0 I 0.2 0.2 I 0.4 I > 2.5 I 0.4 o+ 0.1 0 . 1 I 1.3 I > 2-0 1 0.1 0.1 1.8 O+ ~7.5 3.6 1 6.1 I > 1.5 I 0.3 o+ o+ o+ 0.3 5.3 o+ o+ 1.2 10.9 I 18.2 I > 1.0 I 1.5 0.4 0.3 0.3 1.4 11.8 0.1 0.3 3.6 29.4 I 49.0 I > 0.5 I 4.0 1.2 0.9 1.2 5.4 21.5 0.3 0.1 0.2 0.7 6.7 53.6 I 95.1 I > 0.1 I 4.1 1.3 0.9 1.3 5.6 21.8 0.3 0.1 0.2 0 . 1 7.4 56.1 I 100.0 I Calm I I 0.0

I I

I
I I

I I I
I

1 I
I I I

I

Cht:

SAUDI ARAMCO
RED SEA HINDCAST STUDY
Draiw m.

F 4

PWjSt:

Wed Apr

3 I-€

Directional Distribution of Waves Frequency of &curence nd Exceedonce at reference point

B

8.2

Significant wave Beight Directional Distribution Area : -SEI\ Position : Reference Point C (38.333 20.43N)

-

.....................................................................................
I ES I I
I
(m)

I Direction (Waves coming from, relative to north) 60 90 120 150 180 210 240 270 I 0 30
6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5

300

330 I h

i

I I I I

I
I I

I I
I

I

5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.1 Calm

-

I I

-

I I I I I I
I

-

I I
I

0.1 0.3 1.0 2.2 5.0 0.5

0.1 0.2 1.0 0.1

o+
0.1 0.7 0.1

o+
0.1

o+
0.6
1.8 0.2

0.6 0.1

0.1 0.2 1.2 4.0 5.1 0.5

o+
O+

o+
0.1 0.2 0.1
O+

0.3 0.2

0.4
O+

0.2 0.4 0.1

0.1 0.4 1.1 4.5 o+ 9.8 0.4 24.0 2.3 27.6 0.3 1.6

I

0.1 0.4 1.3 9.0 12.2 32.0 45.4 3.6 0.1

I
I

I I
I

I

significant Wave Eeight Directional Exceedance Area : RBDsEA Position : Reference Point C (38.333 20.43N)

-

I 8s (m) I I > I > I >
I

I Direction (waves coming from, relative' to north)
I I
0

1__________---1--_______________________-------------------------------------------.
5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.1 Calm

30

60

90

120

150

180

210

240

270

300

330 I Onmi

I

I
I

I
I I
0.1 0.4 1.4 3.6 8.6 9.1 0.1 0.3 o+ 1.5 0.6 5.5 2.4 10.6 2.6 11.1

I
I

I > I > I >
I I I

> I >

> >

I I
I I

I
I
I

>
>

I I
I

0.1 0.3 1.3 1.4

o+
0.1 0.8 0.9

o+
0.1 0.7 0.7

o+
0.1

0.4 0.6

0.1 0.3

o+
0.4

0.3

0.5

0.1 0.5 1.6 6.1 O+ O+ 15.9 0.2 4 . 5 39.9 0.6 2.8 67.5 0.6 3.1 69.1

I
I I

I
I
I

I I
I

____________-----_______________________--

0.1 0.5 1.8 6.8 19.0 51.0 96.4 100.0 0.1

cislt:
Pr0j-t

SAUDI ARAMCO
RED SEA HINDCAST STUDY
baring m.

F 4

Danlsh Hydraulic Institute
File:

8
0.3

D d L Wed npr hit:

1996

Directional Distribution of Waves

%de:

drk

Frequency of Occrrence and Exceedance at reference point C

.

. .

significant wave H e i g h t Directional Distribution Area : RBDsEpl p o s i t i o n : R e f e r e n c e mint D (37.33E 23.03N)

-

I I I
I

5.5 5.0
4.5

4.0
3.5

I
I

I I I I
I I

3.0 2.5 2.0
1.5

1.0
0.5

I

0.1 Calm

-

6.0 5.5 5.0

I

I
O+ O+

4.5 4.0 3.5

o+ o+

I
I

3.0 2.5
2.0
O+

o+
O+

o+
o+
O+ 0.4 O+

0.1 0.7 2.0

0.1 I

o+
O+

1.5 1.0
0.5

0.4
0.1

0.1

O+

O+

O+

0.1 0.9 0.7

O+

O+ 6.6 0.5 15.2

0.1 0.1

O+

0.1 0.2

1.528.9

o+

0.1

0.1

0.1

0.3

0.1

1.5

1.6 I I

4.8 I 0.0 I

S i g n i f i c a n t W a v e aeight Area : REDSEA

-

D i r e c t i o n a l Exceedance

P o s i t i o n : Reference P o i n t D (37.333 23.0313)

I I I I I

>
> >

4.0
3.5

I
I I I

o+ o+ o+ o+
0.5
1.5 1.5

I
I

o+
0.2 0.9 2.9 9.6 25.5 58.1 95.1 100.0
0.0

> > I >

I >

I >
I >

3.0 2.5 2.0 1.5 1.0 0.5

I I
I

o+ o+
O+

o+
O+

o+
O+ O+

o;r
Calm

I

I I I

0.1 0.4 0.4

o+
0.2 0.4

o+
0.2 0.3

o+
0.3 0.4

0.1 1.0 3.5 4.2

0.5

0.2
0.5 0.6

1.2 1.5

0.1 0.4
0.5

0.1
0.6

0.8

0.2 0.9 2.9 O+ 9.5 0 . 5 24.6 2 . 0 53.5 7 . 0 79.3 k.5 8 0 . 9

I

I I I
I

I
I

I

0

Frequency of Occurence ond Exceedme

at refereme point D

Significant wave Eeight Directional Distribution : REDSEA Area Position : Reference Point E (35.48E 26.00N)

-

.....................................................................................
I E8 (m)

I

300 330 I Omni I 1_____________1_________________________----------------------------------------I 5.5 6.0 I I I I 5.0 5.5 I I I 5.0 I I I I 4.5 I 4.0 4.5 I I I I 3.5 4.0 I 0.1 I 0.1 I I 3.0 3.5 I 0.5 I 0.5 I I 2.5 3.0 I O+ 1.6 I 1.6 I I 2.0 2.5 I o+ 0.2 4.4 I 4.6 I I 1.5 2.0 I o+ o+ o+ o+ o+ o+ 0 . 1 0.4 11.2 I 11.8 I 1.0 1.5 I 0.2 0.1 0.1 O+ 0.1 0.2 2.3 26.7 I 29.6 I 0.5 1.0 I 0.9 0.3 0.1 0.1 0.1 0.7 0.4 0.2 0.2 0.6 5.7 31.3 I 40.6 I 0.1 0.5 I 0.5 0.2 0.1 0.1 0.2 0.5 0 . 4 0.1 0.1 0.5 3.6 4.7 I 11.1 I calm I I 0.1

I Direction (Waves coming from, relative to north) I 0 30 60 90 120 150 180 210 240 270

I

-

Significant Wave Eeight Directional ExcQedance Area : REDSEA Position : Reference Point E (35.403 26.00N)

-

.....................................................................................
I EP (m)
I I I

I Direction (Waves coming f-,
I
0

30

60

90

120

150

relative e0 north) 180 210 240 270

I
300 330 I Omni

I
I

1_____________1_________________________--------------------------------------------

I
I

I I
I

I
I I I

> > > > > > > >

I >

>

> >

I

5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1 ..o 0.5 0.1 Calm

I
I

I
I I
I

I

I
I I

I I
I

I
I I

o+
o+
0.1 0.2 1.2 1.7

I
I

o+ o+
0.3 0.5 0.1 0.3 0.1 0.1 0.1 0.3

O+

O+

O+

O+

I I I

0.1 0.8 1.4

0.1
0.4

0.9

0.1 0.3 0.4

0.1 0.3 0.4

0.1 0.3
0.8

1.3

0.2 0.6 2.9 8.6 12.1

0.1 I 0.6 I 2.2 I 6.6 I 17.1 I 44.4 I 75.7 I 80.4 I

0.1 I 0.6 I

I

2.2 6.8 10.6 48.2 88.8 99.9 0.1

I

I
I
I
I I I

ai rent: -w
Fmject:

SAUDI ARAMCO
W

Danish
F k

Hydraulic

Institute
1996

RED SEA HINDCAST STUDY
Baing rn

1z I

DO^ Wed +

D i e c t i m l Distribution of Woves Frequency of Occwence ond Exceedonce

SCde:

hit.

drk

at reference point E

I

0-5

I

Significant W r Height an Directional Distrib”tion Area : REDSEA Position : Reference m i n t F (34.66B 28.6611)

-

....................................................................................
I 8s (m) I I I
I I
5.5 5.0 4.5 4.0 3.5 3.0 2.5

I Direction (Waves coming from, relative t o north) I I I 0 30 60 90 120 150 180 210 240 270 300 330 I Omni I 1_____________1_________________________--------------------------------------------l

I
I I I

2.0
1.5 1.0
0.5

I I
I

I

0.1 Calm

-

-

6.0 5.5 5.0 4.5 4.0 3.5 3.0

I

I

I

I

I
I I

I I
I
I

I I

I
I

2.5 I 2.0 I 1.5 I 0.5 1.0 I 7.6 0.5 I 11.2 I

o+
0.4
0.1
O+ O+

O+ 0.1

0’.1 0.1 0.5 0.5

0.2 0.9

I I I o+ I o+ 0.7 I 0.5 19.1 I
3.3 30.7 I I

o+
1.2 27.6 47.6 23.6

Significant W v Height ae Directional B cc X-e Area : REDSEA Position : Reference Point F (34.66B 28.66N)

-

.....................................................................................
I 8s (m) I I I I I Direction (Waves coming from, relative to north) I I 0 30 60 90 120 150 180 210 2 4 0 270 1_____________1_________________________--------------------------------------------1
300

330 I Omni

I I

I > I >

> > > >
> > > > > >

5.5 5.0 4.5
4.0

I

I I
I

I I I

I I I
I

I I I I I I
I

3.5 3.0 2.5 2.0 1.5
1.0

I

I I I I I

I
I I

I

o+
O+ O+

I I I
I

o+
1.3 28.8 76.4 23.6

0.50.1 Calm

0.5 I 8.1 I 19.3

0.7

0.4

0.1

O+ O+

O+

0.1

0.1 0.6

0.1 0.6

0.2 1.1

05 19.8 I . 3.850.4 I

I

I

Urnt
FTojCEt:

SAUDI ARAMCO
RED SEA HINDCAST STUDY
Omn m. r ig

ij

Danish Hydraullc
Fk i

Instltuto
19g6

8
8.6

Datte: wed Apr

Directiml Distribution of Woves Frequency of Occureme and Exceedonce at reference point F

%de:

Ht:

drk

a

APPENDIX B
Oceanweather Inc. Final Report

FINAL REPORT

RED SEA BASIN-WIDE STORM AND CONTINUOUS WIND AND WAVE HINDCASTS Submitted to Danish Hydraulics Institute Horsholm, Denmark by Oceanweather Inc. Cos Cob, CT, USA

June 30, 1996 (original submission January 4, 1996) This report is a deliverable under a Subcontractor Agreement between Danish Hydraulics Institute and Oceanweather Inc., subject to Danish Hydraulics Institute Contract No. 45057/00, 27 February, 1995 with Saudi Aramco governing a Red Sea Hindcast Study.

a

TABLE OF CONTENTS

ACKNOWLEDGEMENTS.
1. 2.

INTRODUCTION DATA SOURCES
2.1 2.2 2.3 2.4

1

3
3

Introduction Land Station Synoptic Observations Synoptic Ship Reports Other Data

4
6

7 10 10
11

3.

STORM SELECTION
3.1 3.2

Background Climatology Storm Selection

4.

WIND AND WAVE HINDCAST METHODOLOGY
4.1 Specification of Wind Fields 4.2 Wave Hindcast Model

18 18 24 34
36

5. STORM HINDCASTS

6. CONTINUOUS 3-YEAR HINDCASTS

REFERENCES TABLES FIGURES APPENDICES

38 43

55
71

A. Grid system point listing
B. Sample storm wind field plots
C. Check plots of three year continuous hindcast

D. Formats of delivered wind, pressure and wave data
E. Reprints of selected paper
F. Selected results on additional storms

ACKNOWLEDGEMENTS At Oceanweather, the technical work was carried out by the following staff members: Dr. V. J. Cardone (project director), . Dr. J. A. Greenwood (wave model adaptation), Ms. J. G. Greenwood (graphics and archive production), Mr. A. T. Cox (wind workstation adaptation), Mr. M. J. Parsons (storm selection, wind field analysis, and overall production management) and Ms E. Ceccacchi (wind field analysis).

1 INTRODUCTION .

This report describes Oceanweather Inc's (OWI) contribution to a project of Danish Hydraulics Institute (DHI) whose objective is to develop definitive meteorological and oceanographic extreme and normal design data in all of AFSMCO's stated areas of interest along the Saudi Arabian coast of the Red Sea. The overall project consists of six major Tasks: Task Task Task Task Task Task
1. Review of Available Data and Reports. 2. Normal and Extreme Wind Data and Storm List

3. Normal and Extreme Currents and Water Levels 4. Normal and Extreme Waves
5. Joint Statistics
6.

0

Miscellaneous Information

OW1 contributed to Tasks 1, 2 and 4 by hindcasting winds and waves on a high resolution grid covering the entire Red Sea for 3 0 storm events selected from the last 20 years, and also for a continuous three year period. Historical meteorological data were gathered, reviewed and utilized principally to allow the production of the most accurate representation of synoptic wind fields possible t o be used for both the storm and continuous hindcasts. Synoptic pressure fields were also produced for the storms only since they were also needed to drive the hydrodynamic model of DHI under Task 3.

0

The wave hindcasts were made with a third-generation wave model adapted to the basin on a grid of spacing 0.25O in latitude and longitude (Grid listing in Appendix A.) The hindcast wind fields on the same grid system were delivered in computer readable form for use by DHI in Task 3. Hindcast waves were also delivered in computer readable form so that together with the wind fields they could be used by DHI to drive local ultra-fine mesh wave models set up for each nearshore study target area.
1

This report describes the data gathered and utilized in this study to select storms and develop wind fields (Section 2), outlines the storm selection procedure (Section 3), describes the wind and wave hindcast methodology (Section 4) and gives a concise summary of the storm (Section 5) and continuous hindcasts (Section 6). We omit lengthy expositions of our methods because they are well established and described in open publications, copies of the most relevant of which are provided in an Appendix to this report.

2

2.

DATA SOURCES

2.1 Introduction

The hindcast models used to generate the wave and currentlsurge time series both in storms and for the continuous periods, are .driven by time histories of the surface wind and, in the case of storms, atmospheric pressure. The process of developing such wind fields and pressure fields had to work mainly from conventional historical meteorological data archived through the normal mechanisms established by the World Meteorological .Organization (WMO), since little additional data were made available to this study by its sponsors. A few limited meteorological and oceanographic data sets have been acquired in recent years for very limited time periods in connection with specific coastal engineering projects (e.g. Koch and Shabib, 1989). Such data are useful to study specific processes (e.g. the sea breeze wind cycle offshore Yanbu) in specific areas and possibly to validate and calibrate near shore models, but in general such data have little value for derivation of wind fields in historical storms and for continuous periods because their historical coverage is limited and in general do not include historically high-ranked events. Most of the preserved digital archive of historical conventional meteorological data for the Red Sea fall into three categories: (1) synoptic observations from land stations in countries bordering the Red Sea.
(2) synoptic ship reports in the Red Sea and Gulf of Aden;
(3) climatological summaries of (1) and (2)

3

In addition to these data, weather maps covering the area are
available from some sources.
2.2 Land Station Synoptic Observations.

Synoptic observations from land stations are often archived at local WMO member country weather stations or central archiving centers in manuscript form (WMO standard forms of surface synoptic observations, anemogram traces, climatological summaries). Such manuscript data were, in general, not available t o this study. However, since the early 1970's, computer readable series of observations for many stations which have reported regularly in real time over the global telecommunications system (GTS) have been archived through the interception and recording on magnetic tape of transmissions through main trunk lines of the GTS at major weather analysis centers. The relevant data from this data source were assembled for this study in three parts: (1) The first part had already been obtained from the US N O M National Climatic Data Center (NCDC) prior to this study and archived by OW1 in connection with previous studies and consisted specifically of data for WMO Blocks 40 and 41 which include all stations in Saudi Arabia, Jordan and Yemen. These data are contained on eleven 6250-bpi density 9 track tapes in ASCII format. These data cover the period 1973 1989 inclusive in general but with some gaps in years (e.g. Jeddah is not included for the years 1980-81) and within the time series there often missing hours. Since this source depends on GTS transmissions, the number of available stations and the density of -observations per station varies through this period. Typically, the station data are given as hourly, 3- or 6-hourly observations in WMO synoptic code, which includes surface pressure and surface wind. Table 2.1 shows the typical contents of this file and Figure 2.1 shows the locations of the stations.

-

4

(2) The second part was simply an update of the above data base

through purchase of additional magnetic tapes from NCDC. Since the archiving procedures, data formats and costs for this type of data have changea significantly in recent years at NCDC, data were obtained only for the following stations for the period
1990

-

1993:

.Hodeidab

I . .

L -

=

l i

..

Jeddah Kosseir Jizan -*Yanbu -_ It should be noted that parts 1 and 2 do not provide complete coverage for the stations because they are essentially recordings of those reports transmitted in real time over the GTS, and for some stations and time periods only 50% or even fewer of the observations originally made are actually archived in this source. As a part of this study, completed series of observations were requested from the study sponsor. However, only the following summaries were provided. (3) The third part consists of diskettes of daily summaries of surface observations provided by the Saudi Meteorological and Environmental Protection Administration for the following stations and for the years 1986-1993:

-

Jeddah Jizan Yanbu These data were found to be of limited use for the analysis of wind fields because only the daily maximum wind speed and associated direction are given with no indication of the hour of

observation of that maximum and the averaging interval used for the wind.
2.3

Synoptic Ship Reports.

Much of the ship report data were already on hand at OW1 before the beginning of the study. OW1 maintain a fairly complete copy, on several hundred 6250-bpi density 9 track tapes, of global collections of individual historical ship report data (the socalled Comprehensive Oceanic and Atmospheric Data Set or corns) and have pioneered methods for the interpretation and utilization of surface wind observations from transient ships (Cardone et al., 1990, reprint attached). The marine data file was current through June 1993 at the start of this study. A scan of these data revealed missing data in the relevant Marsden Squares covering the Red Sea only for the year 1985. The density of marine observations archived in the Red Sea varies greatly over the past 50 years. For example, there are few observations archived before 1950, fewer than 5,000 observations per year before 1960, virtually no observations between 1967 and 1975 when the Suez Canal was closed, about 10,000 observations per year between 1975 and 1979, and about 20,000 observations year during the decades of the 1980's and 1990's. At the start of this study the additional data purchased from NCDC to complete the file included: (a) Ship reports in NCDC's latest COADS update from beginning of record (late 1800's) through 1979 to check if the new inputs to COADS in recent years have added ship reports. b) All reports in 1985 in Marsden Squares 68,69.

6

c) All reports in Marsden Squares 68, 69 for the period July 1, 1993 October 31, 1994 (most recent available at the time of order).

-

Since 1976, the density of observations appears to be sufficient, after the application of a sophisticated analysis scheme, to resolve the general evolution of the wind field in the Red Sea on a synoptic basis (6-hourly or 12-hourly); however, the marine observations are confined to an extremely narrow shipping lane down the center line of the basin; hence, the ship reports alone are not sufficient to define either the extreme or normal wind ..and,waveclimate of the study areas, which are located along the eastern coastal margin of the basin and east of the main shipping lane. The ship report file required considerable quality control and tedious editing to eliminate obviously erroneous data.
2.4

Other Data

a This compilation is produced .
by the US NOAA National Climatic Center and includes monthly and annual summaries of air and sea temperature, dew point temperature, scalar wind speed, sea level pressure, wave height, wind and ocean-current roses in user defined 1-degree squares. This source tends to duplicate and supersede many earlier hardcopy sources such as the Marine Climatic Atlas series of the U. S. Navy, and U. S. NCDC SSMO series, and the atlas of Hogben et al. (Global Wave Statistics, 1986). International Station Meteoroloaical Climate Summary. This data base, compiled as a joint effort of the NOAA NCDC, U. S. Navy and U. S. Air Force, contains detailed climate summaries for 640 world-wide stations and 5434 airfields. Each summary contains up to 40 separate tables and up to 108 sub-tables.

7

a

Weather M a w . The only map source available to us covering the area of interest is the microfilm film known as the NOAA elTropicalStrip" file archived at NCDC. A complete file is also maintained by OWI. This file begins in 1969 and provides twice daily (in early years) and later four times daily surface maps which show plotted surface observations, and winds derived from tracking low level clouds on satellite pictures and streamfunction lines. Gridded Winds Fields. This consists of the so-called WMO file of twice-daily gridded 1000 mb winds for the period 1980-1989 on 2.5 degree spacing. These files provide only a crude first guess of the surface wind fields in the Red Sea for storm and continuous hindcasts. ReDOrtS of Previous Studies.
1. Final Report Volume IV-Meteorological & Oceanographic Data

Collection, August, 1980. Yanbu Industrial Complex, Kingdom of Saudi Arabia. Contract NO. GST E 4009 Navigational Safety Surveys. Prepared for Royal Commission for Jubail and Yanbu Directorate General for Yanbu Region
2. Climatic Study of the Red Sea South and Gulf of Aden.

September, 1982. Prepared by US Naval Oceanography Command Detachment, Asheville, N C. .
3. Final Report for Meteorological Monitoring Network at Madinat

Yanbu Al-Sinliaiyah. April, 1982. Prepared for Kingdomof Saudi The Royal Commission for Jubeil and Yanbu under Contract Arabia No. GST #E-4016, Task 11. Ecology and Environment of Saudi Arabia Co., Ltd.

-

8

4. Two papers from Proceedings of the Symposium on Coral Reef

Environment of the Red Sea, January 14-18, 1984. ed. M. A. H. Saad:
4 (a). Ages of tides in the Red Sea and Gulf of Aden, M. I. El

Sabh and T. S. Murty. 4 (b). Weather systems, storm surges and sea state in the Red Sea and Gulf of Aden
5.
--S.

Design Criteria for Yanbu Expansion Facility. February, 1989. P. Koch and H. M. Shabib, Exxon Production Research Company.

_I _

Wave Study Report (Port of Jizan). Date and authorship not provided.
6.

7.

Tide Tables. Yanbu VLCC Terminal, Jiddah, Jizan,

0

9

3 STORM SELECTION .

3.1 Background Climatology

The objective of the storm selection was to identify episodes of significant wind forcing over the Red Sea from as long an historical period as possible. Significant forcing means the type of condition which can be expected to induce an ocean response in terms of waves, currents and surge in the specific areas of interest in the study. The target storm population is to consist of 30 storms. The general climatology of the Red Sea has been described by Morcos (1970) and more recently by Murty and El-Sabh (1984). The particular variable of interest to this study is, of course, the surface wind. The surface wind regimes in the Red Sea are a result of the interplay between the large scale forcing by major extratropical pressure systems and small scale forcing driven by the large desert-sea contrasts in temperature. Both the large scale and small scale forcing are strongly influenced by the high relief on both sides of the basin, which tend to rectify the wind flow t o blow either up or down the long axis of the basin. However, under certain conditions the coastal relief induces downslope winds which blow out over sea nearly perpendicular to the coastline. These **katabatic'* winds may be quite strong at the immediate coast but the wind speeds typically decay rapidly offshore rendering them of little significance for generation of surge or waves in the concession area compared to the along-basin flows. On the large scale, the northern part of the Red Sea, north of say 20°N, is much more influenced by transient effects of travelling depressions and fronts than the southern part. As frontal troughs moves generally from east to west, they induce veering wind response. That is, in the the typical backing

-

10

north, where the winds are generally northwesterly throughout the year, as a depression approaches the surface winds back to southerly in advance of the front and veer back to the climatological normal northwesterly direction after the passage of the front. In the southern part of the basin, the prevailing flow is northwest from May or June to September and southeasterly during October to April or May. Thus, often during the cool season, there is northwesterly flow in the north and southeasterly flow in the south. During episodes of significant forcing, the boundary between these regimes will become quite dynamic. Even during the winter, during a single such episode, strong southeasterlies may penetrate northward to cover the full extent of the basin, then may retreat far to the south as enhanced northwesterly, or Shamal, propagates southward to the entrance to the Gulf of Aden. Superimposed on the large scale pattern are many mesoscale circulations described locally as: Khamsin, Aziab, Khalil Lull, Kharif and Saba and the familiar thunderstorm or squall line. Locally, within about 10-20 km of the coast, there is the pervasive sea breeze circulation which is especially pronounced during clear sky conditions with weak large scale forcing. The sea breeze is further complicated in areas of significant terrain effects. The relief when it is combined with Ifkatabaticff mesoscale and local circulations tend to be especially energetic along the coast, where unfortunately, the synoptic observing stations tend to be located, thus rendering the coastal observations less representative of conditions over the open sea. A concise summary of the normal wind regimes in the Red Sea was given by Murty and El Sabh (1994) and this is repeated here in Table 3.1.
32 .

‘I
L

Storm Selection
11

The identification, screening and ranking of storms was based on two main sources of wind data: (1) time series of wind observations from selected coastal stations; (2) ship reports. As noted in the previous section, coastal station data are available only since 1973 in digital form, and since 1969 as plotted on historical weather maps. The closure of the Suez Canal between 1967-1975 caused a virtual absence of ship reports during that period. Both of these factors mitigate to force the storm selection to be restricted to the period beginning in 1976, because while it might be possible to "identify" storms from earlier years in coastal land station measurements it would not be possible t o specify accurate wind fields in those events because one or more data types would be missing. The storm selection process can not be purely objective because no single complete consistent data source exists which alone might reveal the occurrence of storms for each major region within the Red Sea. Rather, storms must be sought in a process not unlike that followed by a detective in the investigation of a crime. In this case the various measured data and weather maps provide f*cluestt which if collated, screened and interpreted carefully reveal an event of significant wind forcing over a wide enough area of the Red Sea to constitute a "storm". Further interpretation and study of each identified event is necessary to filter only the top 30 events from hundreds of candidates. The various steps followed in the storm selection are described below.
Step 1. Preprocessing of Ship Reports

Ships from Marsden squares 68, 69, 104, and 105 (these 30°N, squares taken together encompass the area 10nN 50°N) were extracted from the updated library of COADS tapes. After ensuring that our files were complete for

-

four 30°N ship these

12

squares for the stated period (January, 1970 -October 31, 1994), the ships were divided into two files holding, respectively, reports north and south of 2 0 ° . The files were further divided into decades. Each ship report of wind speed was then processed following the procedures described by Cardone et al. (1990). Simply, if a ship report is based on estimation of the Beaufort Force, it is "correctedq1using a scientific scale which effectively raises reported wind speeds lower than about 15 m/s and lowers winds speeds greater than about 15 m/s. The adjusted winds are representative of the 20 m elevation and neutral stability. winds measured with an anemometer are corrected to 20 : height and neutral stability using the anemometer height which m is looked up in a table based on ship call sign and the ship report of air-sea temperature difference. If the anemometer height is unknown, an intelligent guess is made (see Cardone et a1.,1990 for further details). Table 3.2 is an example of the processed file of all ship reports archived at 00 UT, 06 UT, 12 UT and 18 UT on February 9, 1981 SteD 2. Ship event summary and tally program. computer summary was made of the ship files. This summary generates a daily one-line summary of selected properties of the ships for each calendar day and includes the average of the reported wind speeds, the ship report with the highest wind speed, the ship report with the highest sea state and a count of the number of ship with wind speeds of 30 knots or greater. A separate summary was produced for the "northern area" and the I1southern areai",with 20° taken as the dividing latitude between these areas. A hard copy listing of the summary for each area covering every day from January 1, 1976 through October 31, 1994 was then visually scanned to identify discrete events above a trial relatively low threshold. For each such event, the ship reports were actually plotted for the days at and round the event peak (over one thousand plots).
A

;i, :

.-

-

0

13

The first pass through this list identified an inordinate number of gveventsgg the mid-1980's with maximum winds always observed in at ships whose call sign began with U (Russian source decks). This was eventually traced to a problem with the COADS ship file, which evidently contains some ships' reports which appear twice a the same synoptic hour, once with the wind speed coded in knots and the second with the same speed coded but with the speed units indicator reporting m J s , with the speed therefore erroneously doubled t o knots. These erroneous reports had to be filtered from the files used for the storm selection and wind field analysis. SteD 3. Land data summary and tally program. After exploring trial thresholds, the station surface wind speed observations (as reported) were searched for occurrences of wind speeds exceeding the following: Station A1 Wajh Hodeidab Yanbu Jizan
Jeddah

Threshold ws > 20 ws > 25 ws > 20 ws > 15 ws > 20 ws > 20 kts kts kts kts
kts

No. of events
94

34 50
27
35

Aden AP

kts

55
296

Trial events were identified for each station when the criteria were exceeded for at least 9 hours (observations were usually available hourly or three hourly). The station event listings were then cross-correlated to find events which were coherent in time at adjacent stations. The event listings were also correlated against the daily ship report listings produced under Step 2, and if no ship report plot had been produced, one was plotted corresponding to the date in the land data listing, to
14

see if any corroboration existed between the land stations and ship reports. Another approach was to sort the land station files by wind speed and date using only nighttime observations to suppress the seabreeze induced daytime maximum wind speed which often in tranquil background conditions could equal or exceed the wind speeds observed in storm events offshore. At this step, is was apparent that the highest ranked wind speeds over say a month or a year in a land station time series were often erroneous and reflected miscoded data or the result of faulty GTS transmissions. This was .easily verified by examining the time series around each maximum and identifying "spikesvt. This problem also plagued a list of annual maximum wind speeds supplied by the sponsor for Jizan, rendering that series virtually useless. SteD 4. Coarse candidate list. The so-called lfcoarse*t first event list established the date or for the following number of events for each decade and area. Based on ship and land data:
1970's North 1970's South 1980's North 1980's South 1990's North 1990's South

.

33
35 108
87

57
32

events events events events events events -

Based only on land data as tabled above: Land 1973-1993 296 648 Total events events

15

The above lists included possible overlapping events between the northern and southern sections.
SteD 5.

Initial: Ranking.

Using the ship report plots and land station data listings the tedious process of ranking the events in terms of wave generation potential proceeded by using a three step scale in which an event was characterized as of low (#1), moderate ( # 2 ) and high (#3) severity. Then, the following distillation was made: (a) Because of the-large number of events, events of rank 1 and events with hopelessly little data were dropped. (b) After cross-correlating events in the northern and southern areas a total of 24 events of rank 3 were identified, four of which were dropped for lack of sufficient date to carry out an accurate wind field analysis (e.g. only one ship report at a synoptic time in the entire basin). (c) -Since there were an insufficient number of events with rank 83, the rank # 2 events were restudied and after the weaker events were dropped and the rest added to the rank # 3 candidates, the following candidate list emerged: events 1970's South 24 events 1980's North 58 events 1980's South 51 events1990's North 31 events 1990's South 21 events Based only on land data distilled as noted above 7 events Land 1973-1993 4 Total 249 events
17 1970's North

16

This list included the potential for overlapping events in category 82 events. Stev 6. Final distillation. To reduce the storm size to the final target size of 3 0 events representing all areas, first the overlapping events from the north and south data sets were identified. Events were then classified as strong south, strong north or strong north and south. This list was then studied independently and jointly by two meteorologists and filtered scan by scan until a list of 33 events was obtained. All of these events were carried well into the individual storm wind field analysis process before the three weakest of these events could be identified and dropped.

j

0

17

4.

WIND AND WAVE HINDCAST METHODOLOGY

4.1 specification of Wina Fields

Basically the same approach was used to develop wind fields for each storm as was used for the continuous 3-year period. The method builds up wind fields directly from synoptic winds measured from ships and land stations. However, the raw observations are recognized to be biased estimates of the true over-water wind at a standard reference height and are therefore modified systematically before use. In the case of ship reports, considerable quality control is necessary to .prevent,erroneous or miscoded data from contaminating the analysis, and winds must be adjusted to standard height using a method which depends on stability, measurement type (i.e. Beaufort estimate or anemometer measurement) and height (Cardone et al., 1990). Winds measured at the coastal station winds reflect, and are often dominated by, local effects such as poor anemometer exposure, surface roughness significantly different from that of the sea surface, local topographic forcing and the strong diurnal cycle typical of land stations in a desert zone. The sea-breeze quickly decays a short distance offshore. All of these effects conspire to make the land data quite poor indicators of the surface winds over the water even a short distance offshore, making such data unsuitable for use directly in purely objective wind analysis schemes. In this study, little of no information was provided by MEPA on those factors which might make an evaluation of the representativeness of each station possible. However, since the wind analysis procedure is interactive, the analyst developed subjective impressions on the representativeness of each station which routinely report, and used that knowledge to suitably modify or weight the land data in the process of the analysis. The method we use is called Interactive Objective Kinematic Analysis (IOKA).

-

18

Jnteractive Obiective K inematic Analvsis. OW1 pioneered and demonstrated the value of classical kinematic analysis techniques to the specification of marine surface winds in historical storms from conventional data, both in mid- and high-latitude settings as well as in tropical settings (e.g. Cardone et. al, 1980; Cardone, 1991; Cardone et al., 1994a). The method is extremely costly and time consuming and there are few skilled practitioners. Kinematic analysis, however, is necessary in the Red Sea because of the poor relationship between.the surface wind field and the.surface and upper air geopotential +84::-.fh fiead. ,WindIfield products of objective analysis systems ..e 9;I . *.* . ..' operated.by major numerical weather analysis and forecast center , . . *; 2 .?. ) such as the U.S. National Meteorological Center (NMC), the U. K. . Meteorological Office (-0) and the European Center for Medium i. : Range Weather Forecast (ECMWF) are simply not accurate enough to provide accurate wave and current hindcasts (as shown for example "'by Cardone et al., 1994a). Recently, OW1 have been able to make the kinematic analysis process more efficient without sacrificing accuracy by developing a system called IOKA, which is implemented on a'wind-workstation (WWS) with a graphical user interface to facilitate the subjective analyst input. W S is described W recently by Cox et al., 1995, reprint attached).
I ,

0
~

For this study IOKA/WWS was adapted to the Red Sea over the 30a N, 32O N 45O N on a grid spaced in domain loo N increments of 0.25 in latitude and longitude. This domain therefore includes the Gulf of Suez, the Gulf of Aqaba and the western half of the Gulf of Aden. Wind fields are output at 3hourly intervals. However, for analysis of storms the basic interactive analysis is carried out at 6-hourly intervals (00, 06, 12, 18 UT), while for the continuous hindcast it was carried out at 12-hourly intervals (00 UT, 12 UT).

-

-

19

IOKA was developed to explore the degree to which computer based procedures may be used to minimize the laborious tasks of drawing and subsequently digitizing streamline and isotach analyses. The method retains manual intervention mainly in the quality control and editing of input data; the option to assimilate traditional gridded kinematic analyses produced by hand, and in the specification of discrete kinematic control points (KCP) to the analysis program to ensure continuity and preservation of wind maxima. IOKA accommodates up to a classes of input: not all inputs need be present in every application; for example, one input source, data from moored buoys, is not available in the Red Sea, and another, the output of a numerical model of tropical cyclone wind fields is not relevant since tropical cyclones do not occur in the Red Sea. With each input, a pre-processor is associated, taking the winds in whatever form they were originally written and turning them into ASCII files of date, latitude, longitude, u- and v-component of wind, wind speed, wind direction (an approximate value to facilitate visual inspection of the ASCII files; not used by subsequent programs). The ship reports of wind are also adjusted t o equivalent winds at 20 meter height using calibrated procedures which take account of whether the wind was measured or estimated. OW1 maintain a list of over 7,000 anemometer heights of merchant vessels. The eight inputs are:
1 Kinematic control points as provided by manual .

intervention in data-sparse areas.
2. New kinema: that is wind speeds and directions derived

from hand-drawn streamline-isotach maps and coded on a latitude-longitude grid.
3. Old kinema: that

is coded winds available from previous
20

kinematic analyses coded in various formats.

4. Tropical cyclone: winds from solution of Oceanweather's

vortex model, used only inside a certain circle centered on the eye of the storm.
5. Buoy or offshore platform winds.
6. Ship reports. Coordinates are given to 0.1 degree.
!

Winds are reduced to equivalent neutral 19.5 m height. In the default setting of the program, ship reports are ignored when they fall inside the circle defined above for cyclone winds.

~

1.

I

7. Gridded PBL Winds, i.e. winds previously calculated from

" i

isobar analyses entered into a computer through a digitizing tablet, or alternatively winds obtained by objective analysis as say a by-product of one of the WMO numerical weather prediction (NWP) centers such as the US National Meteorological Center, The US Navy Fleet Numerical Oceanography Center, the UK Meteorological Office or the European Center for Medium Range Weather.
8. Climatology: mean monthly winds on the boundary of the

analysis domain. These values are given a low weight, but are needed so that smooth values near the boundary may be effectively interpolated and not extrapolated. The method adopted for integrating all of these classes of observations into a smooth field of a scalar F consists of minimizing the quadratic form

21

..

where ut, is the weight assigned to inputs of class k; F is a , measurement of class k.

i, j, are indices for longitude and latitude at the SW corner of the grid box containing the measurement site;

C, =

(long

-

long,)/.25

Cy = (lat

-

lat,)/.25

are interpolating factors in latitude and longitude; and fl is a scale factor with the dimensions of length. The weights and scale factor of IOKA are determined for each application after some experimentation. Since, for the first few storm events analyzed, wind fields were also developed completely by hand analysis and gridded directly on the target grid, the weights and scale factor could be set to allow IOKA/WWS to reproduce essentially the same wind field as produced by the classical method, especially for storm peaks.
IOKA as sketched above, is applied to three fields F: U, V, and U2 + V2. The adopted wind speed is obtained by analyzing U2 + V2 and taking the square root; previous applications find this to yield tolerably lower bias in speed than analyzing wind speed itself, i.e. W = (U2 + V2)'?. The adopted wind direction is obtained by analyzing U and V and taking the inverse tangent.

0

0

Apulication to Storms. In the Red Sea implementation for storms, data types 1, 2, 5 and 6 were active. Type 1 data are input by
22

0

0

the WWS operator as the analysis proceeds. Since there was judged to be no suitable source of background gridded wind fields for Red Sea events, type 2 date were generated by manually gridding streamline-isotach analyses produced drawn by hand at 6-hourly intervals for each of the 30 storms. The gridding interval was irregular so as to provide more coverage for the critical parts of the analysis, but the average spacing was about one-degree. This tedious step seems to negate much of the efficiency introduced by IOKA/WWS but it was absolutely necessary to provide sufficiently accurate background gridded wind field for WWS to operate effectively. Land station data were brought into the wws as a category 5 type, but with much lower weight than given to a buoy, since the land data had already been considered, with subjectively applied adjustments, in the manually drawn kinematic analyses. For application to storms, IOKA/WWS was applied at 6hourly intervals for an approximately 4 5 day period per event representing about 2 days of spinup before the storm peak and two days of spindown after the storm peak.

-

..

The plots of the wind field for a typical storm are given in Appendix B. The case shown is for a moderately high-ranked NW event for the central Red Sea. ADDlication to Continuous Years. As applied to several continuous years, the analysis interval was relaxed to 12 hourly, as determined by the temporal resolution of the available background fields, namely 00 UT and 12 UT. Wind fields at greater temporal resolution as needed to drive the wave model were interpolated within WWS from these synoptic fields. The data W types considered in W S were 1, 5, 6 and 7. Type 7 winds were taken from our WMO file of 2.5 degree 12-hourly global 1000 mb winds. Type 1 data, the kinematic control points, played just as critical a role in the continuous analysis as the storm analyses. To generate these data the analyst had to interact with each and

a

23

every one of the 2,190 12-hourly analyses which comprise the 3year period. Figure 4.1 shows-the screen image (this is a black and white print of a color image, see front cover for color depiction) of a typical map. Input winds are shown as a filled circle a the base of the wind barb. Since this case is taken from the continuous file, background winds are plotted at every 2.5 degree intersection. Winds observed by land stations are plotted for several stations along the Saudi coast. The irregularly spaced observations (filled circles) of wind within the basin are the ship reports. Figure 4.2 shows the KCP points added interactively by the analysis using WWS. Both Figures also show winds plotted at regular 1-degree intervals as subsampled from the file of final winds calculated by WWS on the 114 degree grid. Figure 4.3 shows the final winds plotted at 112 degree resolution. Winds in these plots are adjusted to effective neutral 1-hour averages at a reference height of 20 m. These winds are reduced to standard 10 m height before they were transmitted to DHI for use in the hydrodynamic models and local area wave models.

4.2

Wave Hindcast Model

Both the storm and continuous period were hindcast on a regular grid covering the entire domain of the wind analysis using a highly calibrated third generation (3G) wave model incorporating shallow water effects. The hindcast approach is rapidly becoming the preferred way to develop both extreme and normal wave climate.

e

24

0

The wave hindcast model used in this study is a so-called fullydiscrete spectral wave model. That is, the wave spectrum is resolved in discrete frequency-direction bins, a grid of points is laid out to represent the basin of interest, and a solution is obtained based upon integration of the spectral energy balance equation, a process which successively simulates, at each model grid .point, and for each time step, the physical processes of wave growth and dissipation (through the source terms of the energy balance) and wave propagation. The model uses third generation physics and is believed to be the only such 3G model in use for commercial purposes. The model is called OWI3G for short. Grid Svstem The grid system is a latitude-longitude array as follows: Latitude range: 10°N to 30°N Longitude range: 32OE to 45OE by .25O (81 parallels) by .25O (53 meridians) (see also Appendix

-

_-

The grid system is shown in Figure 4.4
A)

-

Spectral Resolution

0

Direction: 24 bands. Band 1 is centered 7.5O clockwise from . true north, the width of each band is 1 5 O Frequency: Band 1 is centered on 0.039 hz; the bands increase in geometric progression (ratio = 1.10064) to band 23, .32157 hz.

25

Propaaation Scheme The downstream interpolation scheme described by Greenwood et al., (1985) is used throughout. Propagation over a time step at a grid point is implemented within the alternate growth-. propagation cycle in the model integration by forming linear combinations of spectral variances at neighboring points. The weights used are extracted from a precomputed table of propagation coefficients, which vary by latitude only in deep water, and are specific to each grid point in shallow water. The table of interpolation coefficients is calculated based upon great circle wave ray paths in deep water; in shallow water the weights are calculated following a ray tracing study through a digital bathymetry resolved on the wave model grid. The water depth at each grid point is computed as a simple average of all.depths at 5-minute intervals within a grid box contained in the U. S. National Geophysical Data Center (NGDC) ETOP05 global topographic-bathymetric digital data base. However, the ETOPS depths were found to be deficient in the area of the Farasan Bank, and therefore some depths were read off DMA chart 62001 were used t o overlay the ETOPS depths, though even this chart does not appear to well resolve the fine structure of the depth field over this Bank. The depth field is plotted at wave model grid points in Figure 4.5. The limiting water depth for shallow propagation and growth processes is taken according to the conventional definition: kd > n , where k = .006123 m for the .039 hz frequency bin. Thus, propagation by depth-dependent ray tracing is executed at points within 50 km in any direction of points with depths 513 m or less.

26

Sbectral Gro

0

't

.

,

,

...... .. ....._,., ,,,.
1 .

...

-

..

The theoretical basis of the spectral growth algorithm used in OWI3G follows clusely that of the first 3G model developed during the 1980's through the efforts of an international panel of wave modelers, (including Dr. V. J. Cardone and Dr. J. A. Greenwood of Oceanweather) known as the WAM group. The WAM model was originally described by WAMDI-Group (1988). In WAM, separate source terms for the physical mechanisms of atmospheric input, S , wave-wave interaction,S, and dissipation by whitecapping, S , , (and in shallow water dissipation by bottom friction, S ) are , specified explicitly and the source term balance is integrated to yield the-.net development of the spectrum over a time step of integration without arbitrarily forcing of spectral shape or specification of an external limit to growth. However, in the original development of the model, considerable experimentation ".with and tuning of the input and dissipation source terms was carried out (Komen et al., 1984) to achieve growth rates and asymptotic behavior under constant winds in agreement with field data. The wave-wave interaction "apparent" source term, S , , is not considered tunable and is a parameterization of the exact nonlinear interactions as proposed (see e.g. Hasselmann and Hasselmann, 1985). This so-called Discrete Interaction Approximation (DIA) is also described in WAMDIG (1988). OWI3G combines a source term representation and integration scheme based upon WAMDI with OWI's propagation scheme described above. The source terms follow the theoretical forms used in WAMDI but with different numerics and code and with the following modifications. First, a linear excitation source term is added to S,, taken as a downscaled variant of the term used in OWI's 1G model known also as ODGP. This allows the sea to grow from a flat calm initial condition in OWI3G, unlike all cycles of WAM which require an artificial warm start from a prescribed initial
27

spectrum. The exponential input term is the empirical form of Snyder et al. (1981) in which S, is taken as a linear function of . . However, unlike WAMDI in which U is . friction velocity U computed from the 10 meter wind speed U,, following the drag law of Wu (1982), in OWI3G, a different drag law is used developed in the model tuning stage. That drag law follows Wu closely up to about 20 m/sec then becomes asymptotic to a constant at hurricane wind speeds. The dissipation source term, S, is also taken from WAMDI except that the dependence on frequency is cubic rather , the DIA is adopted except that , than quadratic. Finally, for S two modes of interaction are included (in WAMDI the second mode is ignored). Oceanweather's version of 3 6 as just described was developed based upon tuning runs against the fetch-limited growth benchmark for 20 m/s wind speeds under constant winds used to tune WAMDI initially, as well as in repeated trial hindcasts of a welldocumented moderate extratropical cyclone (SWADE IOP-1, see Cardone et al., 1994a) and two intense Gulf of Mexico hurricanes (Camille, 1969; Frederick, 1979). The bottom friction source term is a'simple quadratic law with a specified tunable friction factor. OWI3G uses the same friction factor found in the North Sea version of WAM (NEDWAM) to yield skillful shallow water predictions. That factor, .076, is exactly twice the value originally proposed for WAM, which was based upon studies of pure swell attenuation in the North Sea JONSWAP experiment. An interesting comparison of the performance of Oceanweather's first generation model, OWIlG and OWI3G in an cxtratropical setting is given by Khandekar, Lalbeharry and Cardone (1994, reprint attached), which also includes further documentation of the growth algorithms. A comparison of the pcrformance of OWIlG, OWI3G and the latest cycle of WAM (WAM-4) in extreme storms is given in Cardone et al., (1994 b). Both of these
28

.

...

studies as well as several unpublished studies, indicate that the differences in skill between OWIlG and OWI3G, at least for an integrated property of the spectrum such as significant wave height, HS or peal spectral period, TP, are very slight and subtle despite the large differences in model formulation and computing requirements. However, for the severe **slantingfetch" wave regime of the study target areas in the Red Sea we would :expect the 3G physics to yield more accurate hindcasts. InDut. OWI3G is driven by the time history of surface wind speed and direction (1-hour average at 20 m) at 30-minute intervals. The duration is long enough to ensure that the wave (and current) response is fully spun-up (i-e. lost memory of the cold model start) before peak conditions are specified in all areas of interest in the basin, and also spun-up in source regions of swell which eventually propagate to areas of interest. Output. There are three output files of a hindcast run.

7 . 1

. .

This is a printable-only file-of summary output at selected grid points (1-line per point 'per time step) at either 30-minute or hourly intervals. This file is used mainly used to print results at a few points for quicklook and checking purposes.
2. Fields File. This is a permanent archive digital file of

1. 1-line Summarv Print File.

a

**fields** variables. Basically this file included points and the following variables: mO ml m2
HmO

all grid

TP
VMD

zeroth moment of spectrum (m**2) first moment of spectrum (m**radians sec**-2) second moment of spectrum (m**2 rad**2 sec**-2) significant wave height as 4 sqrt mO (m) peak spectral period, sec vector mean wave direction, degrees toward which
29

Dom

SPD

waves are travelling, clockwise from true North dominant wave direction: the direction that maximizes the mean value of cos(theta-dom)**2 degrees toward which waves are travelling, clockwise . from North angular spreading, as computed by Gumbel, Greenwood and Durand. See Pearson and Hartley, Bimetrika Tabels for Statisticians, vol. 2, pp. 124, 125, 130, 371. The exponent N in a cosine**N(theta) spreading form can be computed as N = 2.*spd/(l-spd); for example, spd = 0.87 yields N = 13.38.

This file contains the full two-dimensional spectrum at hourly intervals. Since this is by far the most voluminous data file, it is typically restricted to a relatively small subset of points. The points archived for this study are listed in Table 4.1. The spectra are given as variance components, m**2, in 24 angular bins (15 degree bandwidth) and for the frequency bins given in Table 4.2. The archive files are written as follows within the fractional growth and propagation steps which comprise the numerical solution : 15 minutes growth 30 minutes propagation 15 minutes growth intermediate print of 1-line summary 15 minutes growth 3 0 minutes propagation 15 minutes growth print 1-line summary archive fields archive spectra

3. Swectra.

Prior Validation The OW1 family of wave models have undergone extensive previous validation. For example in the original ODGP program, OWIlG hindcasts of significant wave height and spectra were compared t o measurements acquired at offshore platforms in five Gulf of Mexico hurricanes, including hurricane Camille. Since then, the ..model was applied to every major storm to affect the U.S. Gulf and East Coast coasts including Eloise and Belle (Cardone and ROSS, 1979). Hindcast and measured directional spectra were evaluated in Eloise and Delia by Forristall et. a1 (1978, 1980). The model skill in specification of peak significant wave height and associated peak period at a site in a storm was evaluated by Reece and Cardone (1982) for hindcasts made in four different types of -meteorological systems (including tropical and extratropical systems) in three different basins. In over 60 individual comparisons in 19 different storms, the model hindcasts exhibited negligible bias (less than 0.5 m in Hs), and r s errors of less than 1 meter in height and 1 second in peak m frequency (TP).
-A popular index

,A %I*

..
I .

of skill used by wave modelers is the scatter index, which is defined as the rms error (or alternatively by the standard deviation of the model-measured differences) divided by the mean of the data in the measurement data set used for the validation, expressed in percent. For the data of Reece and Cardone the scatter index in HS is 11.8%. An updated version of OWIlG, used since 1983 in most of Oceanweather's hindcast studies, provides essentially the same skill in specification of storm peaks as the original ODGP version, with slightly greater skill in specification of time histories. The model has been shown, with a stable calibration (unlike many models which are retuned with each application) to consistently provide a scatter index in storm peak HS and TP in the range of 10-15%, wherever applied, as long as wind fields are accurately specified. For
31

example, Swail et. a1 (1991) report a scatter index of 12.1% in HS and 11.3% in TP for OWIlG hindcasts of severe storms off the west coast of Canada. Cardone and m a n s (1992) report scatter index of only 9%'-in HS and TP in ODGP2 hindcasts of severe storms off the west coast of New Zealand. Comparable skill has been demonstrated in major validation studies conducted in recent years as part of proprietary industry sponsored studies of tropical cyclone generated waves in the Gulf of Mexico, the Taiwan Straits, and parts of the South China Sea. OWI3G, and the WAMDI model upon which it is based, have also been applied in a number of recent model validation studies. In WI (1988), for example, the WAM model was tested against hurricanes Camille (1969), Anita (1975) and Frederic (1979) and was found to be essentially as skillful as OWIlG in specification of HS, TP and frequency spectra. OWI3G was also applied to Camille and Frederic (these cases formed part of the model tuning data set) and ultimately achieved skill comparable to WAM and OWIlG. OWI3G has been recently applied to the two extreme US East Coast storms (Cardone et al., 1994b). The scatter index in storm peak HS for OWI3G averaged over 9 buoys was 14% in one storm and 10% in the second storm. The corresponding statistics for OWIlG, also applied in that study, were 14% and 15% respectively. Khandekar et al., 1994 investigated skill in OWIlG and OWI3G (as adapted in the Canadian CSOWM wave model) in a much more gentle wave regime over a two week period and find the two models to be skillful and virtually comparable in performance with rms errors in time series of HS of . 6 m and in TP of about 2 seconds. The scatter index is 19%-in HS, higher than reported above for peak-peak comparisons because the mean of the measured data for time series (about 3 m) is less than half the level of data of storm peaks. OWI3G has recently been applied in a major offshore industry Joint Industry Project to develop definitive wind and wave criteria throughout the South China Sea (SEAMOS) (Cardone and
32

0

.

Grant, 1994). This study also included a substantial validation activity involving both cyclonic storm regimes and northerly and southerly monsoon surge regimes. The validation indicated that OWI3G provides skillful hindcasts of wave height and period. The mean differences in wave height are typically 0.5 m or less and the standard deviation of the differences are 0.6 m or less. The mean and rms differences for this period are generally less than - 1 second. For event peaks the mean difference between hindcast and measurements was only 5 cm. The scatter index for event peaks is 18% which is remarkable considering that the mean storm peak wave height is only 2.85 m. Based upon the strength of the prior validation study, OWI3G is expected to provide unbiased hindcasts of wave height and period in Red Sea storms and to achieve a scatter index in the range of 10-158 in specification of storm peak wave height and period and about 20% in specification of continuous time series. unfortunately there are no suitable instrumental wave measurements available in the Red Sea events hindcast t o allow further validation in this study.

0

33

0

5.

STORM BINDCASTS

A total of 30 storms were hindcast'. Table 5.1 gives the dates of the storms and the duration. The typical duration was 4-5 days. Plots at 12-hourly intervals of the wind fields in each storm hindcast are provided under separate cover. A concise summary of the storm hindcasts is given in Table 5.2. Table 5.2 gives for each storm, the grid point which exhibited the maximum HS in the event, the maximum HS and associated period and vector mean direction and the associated wind speed. For example, storm # 16 (December 16 22, 1985) provided the highest hindcast HS, in this case 4.93 m at grid point 667, which is located at 22.25N, 38.5 W, in deep water offshore Rabeigh and Jeddah. The associated peak period is 9.19 sec and the associated wind speed is 20.25 mfs, though absolute maximum winds of about 22 m/s occurred at an earlier time step a this point. The associated wave direction is toward 347 degrees, or toward the northwest. This event happens to also provide the highest ranked southeasterly (from which) event offshore Jeddah and Rabeigh, at least in deep water.

-

Of the 30 storms hindcast, the absolute basin-wide storm maximum sea state propagates toward the southeast in 16 storms and toward the northwest in 14 storms. As expected, in most of the former cases, the maximum is found in the northern half of the basin, while in most of the latter cases the maximum is found in the lower half of the basin (storm 16 is an exception in this regard). In the center of the basin absolute storm maxima are specified for both types. For example, Table 5.3 gives the series of absolute hindcast maxima at grid point 643, which is in deep water just offshore Jeddah at 21.75N, 38.753. The highest

I an additional four storms were hindcast after the completion of the base study; these are discussed briefly in Appendix F.

34

4
I.)

ranked event, event 19, gives peak HS of 4.57 m and is a northwest type. The second highest ranked event gives peak HS of 4.22 m and is a southeaster. All but two of the other 28 events are northwesterly type. However, since many events hindcast actually contain both storm types in the center of the basin, it is possible to add another five cases of southeasterly events at this point by taking a secondary maxima in the southeast phase of t h e event. Several of the "additional" hindcasts also provided additional southeaster peaks in the central Red Sea (Appendix F). In general, for other directional sectors, the hindcasts do not provide absolute or secondary maxima, because of the relatively . low energy in other sectors and the small number of storms hindcast.. However, even f o r those sectors, relative sector maximum may be extracted if required from the full time series of hindcast sea states at each point in each storm or alternatively from the continuous hindcasts which naturally include cases of wave generation from all possible directions.

a

35

0

6.

CONTINUOUS 3-YEAR HINDCASTS

The 3-year continuous hindcast was produced using essentially the same hindcast method as used for the storm events. The only exception was that the wind fields were produced on the wind workstation at 12-hour intervals and then interpolated t o 3hourly intervals within the workstation. The 12-hour interval was selected because it is the interval at which background winds are available, namely 00 UT and 12 UT, and is the smallest time step at which the manual intervention could be reasonably tolerated. These fortuitously represent the sea breeze cycle quite well with 12 UT near or at the onshore flow maximum phase of the sea breeze cycle and 00 U T at or near the nighttime minimum phase. The assimilation of coastal wind measurements and the analyst provided kinematic control points at these analysis steps further incorporated the sea breeze effect along the Saudi coast. In the selection of the continuous years to hindcast, both data coverage and representativeness of the climate were considered. 1989 could be considered First of all, only years between 1980 because the background wind fields were available only for the decade of the 1980's. Table 6.1 shows the stations available for the decade. Based on this consideration alone, 1980 and 1981 were considered unsuitable because of the number of land stations available. Table 6.2 shows the number of ship reports available in the decade and the number of reports available at 00 UT and 12 UT. These tables together show that the best coverage of land 1989 with an average of about 8data is during the period 1986 10 observatiohs per day (only 00 UT and 12 UT) available, while the number density of ship reports in the same period is only about 10% less than the decade average. Given the importance of 1988 was selected. Table 6.3 the land data, the period 1986 shows the number of storms identified in the northern and southern areas and the number of events actually hindcast. The average number of events in the decade is 29.6/year, while the

-

-

-

36

average number of events in the years 1986-1989 is 22.3/year. The average number of events hindcast in the decade is 1.9 events per year while the number which occurred during 1986 1989 is 2fyear. We conclude that the 3-year period selected represents the storm climate reasonably well.

-

The 3-year wave hindcast produced exactly the same archive files -as produced for the storms. Appendix C gives check plots of the hindcast at three selected grid points in terms of the time histories of wind speed and direction, significant wave height, peak spectral period and vector mean wave direction.
I

c

-

37

REFERENCES

Cardone, V. J., 1969. Specification of the wind field distribution in the marine boundary layer for wave Report TR-69-1, Geophys. Sci. Lab.. Available forecasting

.

from NTIS AD# 702-490. Cardone, V. J., W. J. Pierson, and E G. Ward, 1976. Hindcasting . the directional spectra of hurricane generated waves. J. of Petrol. Technol., 28, 385-395. Cardone, V. J. and D. B. ROSS, 1979. State-of-the-art wave prediction methods and data requirements. Ocean Wave Climate ed. M. D. Earle and A. Malahoff. Plenum Publishing Corp.,
1979, 61-91.

Cardone, V. J., Greenwood, storm wind Tech., 32,

A. J. Broccoli, C. V. Greenwood and J. A.

1980. Error characteristics of extratropical

fields specified from historical data. J- Petrol. 873-880.

Cardone, V. J. and D. Szabo, 1985. Impact of uncertainty in specification of offshore wind on accuracy of wave hindcasts and forecasts. Proc. Int. Workshop on Offshore Winds and Icing, Nova Scotia, 7-11 October, 1985. Publ. Environment Canada, Downsview, Ontario.
-. Cardone, V. J. and J. A. Greenwood, 1987. A sensitivity study of spectral growth algorithm. Proc. of Int'l Workshop on Wave Hindcasting and Forecasting, 23-26 September, 1986 (ESRF Report 65), 133-143

Cardone, V. J., J. G. Greenwood and M. A. Cane, 1990. On trends in historical marine wind data. J. of Climate. 3, 113-127.

38

Cardone, V. J., 1991: The LEWEX wind fields and baseline hindcast. In: & , R. C. Beal, Ed., The Johns Hopkins University Press, 136-146. Cardone, V. J. and K. C. Ewans. 1992. Validation of the hindcast approach to the specification of wave conditions at the Maui location off the west coast of New Zealand. Preprints of the Third International Workshop on Wave Hindcasting and Forecasting, May 19-22, 1992, Montreal, Quebec, 232-247. Cardone, V. J., R. E. Jensen, D. T. Resio, V. R. Swail and A. T. Cox, 1994b. Evaluation of Contemporary Ocean Wave Models in Rare Extreme Events: "Halloween Storm" of October, 1991; "Storm of the Century" of March, 1993. Submitted to J. of Atmos. and Oceanic Tech. Cardone, V. J., H. C. Graber, R. E. Sensen, S. Hasselmann and M. J. Caruso, 1994a: In search of the true surface wind field during SWADE IOP-1: ocean wave modelling perspective. Submitted to The AtmosDhere-Ocean system Cox, A . T . , J. A. Greenwood, V. J. Cardone and V. R. Swail, 1995. An interactive objective kinematic analysis system. Fourth International Workshop on Wave Hindcasting and Forecasting; October 16-20, 1995, Banff, Alberta, Canada.

.... ..

~

0

Eid, B. M., V. J. Cardone, J. A. Greenwood, and J. Saunders, 1986. "Real-time spectral wave forecasting model test during CASP". -ESRF Report 065. 1987. Proceedings of the International Workshop on Wave Hindcasting and Forecasting, September 23-26, Halifax, NS, 183-195. Forristall, G. Z., E. G. Ward, V. J. Cardone, and L. E. Borgman. 1978. The directional spectra and kinematics of

39

0
...

surface waves in Tropical Storm Delia. J. of Phvs. Oceanou., 8, 888-909. Forristall, G. 2 ; : 1980. A two-layer model for hurricane driven currents on an irregular grid. J. Phvs. Oceanoq., 10, 9, 1417-1438. Forristall, G. Z., E. G. Ward and V. J. Cardone., 1980. Directional spectra and wave kinematics in hurricanes Carmen and Eloise. 17th International Conference on Coastal Engineering, Sydney, Australia. Greenwood, J. A., V. J. Cardone, and L. M. Lawson, 1985. Intercomparison test version of the SAIL model. Ocean Wave Modelinq. The SWAMP Group, Plenum Press, 221-234. Hasselmann, S. and X. Hasselmann, 1985. Computations and parameterizations of the nonlinear energy transfer in gravity wave spectrum. Part I: A new method for efficient computations of the exact nonlinear transfer integral. JPhvs. Oceanoqr., 12, 1369-1377. Khandekar, M. L., R. L. Lalbeharry and V. J. Cardone, 1994. The performance of the Canadian Spectral Ocean Wave Model (CSOWM) during the Grand Banks ERSl SAR Wave Spectra Validation Experiment. Atmosvhere-Ocean, 32, 31-60. Koch, S. and H. Shabib, 1989. Design'criteria for Yanbu Expansion Facility. Exxon Production Research Company. Offshore Division. EPR.9PS.89. Production Research Application Report. Komen, G. J., S. Hasselmann and K. Hasselmann, 1985. "On the existence of a fully developed windsea spectrum". J. Phvs. Oceanoa.,lQ, 1271-1285. 40

0

Large, W.G. and S. Pond, 1981. Open ocean momentum flux measurements in moderate to strong winds. J. Phvs. Oceanoa., 11, No. 3, 324-336. Morcos S. A., 1971. Physical and chemical oceanography of the Red Sea. Mar. Biol. An. Rev., 8, 73-202. Murty, T. S. and M. I. El-Sabh, 1984. Weather systems, storm surges and sea state in the Red Sea and Gulf of Aden. Proc. Symp. Coral Reff Environ, Red Sea. Jeddah, Jan, 1984, 8-38. Reece, A. M. and V. J. Cardone. 1982. Test of wave hindcast model results against measurements during four different meteorological systems. Offshore Technology Conference. . OTC 4323, 269-. Resio, D.T., 1981. The estimation of wind wave generation in a discrete spectral model. J. of Phvs. Oceanoa., ll(4).

..
*.

Ross, D. B. and V. J. Cardone, 1978. A comparison of parametric -and spectral hurricane wave prediction products. Turbulent Fluxes throuah the Sea Surface. Wave Dvnamics. and Prediction, A. Favre and K. Hasselmann, editors, 647-665.
Snyder, R. L., F. W. Dobson, J. A. Elliot and R.B. Long, 1981. Array measurements of atmospheric pressure fluctuations above surface gravity waves. J. Fluid Mech., 102, 1-59. Swail, V. R.; V. J. Cardone, and B. Eid., 1991. An extremes wind and wave hindcast off the west coast of Canada. (Preprint Volume) Fifth Conference on Meteorology and Oceanography of the Coastal Zone, 6-9 May, 1982, 269-281. WAMDI Group, 1988. The WAM model-a third generation ocean wave prediction model. , 18, 1275-1810. 41

0

Whittaker, E. T. and G. Robinson, 1944. Calculus of Observations. Glasgow, 254-255.

Wu, J. 1982. Wind-stress coefficients over the sea surface from breeze to hurricane. J. Geouhvs. Res., 87, 9704-9706.

42

Table 2.1

Land weather obser i stations surrounding the Red g ea for which observations are c h i v e d at the NOAA aational Climatic Data Center, sheville, NC.

STATION I

NAUE

LATITUDE

LONGITUDE
(- ' )

COUHTRy

ELEVATION

1.

')

(W

PERIOD OF RECORD

I OF OBSERVATIONS

404390 405000 410240 411400
*

YANBU AL BAHR KAMARAN

24 09 15 20 21 40

3 0 04

SD

1
6
12
I

1973-09

42 37 39 09 42 35

YE SD

.

30224

JEDOAH A0DUL AZIE QIZAN

1901-90 1913-90
'

26440 39664

h

16 54

SD

3

11

114610
~

1

ADEN INTI. AP

1

12 50

I

4 5 02

I

DY

I

3

I

1903 09

I
L

11 II

624590 624001

I
1 1

EL TOR

1
I
~

20 1 4 27 40

I
I
I

33 31

I
I I

EO
EO

1
I I

.

I

7000

11

2

I
I
I

RAS JIHSAll

33 3 5

6

w

11

624650 624101 624711 626410 626750 630230

KOSSElR
WADI CIHAL DAEDALUS

I

26 00 24 39 24 55 19 35

34 10 35 10

EG

7

EO
E7 l

3

.
f

I

.
11493
64

II

35 52
37 13

4
2 2
10

PORT SUDAN
AQlQ

su
SU

1973-90 1914-90 1973-90

10 14
15 31

30 11
39 27

NASSAWA

ET

N

630430

I

ASSA0

1

1301

I

4243

I

ET

I

.

9

-1

.

11432

I

. llmlted data
C o u n t r y Codes:

I
SO SU

- Egypt Dr - remen YE - Yemen DJ - Djlboutl
EO

!

-

ET

-

Saudl Arabia
Sudan

Ethiopla

Table 3.1 Summary of Wind Regimes over the Red Sea with wind strength shown in Beaufort Force (from Murty and El-Sabh, 1984) Month January to February Normal Wind Regime North of 20°N, winds from N and NW Force 4 South of 16aN, somewhat stronger winds from S and SE

March (transition) North: N and NW Force 4 as far south as 18ON with Khamsin dust storms. South of 16ON, Apr i1 S to SE winds prevail. Between 18ON and 16ON, winds are variable. Gales are rare, but more frequent in the south on S to SE winds. North: N and NW Force 3-4. South of 15ON, easterly squalls may occur June (transition) July t o August September (transition) October

N and NW winds entire length of Red Sea Force 4 in the north, 3 in the south. Winds more variable in the south. Gales rare

North of 20°N, winds N and NW Force 3-4. Winds variable between Z O O and lao. South of 18ON winds from S and SE Force 3. Cold frontal squalls occur in the north. Gales are rare.
As October except S-SE winds in south more steady. Depressions cause squalls and thunderstorms, sometimes violent in transition zone between NW and SE winds between laaN and 20"N.

November to December (transition)

44

Table 3.2

Sample of processed file of synoptic ship reports: f8yymmddhh*t year, month day and hour of is obseryations for ship with call sign "shipid" at indicated l l t l and IIlonga with anemometer at "ht" lal in meters if wind is measured (tlty" 2 or 4 ) , = otherwise the wind speed reported is a Beaufort estimate. All wind speeds are adjusted to 2 0 m neutral stratification depending on **ty",lfhtl@ and air-sea temperature difference (*ltdifs8)

.

n d i

n

adj

press

dry surf rdif shipid hr

w

d w

I hw

dsu

I hsu

1c8

... .. - ~.-

- ... -- 0
I .

1890 06811 2 2 890 06813 2 3 927 06823 2 4 926 06827 2 5 e 9 06857 2 a 6 926 06838 2 7 926 068L2 .2 8 926 06849 2 9 926 06861 2 10 890 06870 2

11.9 11.8 12.4 12.3 1. 31 13.8 14.9 14.3 1. 61 17.0

41.7 43.9 43.9 47.9 47.9
48.5

42.1 49.8 41.4 40.5

~

0

1 7 890 0 9 & 18 890 06851 1 890 06851 9 20 890 06870 21 926 10517 22 890 10517 23 926 10518 24 890 10527 25 926 10546 26 890 10555 27 926 10573 28 aqo 10574 29 ea8 06842 30 926 06823 31 926 06824 32 926 Ob833 33 888 06836 34 927 06842 35 890 06842 36 890 06842 37 avo 06861 38 890 06917 39 926 10508 40 890 10518 4 926 10536 1 42 890 10546 43 926 10555 44 890 10555 45 926 105G 46 926 06824 47 V26 06824 LE BBP 06826 49 avo 06828 50 926 06835 51 a90 06842 52 927 06851 53 926 06989 54 a90 o m 8 55 926 10555 56 926 10555

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

43.0 U.9 42.6 42.5 42.2 41.3 1 . 37.1 17 20.0 38.7 21.0 38.1 23.4 36.9 2 . . 36.3 41 25.5 35.7 25.6 35.2 26.4 34.7 1. 2 5 44.1 12.6 44.9 12.8 46.8 12.6 48.5 13.2 43.6 14.7 42.9 15.2 41.5 1. 86 3 . 96 1. 38.9 97 2 25.1 35.9 2 2. 52 3. 55

14.8 1. 55 15.7 17.4 21.4 21.5 21.7 22.9 24.4 25.8 27.6 27.2 YL.2 12.5 12.5 13.3 13.1 14.4 14.5 14.7 16.2

4q.q 41.7 41.4 40.4 37.8 37.3 38.0 37.1 36.6 35.5 33.9 34.3 42.4 43.5
U.4

81020WO 81020900 81020900 81020900 81020900 8102WOO 81020900 81020900 81020900 81020900 81020WO 81020900 81020900 81020906 81020906 81020906 81020906 81020906 81020906 81020906 81020906 81020906 81020906 81020906 81020906 81020906 81020906 81020906 81020909 81020912 81020912 81020912 81020912 81020912 81020912 81020912 81020912 81020912 8 020912 1 81020912 81020912 81020912 81020912 81020912 8?OZOPl2 81020918 81020918 81020918 8102091~1 81020918 81020918 81020918 81020918 81020918 8102Wl8 81020918

110 0 018 18.5 1001.5 100 0 014 15.6 1013.0 090 0 023 22.5 1013.8 060 91 016 1. 1013.6 070 0 016 16.6 1014.2 050 009 10.1 1015.8 160. 027 28.0 1008.0 100 009 12.9 1014.5 140 014 17.5 1007.8 140 0 008 7.7 1010.1 330 014 O W 350 330 0 020 06 1 070 160 030 010 0 021 15.5 1016.1 060 0 010 11.6 1015.2 13 140 0 021 2 . 1011.4 0 0 0 1009.3 . 120 0 008 9.0 1012.1 017 20.8 1014.5 330 .~ 330 0 017 19.0 1021.1 028 30.2 1012.7 330 320 014 15.9 1024.2 330 0 020 21.6 1015.1 . 320 o 010 7 3 1023.9 300 009 . 1 3 . 2 1018.0 ~. 300 0 019 20.5 1017.4 0 0 1010.2 . 120 019 21.5 1010.5 100 016 16.7 1011.8 150 024 25.4 1008.2 070 0 010 8.5 1018.8 190 0 021 13.5 1010.8 170 0 021 17.7 1009.3 170 0 01 6 13.9 1008.6 1 40 0 023 22.3 1009.3 010 0 012 13.2 1012.1 018 21.6 1012.1 350 320 0 017 18.9 1013.5 028 30.5 1012.3 330 014 1 . 1024.8 320 41 330 0 020 21.6 1015.7 330 0 008 . 5 . 0 1017.1 340 o w 8.4 1016.5 140 018 21 .o 1012.4 090 018 21.1 1013.9 070 0 013 13.7 1015.6 090 0 017 16.3 1013.0 150 019 21.7 1012.0 170 0 021 2 . 1008.6 01 080 0 020 17.3 1013.6 340 017 20.9 1013.2 340 0 019 21.2 1014.2 330 061 015 19.4 1 1 . Y O 009 12.9 1017.7
~ ~~

2. 50 2. 40 2. 50 24.6 23.3 26.4 25.0 24.1 25.2 2. 52

26.0 29.0 25.0 25.3 24.4 25.4 25.0 26.6 26.5 25.0

-1.0 SHIP -5.0 SHIP 00 5 M L . -0.7 -1.1 NCGP 10 . 00 . -2.1 -1.3 0.2 UXM

.

~

.. .

28.0 24.4 36 5 M l . 20.0 26.0 -6.0 UZXY 26.0 26.0 0 0 UKW . 28.5 27.3 1 2 UYXK . 26.2 28.2 -2.0 Ulll 20.6 24.0 -3.4 16.2 24.0 -7.8 S H I P 21.8 26.0 -4.2 23.0 27.0 .L.O W C 19.8 25.0 -5.2 3.0 WD 28.0 25.0 15.8 18.8 -3.0 16.2 19.0 - 2 . 6 M I J 26.4 25.0 1 4 FNLU . 25.7 26.0 -0.3 2 . 26.8 85 1.7 27.0 26.8 0.2 0 8 EWG8 . 26.8 26.0 30.0 24.4 5 6 5YYL . 27.0 23.0 4.0 UKW 27.5 25.5 2.0 UYXK 28.2 27.3 0.9 UIII 28.0 30.0 -2.0 WLK 22.9 26.0 -3.1 16.3 23.0 -6.7 SHIP 20.3 26.0 -5.7 25.0 24.0 10 W P C . 19.6 24.4 -4.8 2 . 20.0 21 21 UOlJ . 21.4 19.4 20 . 25.2 26.3 -1.1 24.2 25.5 -1.3 23.925.0 .1.1 NCGP 26.0 25.0 1.0 UYEN 25.5 26.2 .0.7 27.0 25.6 1.4 UYXK 26.0 24.4 1 6 5WIL . 24.5 28.0 -3.5 16.2 25.0 -8.8 SHIP 19.0 24.0 -5.0 17.6 19.4 -1.8
~ ~ ~~~

3 3 3 4 3 4 4 4 4 1 2 7 20 4 20 4 20 3 20 4 20 4 26 3 14 3 17 2 17 2 19 2 20 4 20 3 20 4 10 2 20 3 11 2 20 4 14 3 20 4 20 4 20 4 20 4 18 3 26 3 17 2 17 2 19 2 14 3 20 4 20 3 20 4 10 2 20 3 14 3 20 4 20 4 20 4 20 3 1 2 6 20 L 17 2 26 3 20 4 20 3 20 4 20 4

20 20 26 20 20 20 20 20 20

060 2 15 . 070 050 2 0.5
~

wo

060

5 20 .

090 5

-

25 . 21 . 14 . 22 .

20 . 10 . 20 .

100 2 1 0 . 140 2 1 0 . 330 350 330 070 160 2 2 2 2 2
1.5 10 . 3.0 10 . 00 .

wo

120 5

om

. 330 5 2 0 5 20 . . 160 6 2 5 3 10 . 00 . 0.0

36 . 22 . 2.5. 10 . 2.5 1 .o
1.1 10 . 1.5 3.6 0.0

060 2 0 0 .

050

140 2 2 5 . 1.0 330 330 330 320 330 320 300 300 120 100 150 070 190 2 05 . 10 . 3 35 . 3 1.5 2 30 . . 2 00 2 10 . 1.5 3 2 1.0 2 1.5 2 2.5 2 0.5 4 25 . 3 25 . 1.5 2.0 2 1.0
1 .o

000 000
150 . .

.

--

L O ...

340 5 000

.

10 . 00 .

000 0 0.0 330 5 2.0 000 00 .
000
000

. .
-

0.0

15 . 2.2 21 . 3.2 05 . 2.5 1.5 36 .
1 .o

1 40 6

20 .
1.5

OM
990 000
000 000 150 000

5

2.0 0.0
00 . 00 .

i n

170 140 350 320 330 - 320 330 330

. .
.

30 . 0.0

3 35 . 3 1.5 2 30 . 05 . 340 2 1 0 . 140 2 1.5
or0 090 1.5 150 2 1.5

000 0 0.0 330 5 2.0 000 - 0 0 .

1.5 3.6
0.5

990

2.0

2.5

wo
wo
000

340 1.5 330 2 2 0 . 340 2 1.5

.

00 .

15 . 21 .

340 5 1 5 .

45

Table 4.1

Grid points of basin-wide wave model grid at which full directional spectra were archived in all hindcasts.

-0

522

38.75 38.75 38.25 38.15 38.75 38.75 37.25 37.15 38.00 37.75 36.50

19.75 21.25 21.50 21.75 22.50
~

619
629
~

e
~

643 682 698 143 745 146

22.75 23.50 23.50 23.50 24.00 24.75

7 15
810

II II II 11 . 1 I

~~

812 863 671

I I I I I

37.00 36.25 35.50

I I I I I

24.75 25.75 26.00

II 1 I 1 I 1 I II

945 946
~ ~~

35.50 35.15 34.50 34.50 34.50 34.75

27.25 27.25 27.50 28.00 28.25 29.00

0

952 974 987 1019

46

R:\CHCIUIfl.TAB. pMAY-199S IStN

Table 4.2

The 23 frequency bins of the third-generation wave mode1,used for the basin-wide wave hindcasts.

Bin

(Nominal Frequency

1

Lower Limit

I

upper L i m i t

I

Bandwidth

1

*r ,

..,

0.0969849

0.0088683 0.0097608

/I
11

~~~~

11 12
~~

I
I

0.1011483 0.1119885
~

0.0969849

0.1067457 0.1174888 0.1293131

I

0.1067457 0.1174888

I I

0.0107431 0.0118244

1
11

13

0.1232593

15 16 17 18 19 20 21

0.1493180 0.1643457 0.1808858 0.1990906 0.2191276 0.2411811 0.2654541

I

0.1423275 0.1566517 0.1724175 0.1897700 0.2088690 0.2298900 0.2530267

1

0.1566517 0.1724175 0.1897700 0.2088690 0.2298900 0.2530267 0.2784919

0.0143242 0.0157658 0,0173525 0.0190989 0.0210211 0.0231367 0.0254652

11

21~

I

0.2921701 0.3215748

1

0.2784919 0.3065200

I

0.3065200 0.3373689

1

0.0280281 0.0308489

11

23

47

Table 5.1

Dates of periods hindcast for 30 storms.

e

e

I t

21 22 23 24 25 26

I

I I
.

8905 8911 9101

I I

0712-1112 2712-0418 2300-2712

I
II

1

I

9103 9111 9201 9202 9302 9311

I

2100-2518 2800-0100 0112-0518 2212-2712 0112-0518 2012-2418 1200-1618

I

e

27
28

29
30

9401

48

.

,..

.

..

... ..

e

/

Table 5.2

Absolute storm peak Significant wave height (HS) and associated peak wave period (TP), vector mean wave direction (VMD), wind speed (WS) (m/s at 20 m ht) over the entire basin-wide wave model grid.

49
R.'ICHCKLIST.TAB, 9-MAY-1995 15:24

Table 5 . 3

Absolute storm peak significant wave height (HS) and associated peak wave period (TP), vector mean wave direction (VMD), wind speed (WS) (m/s at 2 0 m ht) at grid point 6 4 3 , in deep water offshore Jeddah.

a

77012512-012900 77031000-030800 78120900-121218 79051500-051818 80031000-031418 80041112-041600 80060412-060812 80100200-100518

I

12.66 11.88 9.47 14.40 15.30 12.04 12.44 10.48

I

194.8 326.7 335.0 311.0 325.0 339.0 323.0 323.0

I
2.52

7.52 7.13 9.46 7.89 8.11 6.78 7.86

137: 20 144.20 132.54 139.38 139.90 139.82 133.40

1.76 4.02 3.21 2.76 2.20 2.28

81122000-122418 82032400-032918 83022312-022700
~~

9.38 10.86 12.00 9.31 14.00 10.50 19.56 12.50 14.34 16.86

147.6 325.0 329.0 336.0 325.0 347.0 163.0 319.0 321.7
330.0

1.43 2.17 2.48 1.87 3.72 2.26 4.22 2.38 2.46 4.57

7.42 7.53 7.54 8.06 9.61 7.86 8.66 7.13 1.57 9.72
.

344.12 134.75 140.67 168.44 139.72 140.50 348.50 131.01 134.58 140.95

0

84022000-022518 84032000-032818 85113012-120418 85121600-122000 86061200-061518 87020600-021018 87041600-040212

94011200-011618

I

13.06

I

330.0

1

2.79

I

8.00

I

141.62

I

50

Table 6.1

Availability of synoptic data from stations along the east coast of the Red Sea during the decade of the 1980's in archive files obtained from N O M .
> t i

UI

v

I

Yanbu

Qlzan Yanbu

Yanbu

Qlzan Yanbu

I

Qiran Yanbu

I

Qiran Yanbu

Qizan Yanbu

Qlzen Yanbu

a
Table 6.2 Number of synoptic land station reports in the Red Sea during the decade of the 1980'in archive files obtained from NQAA, for synoptic times 0000 UT and 1200 UT only.

Year

No. of Reports at
2364 924

0 0 2 and 122

1980 1981

I F

1982 1983 1984 1985 1986

I

2780 3039 2055 2609 3137

I

I

1988

I

3596

I

a

52

. .

Table 6.3

Number of synoptic ship reports in the Red Sea during the decade of the 1980'in archive files obtained from N O M , for all synoptic times and 0000 UT and 1200UT on1y

.

--.

w.

..

Year

A l l Reports

Reported at 002 and 122

1980

25,000 23,300 21,200

11,500
10,400

0
~

1981 1982 1983 1984 1985 1986 1987
I

10,100

24,000
23,000 16,000 18,000 17,900 21,000
15,000

7,600 8,600 7,400 7,800 7,300 9,500
6,400

.I

.

I

1988 1989

0

53

Table 6.4

Storm events identified in the decade of the 1980's in the llcoarsell list and the number actually hindcast.

Year

North

South

Total
22

Hindcast
4

1980

I

12

I

10

54

a

0

STATION LOCATIONS
30
-

WADIGIM~C

RAS BANAS

DAY k
YANBU AL BAHR

I

\

QlZAN KAMARAN

DJIBOUTI

0

iI n I V

I

I

-LJ--J
I

30

35

40 45 LONGITUDE

50

Figure 2.1. Locations of weather reporting stations bordering the Red Sea whose data are archived in part at the NOAA National Climatic Data Center, Asheville, NC.

55

Figure 4.1

Sample screen image of wind workstation (see also report front cover) used to generate wind fields. Winds plotted at 2.5 degree intervals are from background first guess winds fields, winds plotted at irregular locations within the Red Sea are ship reports, winds plotted along the coast are reports from land stations and winds plotted at 1-degree intervals are result of assimilation of data shown.

56

Figure 4.2

Screen image of wind workstation for the same sample field shown in Figure 4.1 showing the 1-degree winds and, in the interior of the Red Sea the analyst added winds (kinematic control points).

0

58

Figure 4.3

Final surface wind fields at 112 degree regular intervals representing assimilation of.all data and analy.st provided kinematic control points.

60

RED SEA GRID POINTS
30'

0

2' 8

0

0
3' 2 3' 3 3' 4 35' 3' 6 37' 38' 3' 9 40'
4' 1

42'

43'

-.

44'

45'

46'

Figure 4.4

Grid system of the basin-wide wave,model.

62

00

m

I

24'

-

23'

-

22' -

21' -

35'

36'

37'

38'

39'

40'

41 a

Figure 4 . 4 cont.

64

P P

RED SEA GRID POINT DEPTHS

,29'

28'

27'

26'

25' 32' 33' 34' 35' 36' 37' 38'

Figure 4.5

Depth field (m) of the basin-wide wave model.

67

RED SEA GRID POINT DEPTHS

a

a

Figure 4 . 5 cont.

68

.

. ..

-

m
d

+

+

+

e

t

+

+

f , l ,*
tt

O ' 0 ,

+

P ,

e

0 ,

'.E
e%+

t

0 ,

t

+t'$.t
0

%

t

%

+ *% t
t

4

et et
f.+
2 .

4 ,
*i
(Po

O1t

t t

' *t . % % t% t %t
4
+

+t
0 ,
0 ,

t

et
P t ,
0 ,

'Qt
. ,
t

'%t%tOk.+t "
* %
t t
0 ,

t t
+

4 ,
+

'e,t "'e,
a.
e.

*<t aa t
e.

+

t
I

4J E
0

($
t

w

m

% +4

0 ,

+

s;

0 ,

0
0 ,

<% * %

', .

+
t
1 .

t
t

'i% t
4 %

+

t

0 ,

B

t t

s+
+

%+
+

i

"be % .

t

/

I-

1 1

m
m

0

0

I

I

I

I

I

0
46'

40'

41 '

42'

43'

44'

45'

Figure 4 . 5 cont.

Appendix A.

0

Coordinates and water depth (m) of the basin-wide wave model grid. The columns, in order, are I coordinate, J coordinate, grid point serial number, latitude (degrees N), longitude (degrees E), land sea code, water depth (m).

45

46
47 48 49 50 51 52 53 49 50 51 52 53 48 49 50 51 52 53 45 .46 47 48 49 50
51

1 1

5
5

5
5 5 5 5 5 6 6 6 6 6 6 6 6 7 7 7
7

52 53 46 47 48 49 50
51

52 53 43 44 45
46

._

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 -. 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 42 43

10.0000 10.0000 10.0000 10.0000 10.0000 10.0000 10.0000 10.0000 10.0000 10.5000 10.5000 10.5000 10.5000 10.5000

43.0000 43.2500 43.5000 43.7500 44.0000 44.2500 44.5000 44.noo 45 .oooo 44.0000 44.2500 LL.5000

0.0000
0.0000 ... ...

0.0000
0.0000 ~ ~ ~ . . .

0.0000 0.0000 0.0000 0.0000

0.0000
0.0000

0.0000 0.0000
0.0000

0.0000 .....~ 0.0000

0.0000 0.0000

0.0000

44.7500
45.0000 43.noo

io.noo

10.7500 10.7500 10.7500

44.0000
44.2500 44.5000

io.noo io.noo

u.noo

41 ._

7 7 7 7
7

47 48 49 50
51 -.

44
45
A6

7 7
8

52 53

47

__

44
45 46 47 48 49 50 51 52 53 41 42 43

8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 9 9 9

48 49 50 51 52 53 54.

55
56 57 58 59 60 61 62 63

44
45 46 47 48 49 50 51 52

6L
65 M 67 68 69 70

45.0000 0.0000 43.0000 ~ 0.0000 11.oooo ~ . . ~ 0.0000 43.2500 0.0000 11.0000 0.0000 0.0000 43.5000 11.0000 ii.oooo 43.noo 360.0000 127.4444 44.0000 360.0000 534.0000 11.0000 44.2500 360.0000 801.1111 11.0000 44.5000 360.0000 1000.5555 11.0000 44.7500 360.0000 1253.3334 11.0000 45.0000 360.0000 1412.7?78 11.0000 43.2500 0.0000 0.0000 11.2500 43.5000 360.0000 21.7143 11.2500 43.7500 360.0000 233.8889 11.2500 44.0000 360.0000 547.0000 11.2500 44.2500 360.0000 741.3333 11.2500 41.5000 360.0000 944.4445 11.2500 44.7500 360.0000 1173.0000 11.2500 45.0000 360.0000 1412.8889 11.2500 42.5000 0.0000 0.0000 11.5000 i i . ~ o o o 42.noo o.oooo 0.0000 0.0000 43.0000 0.0000 11.5000 11.5000 L3.ZOO 360.0000 40.1429 97.1111 43.5000 360.0000 11.5000 43.7500 360.0000 267.4445 11.5000 44.0000 360.0000 487.1111 11.5000 44.2500 360.0000 685.5555 11.5000 11~5000 44.5000 360.0000 878.5555 11.5000 44.noo 36o.oooo 1093.8889 11.5000 45.0000 360.0000 1376.2222 i i . n o o 42.noo 36o.0000 IO~.LOOO i i . n o o 43.0000 360.0000 178.2500 i i . n o o 43.2500 360.0000 342.4445 i i . n o o 43.5000 360.0000 397.3333 i i . n o o 43.7500 360.0000 490.7778 i i . n o o LL.0000 360.0000 588.W8 i i . n o o 44.2500 360.0000 744.8889 i i . n o o u.5000 360.0000 833.8889 i i . n o o LL.7500 360.0000 984.11 11 45.0000 360.0000 1129.4445 11.7500 0.0000 0.0000 42.0000 12.0000 0.0000 0.0000 42.2500 12.0000 0.0000 0.0000 42.5000 12.0000 0.0000 0.0000 12.0000 42.7500 0.0000 0.0000 43.0000 12.0000 0.0000 ...... 0.0000 43.2500 12.0000 43.5000 360.0000 756.3333 12.0000 360.0000 8 9 7 . m 43.noo 12.0000 44.0000 360.0000 a35 .M67 12.0000 44.2500 360.0000 916.3333 12.0000 44.5000 360.0000 882.1111 12.0000 44.7500 360.0000 953.2222 12.0000
~~ ~~~~~~~ ~ ~~~~~

0.0000 0.0000 360.0000 294.5000 360.0000 510.5000 360.0000 644.6467 360.0000 384.2000 0.0000 0.0000 360.0000 239.0000 360.0000 R5.5555 360.0000 980.4445 360.0000 1042.3334 360.0000 1102,4445

.

. ..

9 10 10 10 10 10 10 10
1 0

53 66 47
La

7 1

72
73

11
11 11 11 11

49 50 51 52 53 45
46

74 75 7 6 77 . 78 79 80

ai

11 11 11 11 12 12 12 12 12 12 12 12 12 13 13 13 13 13 13 13 1 3 13 13 13 13 13 13 13 13 13 14 14 14 14 14 14 15 15 15 15 15 15 15
16

47 4a 49 50 51 52 53 45 46 47 4a 49 50 51 52 53 37 38 39 40 41 42 43 44 45 46 47 4a ' 49
50

a2 a3
84

85 e 4 a7 a8 a9 90 91 92 93 94
9s

96 97 9a
99

100 101 102 103 104 105 106 107

loa
109 110 111 112 113 114 115
116

51 52 53 42
43

16 16
16

16
1 6 1 6 1 6

17

17
17
17 17 17

44 45 46 47 41 42 43 44 45 46 47 40 41 42 43 44 45 46 47 33 34 35 36 37 38

117 iia 119 120 121 122 123 124 125 126 127 128 129 130
11 3

132 133 13L 135 136 137 138 139 140 141

45.0000 12.0000 43.2500 12.2500 43.5000 12.2500 12.2500 43.7500 &.OOOO l2:2500 12.2500 .. 44.2500 U.5000 12.2500 12.2500 44.7500 12.2500 45.0000 13.0000 . .. . .. 12.5000 43.2500 I2.5000 43.5000 12.5000 43.7500 12.5000 44.0000 12.5000 44.2500 12.5000 44.5000 12.5000 44.7500 12.5000 45.0000 12.5000 43.0000 12.7500 43.2500 12.7500 12.7500 43.5000 12.7500 w.7500 12.7500 U.0000 12.7500 44.2500 12.7500 U.5000 12.7500 U.7500 12.7500 45.0000 11.0000 13.0000 41.2500 13.0000 c1.5000 13.0000 13.0000 41.7500 42.0000 13.0000 42.2500 13.0000 42.5000 13.0000 42.7500 13.0000 43.0000 13.0000 43.2500 13.0000 43.5000 13.0000 13.0000 43.7500 44.0000 13.0000 13.0000 44.2500 13.0000 44.5000 13.0000 u.7500 13.0000 45.0000 13.2500 42.2500 13.2500 42.5000 13.2500 42.7500 13.2500 43.0000 43.2500 13.2500 43.5000 13.2500 42.0000 13.5000 42.2500 13.5000 42.5000 13.5000 13.5000 42.7500 43.0000 13.5000 13.5000 43.2500 13.5000 43.5000 13.7500 41.7500 13.7500 42.0000 13.7500 42.2500 i3.noo 42.5000 13.7500 42.7500 13.7500 43.0000 43.2500 13.7500 13.7500 43.5000 14.0000 40.0000 14.0000 40.2500 14.0000 40.5000 14.0000 40.7500 41.0000 14.0000 41.2500 14.0000

360.0000 0.0000 360.0000 360.0000 360.0000 360.0000 360.0000 36o.oooo 360.0000
0.0000 . .

997.w5 0.0000 100. 404.3333 485.2222 555.6667 830.6667 962.3333

na.ws
0.0000 ~ ~
~ ~

360.0000 20. 360.0000 100. 36o.oooo 60. 360.0000 40. 360.0000 50. .. .. 360.0000 60. 75. 360.0000 360.0000 500. 0.0000 0.0000 50. 360.0000 0.0000 0.0000
~

o.oooo

0.0000 0.0000

0.0000 0.0000 0.0000

360.0000 360.0000 360.0000 0.0000
.

50.6667 56.aooo 97.1250 0.oooo
0.0000

0.0000

0

j

0.0000 0.0000 0.0000 0.0000 0.0000

0.0000
0.0000 0.0000 0.0000 0.0000

360.0000 360.0000 360.0000
0.0000 0.0000 0.0000 0.0000 0.0000

17.1250 25. LO.
0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

o.oooo

0.0000 0.0000
360.0000

36o.oooo 360.0000
0.0000 0.0000 0.0000

2a.5000 70. 100.
0.0000 0.0000 0.0000

360.0000 360.0000 360.0000 360.0000 360.0000
0.0000
0.0000

12.0000 59. 100. 35.
10. 0.0000 0.0000

360.0000 360.0000 360.000

75 . 23.
99. 0.

o.oooo

360.0000 360.0000
0.0000 0.0000

46.778 7.5000 0.0000
0.0000

0.0000
0.0000

0 . 0 .
0.0000 0.0000 0.0000

0.0000 0.0000 0.0000

0

17 17 17 17 17 17 17 17 17 17 17 18 18
18 . . 1R ._

39 40 41 42 43 LL 45 L6 47

WI
49

38
39 40
Ll ..

18 18 18
18 ._

-.

..
-

.. ,

-

18 19 19 19 19
10 ..
'

:. -

e

19 19 19 19 19
10

42 43 44 45 . . 46 37 38 39 L O L1 .. 42 43
LL

142 143 14L 145 1 46 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165
~~~

166

45
L6
L7

20 20 20 20 20 -~ 20 20 20 20 20 20 21
~

.,

35
36 37 38
39

__

167 168 169 170 171 lT2 173 174 175 176 177

LO Ll 42 43 LL 45

in
179 180 181 182 183 18L 185 186 187 188 189 190 191 192 193 194 195 1 96 197 198 199 200 501 202 203 204 205 206 207 208 209 210 211 212

51
21 21 21 21 21 21 21 21 21
>l _.

56
31 32

29

33
34 35 36 37 38 39
L.O .

e

21 21 21 21 21 21 21 21 21 22 22 22 22 22 22 22 22

41 42 43 LL 45

46
47

48
49 31 32 33

3 &
35 36 37 38

0.0000 0.0000 41 .so00 14.0000 41 . n o 0 360.0000 30. 14.0000 42.0000 360.0000 50. 11.0000 60. 42.2500 360.0000 11.0000 60. 14.0000 42.5000 360.0000 0.0000 42.7500 0 . 14.woo 30.0000 14.0000 43.0000 360.0000 0.0000 0.0000 11.0000 43.2500 0.0000 0.0000 14.0000 43.5000 14.0000. . 43.7500 ~ 0.0000 0.0000 . ~ ~ ~ 0.0000 0.0000 14.0000 LL.0000 0.0000 0.0000 14.2500 41.2500 16.6250 14.2500 41.5000 360.0000 50. 14.2500 61 .zoo 360.0000 lL.2500 L2.0000 360.0000 65. 42.2500 ~6o.oooo 2 o i . u 14.2500 BO. 42.5000 360.0000 14.2500 86.3333 42.7500 360.0000 14.2500 15.8571 14.2500 43.0000 360.0000 0.0000 43.2500 0.0000 14.2500 0.0000 0.0000 41.0000 11.5000 14.5000 41.2500 360.0000 7.5 45. 41.5000 360.0000 14.5000 65. 14.5000 41.7500 360.0000 60. 14.5000 42.0000 360.0000 14.5000 42.2500 360.0000 327.5555 14.5000 42.5000 360.0004 250. 14.5000 42.7500 360.0000 55.6467 14.5000 43.0000 360.0000 7.5 14.5000 43.2500 0.0000 0.0000 14.5000 43.5000 0.0000 0.0000 14.7500 40.5000 0.0000 0.0000 11.7500 40.7500 360.0000 . 7.5000 14.7500 41.0000 360.0000 15. 11.7500 41.2500 360.0000 25. 14.7500 41.5000 360.0000 45. 14.7500 ~1.7500 360.0000 ss.oooo 14.7500 42.0000 360.0000 100. 14.7500 42.2500 360.0000 200. i4.noo 42.5000 36o.oooo 75. 25. 14.7500 42.7500 360.0000 0.0000 0.0000 14.7500 43.0000 0.0000 0.0000 15.0000 39.0000 0.0000 39.2500 0.0000 15.0000 0.0000 0.0000 15.0000 39.5000 0.0000 0.0000 15.0000 39.7500 0.0000 40.0000 0.0000 15.0000 0. 40.2500 0.0000 15.0000 7.5 40.5000 360.0000 15.0000 20. 15.0000 40.7500 360.0000 20. 41.0000 360.0000 15.0000 30. 41.2500 360.0000 15.0000 45. 15.0000 41.5000 360.0000 50. 11.7500 360.0000 15.0000 . . 42.0000 360.0000 250. 15.0000 42.2500 360.0000 100. 15.0000 50. 42.5000 Mo.0000 15.0000 16.11 11 42.7500 360.0000 15.0000 0.0000 0.0000 15.0000~ 43.0000 0.0000 15.0000 0.0000 . . . ~ 43.2500 . 0.0000 43.5000 0.0000 15.0000 0.0000 43.7500 0.0000 15.0000 0.0000 4L.0000 0.0000 15.0000 0.0000 39.5000 0.0000 15.2500 0 . 39.7500 0.0000 15.2500 0 . LO.OOOO 0.0000 15.2500 16.1111 40.2500 360.0000 15.2500 25. 10.5000 360.0000 15.2500 25. 40.7500 360.0000 15.2500 20. 41.0000 360.0000 15.2500 7.5 41.2500 360.0000 15.2500
~~~~~~ ~ ~~~ ~ ~ ~~~~ ~ ~ ~ ~ ~ ~ ~ ~

22 22 22 22 22 22 22 23 23 23 23 23 23 23 23 23 23 23 23 23 23 23 24 24 24 24 24

39 60 61 42 43

44
45 31 32 33

34
35 36 37 38 39 40 41 42 43

44
45 30 31 32 33

213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233

2%
235 236 237 238 239 240 24 1 242 243 244 245 246 247 248 249 250 25 1 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 2?3 274 275 276 277 278 279 280 281 282 283

15.2500 15.2500 15.2500 15.2500 15.2500 15.2500 15.2500 15.5000 15.5000 15.5000 15.5000 15.5000 15.5000 15.5000 15.5000 15.5000 15.5000 15.5000 15.5000 15.5000 15.5000 15.5000

is.noo is.noo

34
35 36 37 38 39 40 41 42 43 44 45 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

24
24 24 24 24 24 24 24 24 24 24 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26

15.7500 15.7500 i5.noo 15.7500

is.noo is.noo

15.7500 15.7500 15.7500

is.noo is.noo is.noo

44

15.7500 15.7500 16.0000 16.0000 16.0000 16.0000 16.0000 16.0000 16.0000 16.0000 16.0000 16.0000 16.0000 16.0000 16.0000 16 .OOOO 16.0000 16.0000 16.0000 16.2500 16.2500 16.2500 16.2500 16.2500 16.2500 16.2500 16.2500 16.2500 16.2500 16.2500 16.2500 16.2500 16.2500 16.2500 16.2500

41.5000 4i.noo 42.0000 42.2500 42.5000 42.7500 43.0000 59.5000 39.7500 40.0000 40.2500 40.5000 40.7500 41 .OOOO 41.2500 41 .so00 41.7500 42.0000 42.2500 42.5000 42.7500 43.0000 39.2500 39.5000 39.7500 40.0000 40.2500 40.5000 40.7500 41.0000 41.2500 41.5000 41 . n o 0 42.0000 42.2500 42.5000 42.7500 43.0000 39.0000 39.2500 39.5000 39.7500 40.0000 40.2500 40.5000 4o.noo 41 .oooo 41.2500 41.5000 41.7500 42.0000 42.2500 42.5000 42.7500 43.0000 39.0000 39.2500 39.5000 39.7500 40.0000 40.2500 40.5000 4o.noo 41.0000 41.2500 41 .so00 41.7500 42.0000 42.2500 42.5000 42.7500

360.0000 360.0000 360.0000 360.0000

35.. 80. 295.2222

70.

360.0000

30. 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 360.0000 7.7500 360.0000 8.2222 360.0000 7.5000 360.0000 7.5 360.0000 7.5 360.0000 7.5 360.0000 35. 36Ll.0000 .~~~~~~~ 50. 360.0000 756.3333 360.0000 302.11 11 360.0000 30. 360.0000 24.5556 360.0000 10.0000 0.0000 0.0000 0.0000 0.0000 10.0000 360.0000 360.0000 10.0000 0.0000 0.0000 7.5 360.0000 10. 360.0000 10. 360.0000 15. 360.0000 50. 360.0000 360.0000 400. 360.0000 676.4445 360.0000 171. a 9 25. 360.0000 20. 360.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 360.0000 7.5000 360.0000 8.6667 360.0000 7.5000 0.0000 0. 360.0000 10. 360.0000 15. 360.0000 20. 360.0000 40. 360.0000 600. 360.0000 600. 360.0000 250. 360.0000 30. 15. 360.0000 360.0000 15.8889 10.0000 360.0000 0.0000 0.0000 0.0000 0.0000 360.0000 7.5 360.0000 10. 360.0000 10. 360.0000 10. 10. 360.0000 360.0000 10. 360.0000 65. 360.0000 730.5555 360.0000 992.2222 360.0000 534.11 11 360.0000 60. 360.0000 30. 360.0000 25. 360.0000 10. 360.0000 10.0000
~~~~~~

n

26 27 27 27 27

45

2%
285

29
30 31 32 33

284
287 288 289

. .

-.a;:
28
28

27 27 27 27 27 27 27 27 27 27 27 27 28 28 28 28 28 28 28 28

34
35 56 37

290
291

292

293
294

38
39 40 41 42 43 bc 45 29 30 31 32 33 34 35 36 37 38 39

2%
2% 297 298 2w 300 301 302 303 304 305 306

307
308 309 310 311 312 313 314 315 316 317 318 319 320 321 322

40
41 42 43

28
2a

. 2 8
28 29 29 29 29 29 29

44
25 26 27 2a 29 30 31

325
324 325 326 327 328 329 330 331 332 333

29
29

32
33

29 29 29 29 29 29 29 29

34
35

34
37 38 39 40 41 42 45

29
29

29 30 30 30 30 30

:;a:
28
29 30 31 32 :a 33 0 3 4 3 35 ) 36 37 38 39 40 41 42
‘3

334
335 336 337 338 339

UO
341

342
343 344 345 346 347 348 349 350 35 1 352 353 354

16.2500 16.5000 16.5000 16.5000 16.5000 16.5000 .. 16.5000 16.5000 16.5000 16.5000 16.5000 16.5000 16.5000 16.5000 16.5000 16.5000 16.5000 16,5000 16.7500 16.7500 16.7500 16.7500 16.7500 16.1500 16.7500 16.7500 16.7500 16.7500 16.7500 16.7500 16.7500 16.7500 16.7500 16.7500 17.0000 17.0000 17.0000 17.0000 17.0000 17.0000 17.0000 17.0000 17.0000 17.0000 17.0000 17.0000 17.0000 17.0000 17. 0000 17.0000 17.0000 17.0000 17.0000 17.0000 17.0000 17.2500 17.2500 17.2500 17.2500 17.2500 17.2500 17.2500 17.2500 17.2500 17.2500 17.2500 17.2500 17.2500 17.2500 17.2500 17.2500

43.0000 39.0000 39.2500 39.5000 39.7500 CO.MM0 40.2500 40.5000 40.7500 41 .OOOO 41.2500 41.5000 41 .EO0 42.0000 42.2500 42.5000 42.7500 43.0000 39.0000 39.2500 39.5000 39.7500 40.0000 40.2500 40.5000 40.7500 L1 .oooo 41.2500 41.5000 L1.7500 42.0000 42.2500 42.5000 62.7500 38.0000 38.2500 3a.5000 3a.7500 39.0000 39.2500 39.5000 39 .BOO 40.0000 40.2500 40.5000 40.7500 41 .OOOO 41.2500 41.5000 41 . Z O O 42.0000 62.2500 42.5000 42.7500 43.0000 3a.7500 39.0000 39.2500 39.5000 39.7500

-Mo.oooo
Mo.0000 Mo.0000
360.0000 360.0000 360.0000 360.0000 560.0000 360.0000 360.0000 360.0000 0.0000 0.0000 360.0000 360.0000 360 .oooo 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 0.0000 0.0000

0.0000 0.0000 360.0000 360.0000 360.0000 360.0000

0.0000
0.0000 10. 15. 7.5 7.5 10. 30. 250. 650. 425. 50.0000 18.5556 18.f.444 10. 10.6667 10.0000 0.0000 0.0000 20. 15. 10.

IO.
40.
861.5555 400.

400.
90. 10.

0 .
0. 7.5 7.5 0.0000 0.0000 0.0000 0.0000 0.0000 0. 20. 30.

Mo.0000
360.0000 0.0000 0.0000 0.0000 0.0000 0. 0000 0.0000 Mo.0000 360.0000

Mo.0000
360.0000 360.0000 360.0000 360.0000 360.0000

so.

150. 600. 837.66-57 500. 387.5555

Mo.0000

M.

360.0000 20. 7.5 360.0000 10. 360.0000 360.0000 10.66-57 360.0000 . .10.0000 .~ 0,0000 0.0000 0 0000 . 0.0000 0.0000 O.OOO0 3bo.oow 20. 360.0000 277.1111 360.0000 226.4444 360.0000 277.3333 40.0000 360.0000 7‘29.0000 40.2500 ~o.aooo 998.5555 40.5000 360.0000 966.1111 40.7500 Mo.0000 441.5555 20. 41 .OOOO Mo.0000 41.2500 360.0000 10. 10. 41 .5000 360.0000 7.5 ~1.7500 360.0000 7.5 42.0000 360.0000 25. 42.2500 360.0000 0.0000 42.5000 0.0000

31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33

27 28 29 30 31 32 33

34
35 36 37 38 39 40 41 42 43 27 28 29 30 31 32 33

34
35 36 37 38 39 40 41 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 13

c4
45 25 26 27 28 29 30 31 32 33

34 34
34 34 34 34 34 34 34

34
34

34
35 36 37 38

i4 34 u

17.5000 355 17.5000 356 17.5000 357 17.5000 358 17.5000 359 17.H100 360 17.5000 361 17.5000 362 17.5000 363 17.5000 364 17.5000 365 17.5000 366 17.5000 367 17.5000 368 17.5000 369 17.5000 370 17.5000 371 17.7500 372 17.7500 373 17.7500 374 17.7500 375 17.7500 376 17.7500 37? 17.7500 378 17.7500 379 i7.noo 380 17.7~00 381 17.7500 382 17.7500 383 17.7500 3e4 17.7500 385 i7.noo 386 18.0000 387 18.0000 388 18.0000 389 18.0000 390 18.0000 391 18.0000 392 18.0000 393 18.0000 394 18.0000 395 18.0000 396 18.0000 397 18.0000 398 18.0000 399 18.0000 400 18.0000 401 18.0000 402 18.0000 403 18.0000 404 18.0000 405 18.0000 406 107 18.0000 18.0000 408 18.0000 409 18.0000 410 18.0000 411 18.2500 412 18.2500 413 18.2500 414 18.2500 415 18.2500 416 18.2500 417 18.2500 418 18.2500 419 18.2500 420 18.2500 421 18.2500 422 18.2500 423 18.2500 424 18.2500 425

38.5000 38.7500 39.0000 39.2500 39.5000 39.7500 40.0000 40.2500

4OISOOO
40.7500 41.0000 41.2500 41.5000 4i.noo 42.0000 42.2500 42.5000 38.5000

3a.noo
39.0000 39.2500 39.5000 39.7500
40.0000

10.2500 40.5000 40.7500 41.0000 41.2500 41.5000 41.7500 42.0000 37.0000 37.2500 37.5000 37.7500 38.0000 38.2500 38.5000 38.7500 39.0000 39.2500 39.5000 39.7500 40.0000 40.2500 40.5000 60.7500 41.0000 41.2500 41.5000 11.7500 42.0000 42.2500 42.5000 42.7500 43.0000 38.0000 38.2500
38.5000

38.7500 39.0000 39.2500 39.5000 39.7500 40.0000 40.2500 LOSOOO 40.7500 41.0000 41.2500

0.0000 0.0000 0.0000 0.0000 360.0000 30. 360.0000 347.2222 360.0000 316.4U5 360.0000 398.0000 360.0000 893.4445 360.0000 998.6667 360.0000 691.5555 360.0000 100. 10. 360.0000 10. 360.0000 360.0000 10. 360.0~0 7.5 .~~~~~~~ 360.0000 10. 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 36o.oooo 20. . . ~ . ~ ~ ~ ~ 360.0000 382.7778 360.0000 391 .OOOO 360.0000 395.5555 360.0000 597.1111 360.0000 996.1111 360.0000 943.2222 ... .. 360.0000 340.8889 360.0000 163.4444 360.0000 10. 360.0000 10. 360.0000 10. 360.0000 10. 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 360.0000 30. 360.0000 409.3333 360.0000 445.4445 360.0000 550.8889 360.0000 896.5555 360.0000 1085.6656 360.0000 707.1111 360.0000 225.1111 360.0000 20. 360.0000 15. 360.0000 10. 360.0000 7.5000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 360.0000 7.5 360.0000 40. .... .... 360.0000 233.1111 360.0000 396.8889 360.0000 494.7778 360.0000 711.8889 360.0000 W6.6667 360.0000 976.1111 360.0000 160.2222 360.0000 n. 360.0000 20. 360.0000 10. 360.0000 10.
~~ ~~~ ~ ~~~~~~~

'

~~

34 34
35 35 35 35 35 35 35 35

39 40

23
24 25 26 27 28 29 30 31 32 33

35 35
35

35
35 35 35 35 35 36 36 36 36
,
36_
. , ~. ...

34
35 36 37 38 39

426 427 428 429 430 431 432 433 434 435 636 437 438 439

440
441 442 443

u

23
24 25 26

445 lL6 447

448
U9

36 36 36 36 36 36 36 36 36 36 36
37 . .

E I

27

29 30 31 32 33 34

45b 451 052 453
454 455 456 457 458 459 660

35
36 37 38 21

37 37 37 37 37 37 37 37 37 37 37 37 37 37 37 37 37 37 37 37 38 38 38 38 38 38 38 38 38 . . 38 38 38 38 38 38

I 2

U T

462
463 06c 465 466 L57

23 24 25 26 27 28 29 30 31 32

468
469 470 471 472

33 34
35 36 37 38 39 40 41 22 23

473
470 475 476

477 478

4m
480 481 482

21

483

a 4

25 485 26 486 27 487 28 088 29 489 30 490 .~ 31 491 32 492 33 493 34 494 35 095 36 496
~~

18.2500 18.2500 18.5000 18.5000 18 SO00 18.3mO 18.5000 18.5000 18.5000 18.5000 18.5000 18.5000 18.5000 18.5000 18.5000 18.5000 18.5000 18.5000 18.5000 18.7500 16.7500 18.7500 18.7500 18.7500 18.7500 18.7500 ,18.7500 i8.noo 18.7500 18.7500 18.7500 i8.noo 18.7500 18.7500 18.7500 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.0000 19.2500 19.2500 19.2500 19.2500 19.2500 19.2500 19.2500 19.2500 19.2500 19.2500 19.2500 19.2500 19.2500 19.2500 19.2500

41.5000 41.7500 37.5000 37.7500 38.0000 38.2500 38.5000 38.7500 39.0000 39.2500 39.5000 39.7500 40.0000 40.2500 40.5000 40.7500 41.0000 41.2500 41.5000 37.5000 37.7500 38.0000 38.2500 38.5000 38.7500 39.0000 39.2500 39.5000 39.7500
40.0000

360. 0000 0.0000

0.0000 0.0000
360.0000 360.0000 360.0000 360.0000 360.0000

10.0000 0.0000 0.0000

0.0000
17.0000 50. 189.1111 195.8889 6OL.6667

360.0000 360.0000 360.0000 360.0000 360.0000
0.0000 ... ...

252.5556 10. 10. 10. 10.
0.0000 .. .

40.2500 40.5000

4o.noo
41 .OOOO 41.2500 37.0000 37.2500 37.5000 37.7500 38.0000 38.2500 38.5000 38.7500 39.0000 39.2500 39.5000 39.7500 40.0000 40.2500 40.5000 40.7500 41 .OOOO 41.2500 41.5000 41.7500 42.0000 37.2500 37.5000 37.7500 38.0000 38.2500 38.5000 38.7500 39.0000 39.2500 39.5000 39.7500 40.0000 40.2500 40.5000 40.7500

360.0000 7.5 360.0000 10. 360.0000 20. 360.0000 40. 360.0000 190.8889 360.0000 2 5 7 . U 5 360.0000 541.5555 360.0000 905.7778 360.0000 1391.1111 360.0000 880.6667 360.0000 368.8889 360.0000 . 70. - ... -.. 360.0000 10. 360.0000 10. 360.0000 10. 0.0000 0.0000 0.0000 0.0000

0.0000

0.0000

360.0000 10. 360.0000 20. 360.0000 60. 360.0000 203.6667 360.0000 273.8889 360.0000 509.0000 360.0000 818.5555 360.0000 1013.1111 360.0000 1039.6666 360.0000 554.4445 360.0000 182.3333 360.0000 70. 360.0000 10. 360.0000 10. 360.0000 10.

0.0000 0.0000 0.0000 360.0000 70. 360.0000 212.5556 360.0000 213.8889 360.0000 352.2222 360.0000 604.6667 360.0000 1206.0000 360.0000 1500.0000 360.0000 1005.7778 360.0000 622.2222 360.0000 285.3333 360.0000 50. 360.0000 10. 360.0000 10. 360.0000 10.
'

0.0000 0.0000 0.0000 0.0000

0.0000
0.0000

0.0000

38 38 39 39 39 39 39 39 39 39 39 39 39 39 39 39 39 39 39 40 40 40 PO L O 40 40 40 40 40 40 40 40 40 40 40 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 42 42 42 42 42 42 42 42 42 42

37

36
21 22

497 498 499

500
501 502 5 03 50* 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 506 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567

23
24 25 26 27 28 29 30 31 32 33

3L
35 36 37 22

23
24 25 26 27 28 29 30 31 32 33 34 35 36 37 21 22

23
24 25 26 27 28 29 30 31 32 33

3L
35 36 37 38 39 40 41 21 22

23
24 25 26 27 28 29 30

42
42 42 42 43

31
32 33 34 21

19.2500 19.2500 19.5000 19.5000 19.5000 19.5000 19.5000 19.5000 19.5000 19.5000 19.5000 19.5000 19.5000 19.5000 19.5000 19.5000 19.5000 19.5000 19.5000 19.7500 19.7500 19.7500 19.7500 19.7500 19.7500 19.7500 19.7500 19.7500 19.7500 19.7500 19.7500 19.7500 19.7500 19.7500 19.7500 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.0000 20.2500 20.2500 20.2500 20.2500 20.2500 20.2500 20.2500 20.2500 20.2500 20.2500 20.2500 20.2500 20.2500 20.2500 20.5000

7.5000 41 .OOOO 360.0000 0.0000 0.0000 41.2500 0.0000 0.0000 37.0000 -. . . . ~ ~ 37.2500 360.0000 398.5000 37.5000 360.0000 518.U45 37.7500 360.0000 348.2222 ~8.0000 360.0000 431.1111 38.2500 360.0000 606.2222 38.5000 360.0000 1342.5555 38.7500 360.0000 1819.6666 39.0000 360.0000 1081.3334 39.2500 360.0000 647.2222 39.5000 360.0000 413.8889 39.7500 360.0000 208.4444 40.0000 360.0000 80. 40.2500 360.0000 20. 40.5000 360.0000 10. 40.7500 360.0000 .-.-. . . . . . . . . . . 10. .. . 41.0000 0.0000 0.0000 37.2500 360.0000 421 .42% 37.5000 360.0000 611.8889 37.7500 360.0000 5 7 0 . m

39.0000 360.0000 39.2500 360.0000 39.5000 360.0000 39.7500 360.0000 40.0000 360.0000 40.2500 360.0000 40.5000 360.0000 40.7500 360.0000 0.0000 41.0000 0.0000 37.0000 . .~~~~ 37.2500 360.0000 37.5000 360.0000 37.7500 360.0000 38.0000 360.0000 38.2500 360.0000 38.5000 360.0000 38.7500 360.0000 39.0000 360.0000 39.2500 360.0000 39.5000 360.0000 391?500 360.0000 40.0000 360.0000 40.2500 360.0000 0.0000 10.5000 0.0000 40.7500 0.0000 L1.0000 0.0000 11.2500 0.0000 41.5000 0.0000 41.7500 0.0000 42.0000 37.0000 0.0000 37.2500 360.0000 37.5000 360.0000 37.7500 360.0000 38.0000 360.0000 38.2500 360.0000 38.5000 360.0000 39.0000 39.2500 39.5000 39.7500
LO.0000

689.4445 489.1111 357.0000 90. 40. 30. 30. 7.7500 0.0000

0.0000
380.4445 726.3333 658.1111

770.0000
1376.2222 1662.4445 829.8889 609.m8 437.2222 249.6667 193.0000 60. 30. 0.0000 0,0000 0.0000 0.0000 0.0000 0.0000
0.0000

0.0000 30. 120. 750.7T78 962.7??8 1653.6666 1054.8889 539.0000 307.5555 119.1111 34.2500 7.5000
~~~~~~

360.0000 360.0000 360.0000 360.0000
360.0000

40.2500 37.0000

0.0000 0.0000

0.0000 0.0000

0

43 43
L3 .-

22 23
24

568 569
570

20.5000 20.5000
20.5000 -... ...

37.2500 37.5000
37.7500

43 43 43 43 43 43 43 43 43 44 44 44
44

44

44
44 44 44 44 44
45

571 26 572 27 573 28 574 29 575 .30 576 n 31. s 32 578 33 579 21 580 22 581 23 582 24 583 25 581 26 585 27 586 28 587 29 588 30 589 31 590
17

IS

591

4 ; 45 45 45 45

45 45 45 45

45 LS ._ 21, ~~

18 19 20 21 22 23 25 26 27 28
29 -.

592 593 594
595

596 597
5013 -. .

20.5000 20.5000 20:SOOO 2o.5000 20.5000 20.5000 20.5000 20.5000 20.5000 20.7500 20.7500 20.7500 20.7500 20.7500 20.7500 20.7500 20.7500 20.7500 20.7500 20.7500 21.0000 21.0000 21.0000 21.0000 21.0000 21.0000 21.0000
21.0000 - ..

38.0000 38.2500 38.5000 38.7500 39.0000 39.2500 39.5000 39.7500 40.0000 37.0000 37.2500 37.5000 37.7500 38.0000 38.2500 38.5000 38.7500 39.0000 39.2500 39.5000 36.0000 36.2500 36.5000 36.7500 37.0000 37.2500 37.5000
37.7500 .

360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 36o.oooo 360.0000 360.0000 360.0000
~~

685.4445 562.3333 M9.4445 1190.4445 1454.0000 931.7778 755.~5 491.2222 167.3333 18.1111
0.0000 0.0000
0.0000

0.0000 0.0000 0.0000

360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000
0.0000 0.0000 0.0000 0.0000

1R.W

464.3333 690.0000 1495.2222 1548.33% 985.8889 763.5555 392.3333 67.5556
0.0000
0.0000

0.0000 0.0000

o.oooo
0.0000

o.oooo
0.0000

599 600 601 602

45 . .

45 45 45 45 45 45 45 45 46 46 46 46 46 46 46 46 46 47 47 47 47 47 47 47 47 47 47 47 47 48 48 48 48

30 31 32 33 34 35 36 37 21 22 23 24 25 26 27 28 29
~~

46 ~. 30

47 ..

19 20 21 22 23 24 25 26 27 28 29 30 20 2 1 22 23

21.0000 21.0000 21.0000 21.0000 603 21.0000 .~~ 604 21.0000 605 21.0000 606 21.0000 607 21.0000 608 21.0000 609 21.0000 610 21.0000 611 21.0000 612 21.2500 613 21.2500 614 21.2500 615 21.2500 616 21.2500 617 21.2500 618 21.2500 619 21.2500 620 21.2500 621 21.2500 622 21.5000 623 21.5000 624 21.5000 625 21.5000 626 21.5000 627 21.5000 628 21.5000 629 21.5000 630 21.5000 631 21.5000 632 21.5000 633 21.5000
~

38.0000 38.2500 38.5000 38.7500 39.0000 39.2500 39.5000 39.7500 40.0000 40.2500 40.5000 40.7500 41.0000 37.0000 37.2500 37.5000 37.7500 38.0000 38.2500 38.5000 38.7500 39.0000 39.2500 36.5000 36.7500 37.0000 37.2500 37.5000 37.7500 38.0000 38.2500
38.5000

360.0000 360.0000 36o.oooo 360.0000
540.0000

360.0000 36o.oooo 360.0000 360.0000
0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

160.1111 571.5555 818.8889 1635.4445 1252.0000 768.6667 687.4445 234.2222 7.5000
0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000
0.0000 0.0000

475.8889 802.1111 1037.1111 1643.2222 1065.7778 734.4445 662.2222 189.6667
0.0000 0.0000 0.0000

31 . _

z 5 636
637 638

bu

21.5000

I;O 1kO
21.7500 21.7500 21.7500

38.7500 613.2222 39.0000 237.8889 0.0000 0.0000 39.2500 0.0000 0.0000 39.5000 0.0000 36.7500 0.0000 37.0000 360.0000 280.0000 37.2500 360.0000 565.4445 37.5000 360.0000 732.5555

0.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000

164.6250 615.5555 818.3333 1215.8889 1508.0000 971.7778
799.7778

48

48 48

La .-

24 25 26 27

48 48 48 49 49 49 49 49 49 49 49 49 19 49 49 49 50
50

IS
21

639 640 641 642 643

29
30 22 23 24 25 26 27 28 29 30 31 32 33 19 20

6 U
645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660

50 50 50 50 50 50 50 50

22 662 23 663 24 664 25 665 26 666 27 667 28 M8 50 669 .~ 29 50 % 670 I 51 17 671 51 18 672 51 19 673 51 20 674 51 21 675 51 22 676 23 677 51 51 24 678 51 25 . 679 51 26 680 51 27 681 51 28 682 51 29 683 51 30 681 31 685 51 52 16 686 52 17 687 52 18 688 52 19 689 52 20 690 52 21 691 52 22 692 23 693 52 52 24 694 52 25 695
52 .26

I1 M1

52 52 52 52 53 53 _53 53 53 53
~~

28 29 30 13

I7

696
~~

11 .~

15 16 17 18 20 21

53 ._ 19 ..

697 698 6W 700 701 702 703 704 705 706 707 .~

21.7500 21.7500 21.7500 21.7500 21.7500 21.KOO 21.7500 22.0000 22.0000 22.0000 22.0000 22.0000 22.0000 22.0000 22.0000 22.0000 22.0000 22.0000 22.0000 22.0000 22.2500 22.2500 22.2500 22.2500 22.2500 22.2500 22.2500 22.2500 22.2500 22.2500 22.2500 22.2500 22.5000 22.5000 22.5000 22.5000 22.5000 22.5000 22.5000 22.5000 22.5000 22.5000 22.5000 22.5000 22.5000 22.5000 22.5000 22.7500 22.7500 22.7500 22.7500 22.7500 22.7500 22.7500 22.7500 22.7500 22.7500 22.7500 22.7500 22.7500 22.7500 22.7500 23.0000
~~ ~ ~~ ~ ~~ ~

37.7500 38.0000 38.2500 38.5000 38.7500 39.0000 39.2500 37.0000 37.2500 37.5000 37.7500 38.0000 38.2500 38.5000 38.7500 39.0000 39.2500 39.5000 39.7500 40.0000 36.5000 36.7500 37.0000 37.2500 37.5000 37.7500

360.0000 1090.2222 360.0000 1 4 0 6 . m 8 360.0000 958.6667 360.0000 7U.7778 ~~~~~~.~ . 360.0000 247.6667 360.0000 42.4000 0.0000 0.0000
~~

360.0000 .... ....

539.Md7

360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000
0.0000 0.0000 . . 0.0000 0.0000

612.5555 791.2222 967.2222 985.5555 830.2222 724.4445 136.11 1 1 7.5000
0.0000
0.0000 ~ ~ ~ ~ ~ . .~ . ~

0.0000 0.0000

0.0000 360.0000 360.0000 360.0000 360.0000 360.0000 38.0000 360.0000 .. . . .. . 38.2500 38.5000 38.7500 39.0000 0.0000 ... ... 39.2500 0.0000 36.0000 36.2500 360.0000 36.5000 360.0000 36.7500 360.0000 37.0000 360.0000 37.2500 360.0000 37.5000 360.0000 37.7500 360.0000 38.0000 360.0000 38.2500 360.0000 38.5000 360.0000 38.7500 360.0000 39.0000 360.0000 0.0000 39.2500
0.0000 39.5000 ~ ~ ~ . . . .

0.0000 482.2500 627.0000 699.8889 930.6667 1012.5555 922.6667

0.0000 ... ...

0.0000 115.6000 257.2222 573.2222 662.2222 829.U45 1253.5555 122l.MM 827.3333 678.1111 629.1111 319.8889 48.6667
0.0000

23.0000

23.0000 23.0000 23.0000 23.0000
23.0000
~~

53 53

ma

23.0000

709

23.0000

35.7500 36.0000 36.2500 36.5000 36.7500 37.0000 37.2500 37.5000 37.7500 38.0000 38.2500 38.5000 38.7500 39.0000 39.2500 35.0000 35.2500 35.5000 35.7500 36.0000 36.2500 36.5000 36.7500 37.0000

0.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000 360.0000

0.0000 0.0000

37.8889 178.4444 478.2222 597.0000 804.6667 1057.5555 1369.6666 962.3333 740.7778 577.2222 360.0000 388.2222

0.0000 ... ...

0.0000 ~ ~ ~ . . .

360.0000 162.3750 360.0000 311.7778 360.0000 496.11 1 1 360.0000 673.11 1 1 360.0000 749.U45 360.0000 1009.1111

._. .^ ,

53 53 53 53 53 53 53 53 53 53 53 53 54 54 54 54 54 54 54 54 54 54 54 54 54 54 55 55 55

22

23 24 25 26 27 28 29 30 31 32 33 15 16 17 18 19 20 21 22 23 24 25 26 27 28 15 16 17 55 _. 18 19 55 20 55 21 55 22 55 23 55 24 55 25 55 26 55 27 55 28 55 29 55 15 56
5. 6 .

710 711 712 713 714 715 716 717 718 719

720
721 722 7% 724 725 726 727 728 729 730 731 732 733 734 735 736 737

738
739 740 74 1 742 743 744 745 746 747 748 749

n o
n i 752

23.0000 23.0000 23.0000 23.0000 23.0000 23.0000 23.0000 23.0000 23.0000 23.0000 23.0000 23.0000 23.2500 23.2500 23.2500 23.2500 23.2500 23.2500 23.2500 23.2500 23.2500 23.2500 23.25500 23.2500 23.2500 23.2500 23.5000 23.5000 23.5000 23.5000 23.5000 23.5000 23.5000 23.5000 23.5000 23.5000 23.5000 23.5000 23.5000 23.5000 23.5000 23.noo

37.2500 37.5000 37.7~00 38.0000 38.2500 38.5000

.

16

56 56 56 56 .~ 56 56 56 56 56 56 56 57 57 57 57 57 57 57 57 57 57 57 57 57 57 57 57 57

3 17 n %noo
18 19 20 21 22 23 24 25 26 27 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 n 4

xnoo

n s

756 757

na
759 760 761 762 763 761.. 765 766 767 768 769

23.noo 23.noo 23.7500 23. n o o 23. n o 0

anoo u.noo u.noo
23.noo

u.noo
24.0000 24.0000 24.0000 24.0000 24.0000 24.0000 24.0000 24.0000 24.0000 24.0000 24.0000 24.0000 24.0000 24.0000 24.0000 24.0000 24.0000
~~~~~

m 771 m

m
774

m T16 rn

m
TIP 780

360.0000 1389.0000 360.0000 1111.3334 36o.oooo 850.2222 364.0000 669.0000 360.0000 516.6447 360.0000 292.5555 y . o 36o.oooo 36.4286 1n o 0.0000 0.0000 39.0000 39.2500 0.0000 0.0000 39.5000 0.0000 0.0000 39.7500 0.0000 0.0000 40.0000 0.0000 0.0000 35.5000 360.0000 50.8000 35.7500 360.0000 1LO. 1 1 1 1 36.0000 360 .OOOO 515.3333 36.2500 360.0000 697.4445 36.5000 360.0000 862.1111 3 6 . n o o 360.0000 1087.4465 37.0000 360. 0000 1470.2222 37.2500. 360.0000 1329.3334 .. . . _ . 37.5000 Mo.0000 854.2222 3 7 . n o o 360.0000 749.5555 38.0000 360.0000 639.8889 38.2500 360.0000 498.2222 38.5000 360.0000 210.8889 0.0000 .....~ 0.0000 38.noo 90.3no 35.5000 360.0000 35.7500 360.0000 216.1111 36.0000 360.0000 377.3333 36.2500 360.0000 741.2222 36.5000 360.0000 1097.6666 3 6 . 7 ~ 0 0 360.0000 1368.8889 37.0000 360.0000 1136.0000 37.2500 360.0000 669.2222 37.5000 360.0000 710.5555 37.7500 360.0000 638.7778 38.0000 360.0000 596.3333 38.2500 360.0000 297.7778 86.2500 38.5000 360.0000 0.0000 0.0000 38.noo 0.0000 0.0000 39.0000 35.5000 360.0000 104.5000 3 5 . n o o 360.0000 302.2222 36.0000 360.0000 816.0000 36.2500 360.0000 967.7778 36.5000 360.0000 1 0 5 0 . m 9 36.7500 360.0000 967.2222 37.0000 360.0000 869.2222 37.2500 360.0000 720.3333 37.5000 360.0000 607.7778 3 7 . n o o 360.0000 188.0000 38.0000 360.0000 327.7778 38.2500 360.0000 159.8889 0.0000 38.5000 0.0000 0.0000 0.0000 35.0000 0.0000 0.0000 35.2500 0.0000 0.0000 35.5000 3 5 . n o o 360.0000 311.7778 36.0000 360.0000 772.6667 35.2500 360.0000 10 18.0000 36.5000 360.0000 1020.6667 xnoo 360.0000 932.8889 37.0000 360.0000 774 .3333 37.2500 360.0000 611 .OD00 37.5000 360.0000 571 .OOOO 37.7500 -. ...~. 360.0000 285.2222 79.0000 38.0000 360.0000 0.0000 0.0000 38.2500 0.0000 38.5000 0.0000 0.0000 0.0000 38.noo 0.0000 39.0000 0.0000
~~~~~~

58 58

58
58 58 58 58 58 58 58 58 58 59 59 59 59 59 59 59 59 59 59 59 60 60 60 60 60 60 60 60 60 M 60 61 61 61 61 61 61 61 61 61 61 61 61 61 61 61 61 61 62 62 62 62 62 62 62 62 62 62 62 63
~~

14 15 16 17 18 19 20 21 22 23 24 25 13 14 15 16 17 18 19 20 21 22 23 13 14 15 16 17 18 19 20 21 22 23 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 12 13 14 15 16 17 18 19 20 21 22 12 11 13 14 15 16 17 18 19

781 782

783

784
785

786
787 788 789 790 791 792

793
794 795 796 797 798 799 800 801 802 803 801

805
806 807 808 809 810 81 1 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843

63
63

844
845 846 847 848 849 850 85 1

63
63 63 63 63 63

24.2500 24.2500 24.2500 24.2500 24.2500 24.2500 24.2500 24.2500 24.2500 24.2500 24.2500 24.2500 24.5000 24.5000 26.5000 24.5000 24.5000 24.5000 24.5000 24.5000 24.5000 24.5000 24.5000 24.7500 24.7500 24.7500 24.noo 24.7500 24.7500 24.7500 24.7500 24.7500 24.7500 24.7500 25.0000 25.0000 25.0000 25 .OOOO 25 .OOOO 25.0000 25.0000 25.0000 25.0000 25.0000 25.0000 25.0000 25.0000 25.0000 25.0000 2s. 0000 25.0000 25.2500 25.2500 25.2500 25.2500 25.2500 25.2500 25.2500 25.2500 25.2500. 25.2500 25.2500 25.5000 25.5000 25.5000 25.5000 25.5000 25.5000 25.5000 25.5000 25.5000

35.2500 35.5000 35.7500 36.0000 36.2500 36.5000 36.7500 37.0000 37.2500 37.5000 37.7500 38.0000 35.0000 35.2500 35.5000 35.7500 36.0000 36.2500 36.5000 36.7500 37.0000 37.2500 37.5000 35.0000 35.2500 35.5000 35.7500 36.0000 36.2500 36.5000 36.7500 37.0000 37.2500 37.5000 34.0000 34.2500 34.5000

0.0000 0.0000 360.0000 184.444 360.0000 483.8889 360.0000 774.3333 360.0000 1023.8889 360.0000 1028.11 11 360.0000 865.445 360.0000 671.44445 360.0000 408.445 360.0000 168.5714 360.0000 34.4000 0.0000 0.0000 0.0000 0.0000 360.0000 1 4 0 . U 360.0000 516.0000 360.0000 ~U3.8889~ ~ . . ~ . . . .. . ~ . 360.0000 926.1111 360.0000 1055.1445 360.0000 1034.5555 360.0000 825.4445 360.0000 403.2222

360.0000 592.4445 360.0000 673.5555 .... .... .~ . ... .. 360.0000 984.5555 360.0000 1106.5555 360.0000 978.7778 360.0000 729.3333 360.0000 200.444 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

u.mo

0.0000 0.0000

0.0000 0.0000

35.0000 35.2500 35.5000 35.7500 36.0000 36.2500 36.5000 36.7~00 37.0000 37.2500 37.5000 37.7500 38.0000 34.7500 35.0000 35.2500 35.5000 35.7500 36.0000 36.2500 36.5000 36.7500 37.0000 37.2500 34.5000 34.7500 35 .oooo 35.2500 35.5000 35 . n o 0 36.0000 36.2500 36.5000

360.0000 355.4445 360.0000 584.8889 360.0000 593.5555 360.0000 657.5555 360.0000 1134.8889 360.0000 1306.3334 360.0000 1029.1111 360.0000 533.5555 360.0000 132.7778

0.0000 0.0000
0.0000

0.0000
0.0000

0.0000
0.0000

0.0000

360.0000 301 .SO00 360.0000 539.3333 360.0000 605.1111 360.0000 664.8889 360.0000 769.1111 360.0000 1630.0000 360.0000 1603.7778 360.0000 1220.2222 360.0000 363.M67 360.0000 105.2500 0.0000 0.0000 0.0000 0.0000 360.0000 445.3333 360.0000 672.5555 360.0000 723.2222 360.0000 840.6667 360.0000 1003.0000 360.0000 1320.1111 360.0000 1035.0000 360.0000 377.0000

0

63
63

63 63
64

20 21 22

23

--.
"

11 12 13 14 15 64 16 66 17 64 18 64 19 64 20 64 21 65 5 . . 65 6 65 7 65 8 65 9 65 . 10 _ 65 11 65 12 65 13 65 14 65 15 65 16 65 17 65 18 65 19 65 20 65 21 22 65 65 23 65 24 65 25 9 66 10 66 11 66 12 66 13 66 14 66

64 64 64 64

852 853 854 855 856 857 858 859 860 861 862 865 861 865

25.5000 25.5000 25.5000 25.5000 25.7500 25.7500 25.7500 25.7500 25.7500 25.7500 25.7500 25.7500 25.7500 25.7500

8~
867

zs.noo

36.7500 37.0000 37.2500 37.5000 34.5000 34.7500 35.0000 35.2500 35.5000 35.7500 36.0000 36.2500 36.5000 36.7500 37.0000
33.0000 .. ..
~~~

868
869 870 871

~~

~~

a n

873 874 875 876

8?7
878
879

880 881 882 883 884 885 886 887

888
889 890 891 892 893 894 895 896 897 898 899 900
901

66 15 66 16 17 66 18 66 6 6 ' 19
67 67 67 67 67 67 67 67 67 -. 67 67 68 9 10 11 12 13 14 15 16 17 19

0

i a
8

902 903 904 905 906 907

908
W9 910 911 912 913 914 915 916 917 919 920 921 922

a
68
60

__ 68

io
11 .. 12 13 14 15 17 18 5 6

9

68

68
68 68 68 69 69

26.0000 26.0000 26.0000 26.0000 26.0000 26.0000 26.0000 26.0000 26.0000 26.0000 26.0000 26.0000 26.0000 26.0000 26.0000 26.0000 26.0000 '26.0000 26.0000 26.0000 26.0000 26.2500 26.2500 26.2500 26.2500 26.2500 26.2500 26.2500 26.2500 26.2500 26.2500 26.2500 26,5000 26.5000 26.5000 26.5000 26.5000 26.5000 26.5000 26.5000 26.5000 26.5000 26.5000 26.7500 26.7500 26.7500 26.7500 26.7500 26.7500 26.7500 26.7500
~~

33.2500 33.5000 33.7500

34.0000
3L.2500 34.5000 34.7500 35.0000 35.2500 35.5000 35.7500 36.0000 36.2500 36.5000 36.7500 37.0000 37.2500 37.5000 37.7500 38.0000 34.0000 34.2500 34.5000 34.7500 35.0000 35.2500 35.5000 35.7500 36.0000 36.2500 36.5000 34.0000 34.2500 3.4.5000 34.7500 35.0000 35.2500 35.5000 35.7500 36.0000 .... ... M.~OO 36.5000 33.7500 34.0000 Y.~SOO
~~

360.0000 43.3333 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 360.0000 407.4286 360.0000 629.7778 360.0000 765.5555 360.0000 862.7778 360.0000 1006.0000 360.0000 1168.445 360.0000 1171.2222 360.0000 582.445 360.0000 3 1 . U 360.0000 11.8000 0.0000 o.oooo 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 . . . ~ 0.0000 ~ ~ 360.0000 550.8889 360.0000 717.5555 360.0000 765.2222 360.0000 994.2222 360.0000 1017.5555 360.0000 991.8889 360.0000 894.8889 360.0000 490.2222 360.0000 10.8750 0.0000 0.0000
~~~~ ~~~~~~

0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 360.0000 665.5000 360.0000 674.8889 360.0000 690.7778 360.0000 932.6667 360.0000 W8.8889 360.0000 930.3333 360. 0000 737.445 360 .OOOO 717.7778 360 .OOOO 224.rn8 0.0000 0.0000 360.0000 690.6000 360.0000 858.8889 3bo.0000 865.7778 360.0000 972.2222 360.0000 1008.7778 360.0000 910.4445 360.0000 702.5555 360.0000 432.6667 3M1.0000 191.8889 .~~~ .. .. 3o.8000 0.0000 0.0000

0.0000 0.0000

0.0000

0.0000

m.oooo o.oooo

o.oooo
~

3L.5000

34.7500 35.0000 35.2500 35.5000 36.0000 36.2500 33.0000 33.2500

360.0000 36o.0000 360.0000 ..~ ... .. 36o.oooo 360.0000 360.0000 360.0000

422.7778 733.5555 931.5555 ioi4.3333 940.8889 RP.OOOO 656.8889 29.4286
0.0000 0.0000

26.7500 26.7500 27.0000 27.0000

360.0000
0.0000 0.0000 0.0000

0.0000

0

69 69 69 69 69 69 69 69 69 69 69 69 69 69 69

7 8 9 10 11 12 13 14 15
1 6

m m

70 70 70 70 70
70

m

70 7 1 7 1 71 7 1 71 71 71 71 71 71 7 1 72 72
72 '72

17 18 19 20 21 8 9 10 11 12 13 14 15 16 17 7 8 9
10 ..

923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950
951 . .
~

72 72 72

72 72
73 73

73 73 73
73

73 73
73

73 73 73
73

74 74 74 74 74 74 74 74
75 . .

75 75 75 75

952 953 954 955 15 . . 956 16 957 17 958 7 - 959 8 960 ' 9 961 10 962 11 963 12 966 13 965 14 966 15 967 5 968 6 969 7 9m 8 971 9 ,972 10 973 11 974 12 975 13 976 14 972 15 978 1 6 979 17 980 5 981 6 982 7 Pa3 8 964 9 985 10 986 11 987 12 988 5 989 6 990 7 991 8 992 9 w3

11 12 13 14

27.0000 27.0000 27.0000 27.0000 27.0000 27.0000 27.0000 27.0000 27.0000 27.0000 27.0000 27.0000 27.0000 27.0000 27.0000 27.2500 27.2500 27.2500 27.2500 27.2500 27.2500 27.2500 27.2500 27.2500 27.2500 27.5000 27.5000 27.5000 27.5000 27.5000 27.5000 27.5000 27.5000 27.5000 27.5000 27.5000 27.7500 27.7500 27.7500 27.7500 27.7500 27.7500 27.7500 27.7500 27.7500 28.0000 28.0000 28.0000 28.0000
28.0000 _. . .. . .

33.5000 33.7500 34.0000 34.2500 34.5000 34.7500 35.0000 35.2500 35.5000 35.7500 36.0000 36.2500 36.5000 36.7500 37.0000 33.7500 34.0000 34.2500 34.5000 34.7500 35 .oooo 35.2500 35.5000 35.7500 36.0000 33.5000 33.7500 34.0000 34.2500 34.5000 34.7500 35.0000 35.2500 35.5000 35.7500
36.0000

0.0000

0.0000

360.0000 360.0000 360.0000 360.0000
360.0000

360.0000 360.0000 360.0000
0.0000 0.0000
0.0000 0.0000

0.0000 0.0000

360.0000 360.0000 360.0000 360.0000 360.0000
360.0000 360.0000

360.0000 0.0000
0.0000

0.0000 0.0000 729.5555 881.1111 1000.3333 1014.6667 919.4445 675.1111 572.7778 73.8571 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 767.6667 1017.2222 1008.6667 931.7778 819.8889 655.0000 178.3333 8.6000 0.0000
0.0000

360.0000. 140.5556 360.0000 826.0000 360.0000 1022.7778 360.0000 950.3333 360.0000 649.5555 360.0000 742.8889 360.0000 395.7778 360.0000 26.8571 0.0000 0.0000
0.0000
0.0000

33.5000 33.7500 34.0000 34.2500 34.5000 34.7500 35 .oooo 35.2500 35.5000
33.0000

360.0000

10.0000 360.0000 89.0000 360.0000 329.2500 360.0000 928.0000 .....~~~ 360.0000 765.5555 360.0000 560.3333 360.0000 293.1111 360.0000 51.3333

28.0000 28.0000 28.0000 28.0000 28.0000 28.0000 28.0000 28.0000 28.2500 28.2500 28.2500 28.2500 28.2500 28.2500 28.2500 28.2500 28.5000 28.5000 28.5000 28.5000 28.5000

33.2500 33.5000 33.7500 34.0000 3k.2500 34.5000 34.7500 35.0000 35.2500 35.5000 35.7500 36.0000 33.0000 33.2500 33.5000 33.7500
3L.0000

34.2500 34.5000 34.7500 33.0000 33.2500 33.5000 33.7500 34.0000

360.0000 __. .. . ..

7.5000 . .
0.0000
0.0000

~

~

~

0.0000

0.0000 0.0000

0.0000

a

75

%
75

10

WL

28.5000 . .- ~ 34.2500 . .

n
76 76 .76 76 76 76 76
75 .

11 12 13 4 5

995
996

w7 WE
999

7 io01 8 1002 9 1003 i o 1006
91

i iooo

28.5000 28.5000 28.5000 28.7500 28:7500 28.7500 28.7500
~

z8.noo
28.7500 28.7500 2a.7500

1005

34.5000 34.7500 35.oooo 32.7500 33.0000 33.2500 33.~000 33.7500 34.0000 u.2500 34.5000 32.0000 32.2500

0.0000 360.0000 360.0000 0.0000 360.0ooo 360.0000

0.0000 737.8571 189.2000

o.oooo

o.oooo o.oooo o.oooo 0.0000 o.oooo
o.oooo

8.3333 7.5556

o.oooo o.oooo o.oooo 0.0000 o.oooo
o.oooo
0.0000 0.0000 0.0000 ~ . 7.5556 7.5000 0.0000 0.0000 0.0000
~~~~~~

77 77
TI

TI

a

1011 77 1012 77 1013 TI 7 1014 8 9015 77 9 1016 77 10 1017 TI 11 1018 77 12 1019 TI 13 1020 TI 3 1021 78 4 1022 78 5 1023 78 7 8 6 1024 78 7 1025 rn a 9026 . .78 9 1027 78 10 1028 78 11 1029 78 12 io30 78 13 1031 m 1 9032 .. . 2 1033 79 79 3 1034 4 1035 79 79 5 1036 79 6 1037 79 7 1038 79 8 1039 79 9 1040 79 10 1041 79 11.1042 79 12 1063 79 13 1OLL 80 2 1045 80 3 1 w 80 4 1047 80 5 1048 80 6 1069 80 7 1050 80 8 1051 80 9 1052 80 10 1053 80 11 1054 80 12 1055 80 13 1056 81 1 1057 81 2 1058 81 3 1059 81 4 1060 61 5 1061 81 6 1062 81 7 1063 8 1064 81
L 5 6
~

..

1 2

3

1008 1009
1010 . .

29.0000 29.0000 29.0000. .. . . 29.oooo 29.0000 29.0000 29.0000 29.0000 29.0000 29.0000 29.0000 29.0000 29.0000 29.2500 29.2500 29.2500 29.2500 29.2500 29.2500 -. 29.2500 29.2500 29.2500 29.2500 29.2500 29.5000 29.5000 29.5000 29.5000 29.5000 29.5000 29.5000 29.5000 29.5000 29.5000 29.5000 29. 5000 29.5000 29.7500 29.noo 29.7500 29.7500 29.7500 29.7500 29.7500 z9.noo 29.7500 29.7500 29.7500

__

32.5000 _ _ . .. . .

0.0000 0.0000
0.0000 . ~

~

~

mnoo

30.0000 30.0000 30.0000 30.0000 30.0000 30.0000 30.0000 30.0000

32.7500 36o.oooo 33.0000 360.0000 0.0000 33.2500 33.5000 0.0000 0.0000 33.7500 0.0000 0.0000 ... ... ... ... 34.0000 0.0000 0.0000 34.2500 0.0000 0.0000 34.5000 34.7500 360.0000 117.0000 0.0000 0.0000 35.0000 0.0000 0.0000 32.5000 7.5000 . ... .. 32.7500 360.0000 0.0000 0.0000 33.0000 0.0000 0.0000 33.2500 0.0000 0.0000 33.5000 n.7500 o.oooo o.oooo 0.0000 0.0000 34.0000 0.0000 0.0000 34.2500 0.0000 0.0000 34.5000 34.7500 360.0ooo 162.6250 35.0000 0.0000 0.0000 32.0000 0.0000 0.0000 0.0000 0.0000 32.2500 7.5000 32.5000 360.0000 0.0000 0.0000 32.7500 0.0000 0.0000 33.0000 0.0000 0.0000 33.2500 0.0000 0.0000 33.5000 0.0000 0.0000 ~.~~~~ 33.7500 0.0000 34.0000 0.0000 34.2500 0.0000 34.5000 0.0000 34.7500 0.0000 0.0000 35.0000 0.0000 0.0000 32.2500 .5000 32.5000 360.0000 0.0000 0.0000 32.7500 0.0000 0.0000 33.0000 0.0000 0.0000 33.2500 0.0000 0.0000 33.5000 0.0000 0.0000 33.7500 0.0000 0.0000 34.0000 0.0000 0.0000 34.2500 0.0000 0.0000 34.5000 0.0000 0.0000 34.7500 0.0000 0.0000 35.0000 0.0000 0.0000 32.0000 0.0000 ... ... 0.0000 32.2500 0.0000 0.0000 32.5000 0.0000 0.0000 32.7500 0.0000 0.0000 33.0000 0.0000 0.0000 33.2500 0.0000 0.0000 33.5000 0.0000 0.0000 33.7500
~ ~

81 81 81 81

ai

9 1065 10 1066 11 1067 12 1068 13 1069

30.0000 30.0000 30.0000 30.0000 30.0000

34.0000 34.2SOO 34.5000 34.7500 35.0000

0.0000 0.0000 0.0000 0.0000

-0.0000 0.0000
0.0000 0.0000

o.oooo

o.oooo

.

1069 grid p i n f s of which 733 are sea

0

Appendix B.

Final winds €or a sample storm. Winds are plotted at 6-hourly intervals, and at 112 degree intervals for clar-ity

.

L

a

a

a

a

i

0

Appendix C.

Check plots of results of continuous 3-year hindcast at three grid points. Variables plotted are from top to bottom: wind speed (1-hour average at 20 m ht), . wind direction (degrees from which), significant wave height (m), peak spectral period (sec), vector mean wave direction (degrees, toward which)

15

1986 DATA

R GRID P O g T 334
I

6,

.. , ,

1/I /86

711186

1011186

15,

1986 DATA

R GRID POINT 643
I

10
(/1

3 5
I

I

I

I

I

I

I

I

I

I

I

I

300
0200

3

100
0 4[

3L
L

I

300 I
7 .

"

I

'i

'

.....

....

/.

- r o w
A

VMD
.I.

HS
- r o

:o 0 s ,::'.'!>
Q)

0

0 0

0 0

1 ' T ' ' ;

........
....

-.

..

...
. .

2 7 .

.....

-. ....

-*>.::7-.

.,. .

151

1987 DATA I

1 GRID POINT 334

I

I

I

I

I

I

I

I

I

I

I

e

15
10

1987 DATA

.' '3 GRID POINT 643

a
I I I

I

I

I

I

I

I

I

I

I

4 7

I

10 8 6 P 4

2 0

1I1 187

411I87

711I87

1011187

20 c

1987 DATA

R GRID POINT 877

I
I

10,

I

I

I

I

I

I

I

I

I

I

I

300

;

I

v

;

>

100
0I .

I

I

I

I

1I1 187

711I87

e

I

I

. .

1 011 I87

e

2 0

A

s 00

0

- r o b o o c

VMD

TP
0 r u i L a
I '

-I

' , , ,' '' - ' . :.>

'

0

' ' ' '4.
I -

c

0

0 0

-

\ND
w
0 0

w

0 0

0

.> -. .? - . --.. ......... ..... - A ........... - ; _-.
'

.............-. ....

,p

> . ....... 3 .
,

..... -6'. .......... .... . _iW,

..-=

_.I

...

s al

..:..A
' -

. . "'-3 .......

z

?.--.
-. ...... .... -

.......,__ . . . .........-... . .- . ....... .-... a

z
. .

-

:3
.......
-%
-2

s al
,.

3 A

.........

s
3
. . . .-.Lc:

-

.....

I

I

I

I

I

I

I

I

I

I

I

- 1

I

I

I

I

I

I

1

I

I

I

I

I

\

111/88

711I88

e

1011188

10,

I

I

I

I

I
1

I

I

I

I

I

I

I

J
ij
.. ..

300

!

!

:

\

I

..

Appendix D.

0

The following is sample documentation of wind fields, pressure fields, wave fields and wave spectra delivered to DHI.

0

To: DALE KERPER. DHI From: J. A. Greenwood. Oceanweather Date: 16 June 1995 Subj: RED SEA; Wind fields for the following storms: no3 8112 8801 9201 8003 6203 8905 9202 8004 8302 8911 9302 8006 8402 9101 9311 8010 8511 9103 9401 8102 8606 9111

The ‘.PCH files contain wind direction and speed for %hour time steps in the six storms. The wind directions are in degrees, to which the wind is blowing, clockwise from true south; the wind speeds are in d s and reduced to an analysis height of 10m. The 136 cards for each time step are formatted as follows:
1 date of wind directions 2 t0.68 wind directions 69 date Of wind speeds 70 to I 36 wind speeds (7’3.18) (16F5.1) (7’3.18) (16F5.2)

The winds shown on land do NOT reflect orography: they are extrapolated over-sea winds, included as an aid to interpolation.

_.

.. :

. .. . ..

... ..

,

..

. .. ... .

To: DALE KERPER. DHI Fmm: J. A. Greenwoo& Oceanweather Date: 16 June 1995 Subj: RED SEA: wave statistics for 30 Red Sea Storms
The endosed CD-ROM contains sixty f l s ie:

Fieldsfiles.'.pfi.can be read in the format (14,16.15.2F6.2.3F7.4.F6.3,F7.3.2F6.1,F7.4). From lefl to right, the quantities punched are: 14 date: year & month 1 date: day, hour, & minutes (minutes are zero) 6 1 grid point number: points with significant height less than 4 cm are 5 omined. F 6 2 north latitude. degrees F6.2 east longitude, degrees R . 4 mO: units m e t e r s 7 F7.4 m l : units m e t e r s 7 radians sec--1. If moment is wanted in m 7 hz. divide by 2'pi . ~ 7 . 4 units m e t e r s 7 r a d 7 sec*-Z. If moment i wanted in m y h z 7 . m2: s divide by ( 2 ' p i ) 7 . F6.3 significant height, meters, computed as 4.00O'sqrt(mO). F7.3 peak period, sec F6.1 vector mean direction. to which waves are traveling. clockwise from tWe north, degrees F6.1 dominant direction: the direction that maximizes the mean Value Of cos(theta4om)7. F6.2 angular spreading, as computed by Gumbel Greenwood & Durand. See Peanon & Hartley. Biometrika tables for statisticians. vol.2. pp. 124, 125, 130. 371. The exponent N in a cos-N(theta) spread can be computed a s N = Z.'sp/(l .-sp): for example, ~ ~ 0 . yields N43.38 . 87 Spectra files, 'sps contains spectra at 36 grid points: those you requested plus points 278. 744. 81 1, 963, 975, added to facilitate interpolation. Spectra are only punched when the variance component at some direction, at some frequency not s exceeding 2 9 2 hz. i greater than .00001 m 7 . Spectra have been compressed ,by deleting directions in which all components would punch as zeroes. Toreconstitute a spedrum. the following code may help:
' ' '

0

.

.

10 FORMAT (F5.1.2315) ,F5.2,T39,F52.T47,F5.O,T58.F5.3. 11 FORMAT (T6,F5,O.F7.O,T23,F5.O,T31 S T69,F6:3.T82,F6.2.T108.12) real spec(23.24).date(2).lat.long integer jspec(23) c clear array SPEC do12lb=1.24 do 12 ia = 1.23 12 spec(ia.ib) = 0. read (10.1 1) date.gp.lat.long.deph,hsig.tpeak,vmd,ndirec do 14 Ib = 1.ndirec read (10.1 0) direcjspec Mirec = nint((direc+7.5)/15.) do 13 ia = 1.23

0

. .
,.

._ -. . ... ,-

. .

. .

. . .. .... . . .

..-

. . . .-

-

0 ...
\

_-

-.

0

13 14

spec(ia.kdirec) = le-5*jspec(ia) continue

SPEC, so constructed, will contain variance components (NOT spectral densiies). in m-2. The 24 angulaEbins start at 0 degrees (true north, to which waves are traveling) forthe counterdockwise end of the first bin. The directions punched are the midpoints of bins. The 23 frequency bins are defined as follows (frequencies in hertz): bin nominal lower upper bandwidth i frequency h! limit

1 2 3 4

0.0390000' 0.0429251 0.0472451 0.0520000 5 0.0572334 0.0629935 0.0693333 0.07631 12 0.0839914 0.0924444 0.1017483 0.1119885 0.1232593 0.1356644 0.1493180 0.1643457 0.1808858 0.1990906 0.2191276 0.241 1811

0.0371742 0.0409155 0.0450333 0.0495656 0.0545540 0.0660874 0.0727386 0.0800592 0.0881166 0.0969849 0.1067457 0.1174888 0.1293132 0.1423275 0.1566517 0.1724175 0.1897700 0.2088690 02298900

0.0409155 0.0450333 0.0495656 0.0545540

0.0037413 0.0041178 0.0045323 0.0049884 0.0600444 0.0054904 0.0066512 0.0073206 0.0080574 0.0088683 0.0097608 0.0107431 0.0118244 0.0130144 0.0143242 0.0157658 0.0173525 0.0190989 0.0210211 0.0231367

.,~~, ."."l
..?

6 7 8 9 10

0.0600444 0.0660874 0.0060430
0.0727386 0.0800592 0.0881 166 0.0969849 0.1067457 0.1174888 0.1293131 0.1423275 0.1566517 0.1724175 0.1897700 02088690 0.2298900 02530267

~..

11 12 13 14 15
16 17 18 19 20

21 0.2654541 0.2530267 0.2784919 0.0254652

22 .0.2921701 0.2784919 0.3065200 0.0280281
23 0.3215748 0.3065200 0.3373689 0.0308489

0

DALE KERPER, DHI From: Michael Parsons, oceanweather Date: 23 June 1995 Subj: RED SEA; Pressure Fields for storms 8512 and 9201 The * . A X files contain date, lat, long, and pressure every six hours for the two storms. To convert pressure to millibars add 1000 to the value in the ascii file. The file is formatted as follows:
format(i10,3f10.4)ymdh,lat,long,pres

To:

0

The pressures shown on land are used as an aid to interpolation.

//
..

. . . .. ........

.....
. .

0
. .

.... -. . . . .

. .... .. . ..... :., . . . . . .. . . . . . . . .. . .. ... .. .
...
: ' -

. . . . . . . ..:..,:.

. . ...>.,....a .. .:......i:. .

..!

.......

., .

-

....

0

0

Appendix E Reprints of selected papers which describe . Oceanweather's current practices regarding processing of historical ship report data (Cardone et al:, 1990), interactive kinematic objective analysis of surface marine winds (Cox et al., 1995), 3-G wave modeling (Khandakar et al., 1994), comparative analysis of OWIIs wind fields and wind fields produced by alternative sources, (Cardone et al., 1995) and evaluation of accuracy of OWI's 1-G and 3-G wave models in severe storms (Cardone et al., 1996).

On'Trends i Historical Marine Wind Data' n
VINCENT J. CARDONE A N D JUUET G. GREENWOOD
Ocmnmuhcr. Im.. Cos Cob. Connmim

h h R K A. CANE
LMlonr-Dohmy Geological Observaory. Palisades. New York

(Manuscript d v e d 28 February 1989. in final form 30 A u m 1989)
ABSTRACT Compilations ofsurhcc winds from ship rcponr dna 1854 show a n u m k oflong pricd w i a t i o w including a m o d toward nrcngihening winds over the pan thrrc decade* Some indgaton indicate that thae variations arc d c h a n g ~ the dimate symm. in while othcn that they are a d a m ofthc evolution of mearurrment technique. In an ancmpt to molvc lhis isrue, we have uamined individual ship rspons from thrrc regions with high data densities South China Sg, North padfis and North Atlantic shipping lanes We find that the apparent stdace wind rtrrngtheningfmm the 19501lo the PICynt is a mnrcquene of the i n d g uy ofancmomnm in place of sea-state epimata. Thc spxific muses are the opn;llional uy of an incorrect coovmion fmm k u f o n lone 10 wind speed, and the widespread assumption that the height of shipborne ancmomCtcnb IO m,whcrras thc actual m- height is 19.3 m. Cornstingthe mnmsion d e and vning the height lo be 20 m largely eliminates the Ucnd. The adjunmmt oianemomncr winds for suatifmtion effsu funher increases the condncncy bet.uccn the measured and ertimatcd win& A formula for mmcring the wind speds available i &dard amaged producn is prrrenlcd, in addition lo wind Ipcd the d o of n ertimatcd lo measured okcrntions and the avcragc airtemperature da a arc required. i m Evc~wilhthcmmcrioetheprr-I9Mwindsappsarlobe~ath~thcpml-I9Mwin&Thcmon~~y uplanation is thc a k n a ofunivmal nandardr for xa natc and Beauron f m beforc 1946. Pmriblc mncdie arc discus4 but the m is daunting. Though this caulionay tale d m not rdc out the rriacncc of& mnds k in surface winds it d m impugn thdr dnstability.

1. Introduction (NCDC) TDF-11 ship file. In the tropical Atlantic Scientific interest in historical surface wind data has O w n , a new surface wind and temperature database increased substantially in recent years, mainly because covering 1964-79 was developed from the same data of the importance of windstress forcing in the dynamics source by Picaut et al. (1985) and updated to 1984 by ot of the coupled tropical ocean and atmosphere. Surface Servain and Seva( 1987). The m s recent compilation, wind stress is the essential driving for ocean models, COADS (Comprehensive Ocean-Atmosphere Data and the needs of Ocean models compel much of this Set; Woodruff et al. 1987), provides global monthly means of surface variables over the period 1854-1979, interest. objectively calculated f o over 70 million ship obrm The starting point for nearly all climatological marine wind data prpducts is machine-readable fUes of servations synthesized through cornputer-bad quality historical surface weather reports kom transient and checks and duplicate report elimination procedures n weather ships maintained at major national meteoro- from the over 100 million observations contained i logical centers. For example, the FSU (Florida State a large number of machine-readable marine wind dataUniversity) tropical pacific ocean monthly mean surface winds (Goldenberg and O'Brien 1981) for the Almost a decade ago, Busalacchi and OBrien ( 198 1 ) yem 1961-83 were derived by subjective analysis of first demonmated that available wind data for the ship repom in N O M S N a t i o d CLimatic mta Center VOPid pacific, war as it was (and is), have enough information to allow a simple model to reproduce significant features of the El Niiio signal. Progress since ' LamOnt-hhvry C + d @ dobylvatory Comhtion l+~mter then has brought us to the threshold of operational 4553. ocean general circulation models for the tropical Paciiic (Leetmaa 1987) and the tropical Atlantic (Merle and Comeapding ovrhor nddreu: h Vincent J. monc, . o~ean- Morliere 1988 1. In addition, coupled models, generating initial conditions f o wind data alone, have b a n rm weather Inc.. Suite I . 5 River Road, Cor Cob, CT 06807.

114
I O U , R N A L OF C L I M A T E

.-_.

VOLUME 3

used to predict El Niiio (Cane et al. 1986; Barnett et al. 1988). Climate problems typically call for the integration of Ocean models overaveral decades or more. For example, establishing the skill of.El Niiio forecasting models requires retrospective forecasts over several events ( i Cane et al. 1986). Nevertheless, only a vz handful of ocean model simulations extending over more than a few years and driven by actual wind data has beta reported. (For the Pacific: Latif 1987; Busalacchi and OBrien 1981; Busalacchi et al. 1983; m e r 1989; Posmentier et al. 1989. For the Atlantic: Servain et al. 1985; Reverdin and Du Penhoat 1987). None venture into the data-poor years before 1947. Though not all call attention to it, all either remove the secular trend in the wind forcing or show trends in the simulated ocean variables that.= not present in the observations. An explicit discussion of these trends is given in Posmentia et aI. (1989). They conclude that, within the framework of widely accepted Ocean physics, the strengtheningof the PadGc trades indicated by the wind data is inconsistent with the absence of a similar trend in sea surface temperature and s a level. They therefore e suggest that the trend in the winds is spurious. A similar conclusion was reached by Wright (1988). Others s u m the trend is real. WhysaU et al. (1987) assert that the FSU winds exhibit a strengthening of the zonalwind in the central and castem tropical Pacific Oceatiof about I m s-’ overthe period 1961-83. Further, using a wind d a w compiled by the British Meteorological Office of monthly means in 5-degree squares over the period 1920-83, they find support for extending the trend back to the 1930s and argue for the existence of a major climatic anomaly in surface a winds during roughly the years of World W r 11. Ramage ( 1984) and Peterson and Hasse ( 1987)discuss the possible introduction of spurious very longterm trends into historical marine data by changes in the procedure used to derive wind speed equivalents ofBeaufort force estimates. It appears From these studies that between the middle of the nineteenth and twentieth centuries, the major changes in Beaufort force estimation procedure (at first related to sail characteristics, and later, for steamships, to the appearance of the sea) and lack of standard Beaufort force sea state and.Beaufort force wind speed equivalency relationships will probably forever prevent the separation of spurious trend from climatic change. In a funher study, Ramage (1987) studied COADS monthly wind speed summaries in 24latitude by 2degree longitude boxes containing the major South China Sea trade route,and found about a 1 m s-’ positive bias in wind s p n d averaged over 1950-79 relative to 1900-39. This differena was attributed mainly to two effects: 1 ) the 1946 formalization of the sea-state equivalents OS Beaufort force; 2) beginning after 1950, the in-ng number of marine wind observations based upon anemometer measurements.

The objective of the present study is to investigate whether, by working f o individual ship mt rather rm r s than previously averaged summaries, it can be shown that trends arise from inhomogeneous wind observation methods, and to what extent any spurious trends found can be removed by adjusting the individual historical ship reports for such measurement biases. This pilot study uses ship observationsin threeOcean regions characterized by relatively dense coverage of reports. The regions are the same South China Sea (SCS) trade route investigated by Ramage ( 1987) and segments of the most heavily traveled midlatitude mnsNorth Pacific and transNorth Atlantic shipping lanes. For several 2degree squares in each area,we compute the march of mean seasonal wind speed anomalies since 1900 in three ways: first, directly from the individual wind speed observations as given in the TDFI I data source; second, alter the individual reports are adjusted to a common reference level of 20 meters, taking into account differences between measurement types; finally, after applying a stability correction. Trends are then calculated and compared within different historical periods beginning in 1950. Trends ax not calculated through 1950 because the pre-1950 winds appear to be consistently weaker than the post1950 winds. This is discussed funher in, section 6. The analysis outlined above was carried out basically on two datasets. At first, we used a l l of the ship data in each region,after eliminating duplicate reports, taking the reported measured-estimated indicator code of the report at face value in the assignment of wind observation type. Second, we used a dataset for South China Sea only, eliminating ship reports in wurce decks in which the measured+stimated indicator code was apparently not correct. The analysis of all of the ship data in all three regions is described in section 4. while the analysis of the quality controlled dataset for the South China Sea is given in section 5. Fim, we describe our data source (section 2) and the wind speed adjustment procedures (section 3).
2. Data

All ship reports pnmssed in this study are contained in the NOAA NCDC TDF-I I archives. Basically, the NCDC data are maintained in three separate files sorted by historical time periods: 1854-1969 (also referred to as “NCDC Atlas” file); 1970-79 (usually referred to as the “decade of the ’70s”file); and the period 1980present. Each of these large files condm of Rporrs taken from a number of source “decks” (Socaued because marine data were originally stored on computer cards before being transferred to tape). Figure 1 shows the locations of the 2degree squares treated in each s h i p ping lane and indicates the total number of Ship repoRs in each square in the period 1900-84. For the second analysis refemd to in the previous d o n , the South China Sea database was updated through 1987.

-

J m m Y

1990

CARDONE. GREENWOOD A N D CANE

115

I60

I

I70

-40

-30

FIG.’I. Total n u m b ofship ~CPON ( 1900-84) in Iwodegrce IatiNde-lOngiNde swam studied: (a) South China Sca; (b) N o d psdfc Ocean; (c) N o d Atlantic Ocean.

I

Though amactive in many ways, we did not utilize the COADS source for this study. The COADS individual report files had not systematically retained from all source decks the character group containing the ship call sign, which might be needed to identify ship anemometer height. Apparently, the parsofthe ship report including the wind group and the measured-estimated indicator were not subjected to the substantial quality control procedures applied to many other elements of the observation. Fortunately, the NCDC sources constitute by far the largest part of COADS. On a global basis, the NCDC “Atlas” and “decade of the ’70s” files contain about 57 million reports compared to the final COADS tally of about 70 million r p r s More speeot. cifically, in the S-CS trade route, Ramage ( 1987) notes a total of 140 157 observations M e e n 1900-39 and 344 964 observations between 1950-79 in COADS. eot, AAer elimination of duplicate r p r s following basically the procedures used in the development of COADS, our file contains 114 016 observations between 1900-39, and 252 851 observations between 1950-79. It is difficult to imagine how our results could be changed by the larger COADS dataset.

3. Calculation of wind speed a. Adjustment o wind speed to 20-m refernnee level f The procedure used in this study to adjust ship wind speed reports to a standard level of 20 m has been used

for nearly two decades in an analysis scheme used to specify synoptic marine surface wind fields for numerical ocean surface wave models (e.g., Cardone 1969; Cardone et al. 1982). In the procedure, all ship reports of surface wind speed are adjusted to the effective neutral stability 20-m equivalent wind speed. If the reported speed is based upon a Beaufort estimate, Cardone’s ( 1969) Beaufort number-wind speed equivalency scale is used, and no further adjustment for stability is made. If the reported speed is measured, the speed is adjusted for anemometer heights different from 20 m and stability. The effective neutral wind speed is defined as U J Z ) = ( u * / k )~ e I z / & ( U . ) l l (1) where U is friction velocity,k is the von -in . constant, and & is a roughness parameter, generally a function of V . in the marine surface layer. Ifthe marine surface layer is neutrally stratilied, the effective and actual 20-m mean wind speedscoincide. In a non-neutrally stratified surface layer, U,is linked to the actual . wind through U, which may be calculated from a single layer wind measurement and the air-sea temperature difference using stabilitydependent profile forms and a roughness parameter form. (For a more detailed description of the algorithm used, see Cardone 1969; Ross et al. 1985,) In fact, this is the way wind speed observations from ships based upon anemometer measurements are gen-

I

. _ I

:.,-

,116

JOURNAL OF CLIMATE

VOLUME 3

and 25 m. The call sign group beginning with "4" has the lowest average height, because this group consists mainly of NOAA NDBO moored buoys, whose anemometers are maintained at 5 or IO m, depending on hull type and payload. Anemometer heights on modem fet of container vessels vary upwards of 25-30 m. les while heights on exploratory drillships and semi-sub mersibles range between 40 and 80 m. There are few, if any, reports from buoys, drillships or semisubmersible platforms in the regions studied. As pan of the present study, a sensitivity analysis was carried out in which the actual heights were assigned to measured winds whenever an anemometer height could be found to match the report call sign, and the wind corrected to 20 m, all unknown heights being assigned to 20 m. The results are virtually the same as the analysis, whose results are given in this paper, in which the heights ofall measured winds were assigned to 20 m. We attribute this mainly to the small proportion of the total population of measured winds for which an assignment of anemometer height is posT w I. hemometer heights on mnrins Wr s ~vencis o m buoys rip)soncd by leading CharaM on d dgn (In).Table g i number sible. For example, for the total South China Sea ship ~ of heights on fl, avaagc height, standard deviation, minimum and report database, of the 84 17 I measured ship repom ie maximum height within p u p . The mean height ofthe 2964 entries in the past 40 years, only 68 15 call sign-height matches k 1 . m. 93 were found. Thus, the direct calculation did not yield a conclusive result on the effect of varying anemometer height. Because a misassignment of 2 5 rn ryPjcally results in an ermr of less than 596, we expect the effect 0 of randomly distributed heights on climatologies to be 0 I 0 negligible. However, the use of a mean of 20 m instead 2 0 of 10 m, which has traditionally been assigned to ma.3 5 41 30 23.2 9.4 rine surface winds, corrects an error that is significant 4 93 . 4. . 3 5 50 23 because it systematically biases anemometer winds 5 44 26.7 92 . 5 42 6 21.5 1. 03 63 3 43 high.
~~

erally treated. That is, U is 6rst calculated from the . reported wind speed and air-sea temperature M e r . ena$ then U is calculated from ( 1). Anemometer height is assigned, if &sihle, through a reference ls it of ship call sign vesus anemometer height which contains about 3000 such heights, synthesized from various individual source lists, mainly W M O ( 1976). Sil for tl, a large number of observations in which the ship code wind indicator indicates that wind speed was indeed measured, either the call sign or height is missing. For such cass,the adjustment algorithm assigns a height of 20 m, which is close to the average of all known heights. Table I gives the distribution of ship anemometer heights in the list as a function of the leading character in the call sign. The average height overall is 19.3 m, while the average height for most call sign groups (call sign is related roughly to country of registry, e.g., J-Japan, U-USSR) ranges between 15

0

0

7

0
I' 17 30 , 8 40 26 119' 191 9 29

8
9
A

B C D E F
G

28.0 23.6 27.6 24.6 20.7 23.4

00 . 3I . 6.7

28
18

28 .
6.4 3.0

9 2 1

IO
14

28 29 38 28 3 0 28
4 1

6. Beaufort force wind speed equivalency scale
Reconciliation of measured wind speeds over open sea with wind speed equivalents of Beaufort force has been a subject of continuing interest to the Commission for Marine Meteorology (CMM) of the WMO. The first standard Beaufort force wind-speed equivalency scale was introduced in 1946, based upon data first published in 1906. A number of revisions have been proposed (Verploegh 1956; Cardone 1969; Kaufeld 1981). and in 1970 the WMO introduced a revixd scale (WMO 1970) of wind speed equivalents for UY in scientific studies, though the operational scale (in use since 1946) was not altered. The issue remains of critical importance, since even today approximately 50% ofwind observations received from merchant ships in the North Atlantic and about 20% of wind observations in the North Pacific are derived from Beaufort estimates. Figure 2 compares several of the various scales referred to above. The official scale is purported to relate to IO-m level wind speeds. Cardone ( 1 9 6 9 ) reviewed and synthesized series of simultaneous observations

3
7 12 6 12 8 I3 20

52

50
45

H
I J

18
530 33 I8 0 77

K L
M N

.
3. 44

3. 01
20.1
~~~~

20.9 13.0
5.7

28 50 80 80

0

0 P

3
2 0 0 223 4 1 I33 58 I62 0 n 46

38.7 2. 20

23 . 99 . 36 . 42 . 36 . 172 94 . 47 .

8 36 I5
5 II 3

42 40 29 26 2 1 31 80

Q R S
T U V W X

13.8
15.0

13.8 26.7 20.I
1. 25

6 8
6 "

80
26

Y
~

z

J W ~ Y 1990

CARDONE. GREENWOOD A N D CANE

1 I7

FIG. Comparison of Beaufor#number-wind rpeed equivalenu. 2.

0

from British and Canadian weather ships, in which one observer read the anemometer dial and another observer, separated from the first, &mated the Beaufort force. Cardone's scale relates the Beaufort number to the 20-m level (the average anemometer height on the ships involved). Kaufeld ( 1981) compared Beaufort-based wind speed observations h m transient ships and wind measurements from six N o d Atlantic Ocean weather ships for neighboring and simultaneous observation pairs. His scale was supported both from comparison of cumulative frequency d h b u t i o n s of wind speed and from direct correlation of observation pairs. Kaufeld's scale presumably relates to the wind speed at 25-m, the average anemometer height of the ensemble of weather ships used in his analysis. The wind speed reference level for the W M O scientifx scale is not d f i e d . While the newer scales shown in Fig. 2 differ slightly from each other. thev are all consistent with the idea that the operational- (1946) scale understates light winds and overstates strong winds. Some of the small

Kaufeld's scale and Cardone's scale at higher winds (Beaufort 6 to I I ) may be due to anemometer height differences between the two underlying datasets.For a logarithmic profie, using a roughness law consistent with the drag law of Large and Pond ( I98 I ) , the over-water wind speed at 25 m at neutral stability is ahout 3% greater than the wind s p e d at 20 m, while the wind speed at 20 m is about 8%greater than the wind speed at IO m. However, the differences between the operational scale and the newer calibrated scales cannot be explained in terms of refcrence level differences alone. The difference between Kaufeld's scale and Cardone's scale at Beaufort force 12 probably arises from the very limited data there. The dataset analyzed by Cardone, for example, had only nine occurrences of Beaufort force 12 out of 5,499 data pairs. Wind speed repons based upon Beaufort force es timates are "corrected" in this study to 20 m, with a . double conversion. The reported wind speed. derived presumably from the operational xale, is taken back to Beaufort force number. This force estimate is then

diffmnas be

I I8

JOURNAL OF CLIMATE

VOLUME3

converted to an equivalent 20-m wind speed udng Cardone's scale. No further correction for stratiliation is made on the assumption that Beaufort estimates already incorporate this effect. This i because the visual s estimate is dominated by the appearance of the sea surface, which is more a result of surface stress than 20 m wind. The conversions, made from tables in practice, are well approximated by (Cardone 1969):
U0 = 2

2.161J,"9,

i where U,s the rqwrted(Beaufort) wind speed in knots.
4. Trend analysis in three regions on the full dataset

The results presented in this section are taken from our analysis of the full TDF-11ship report archive a s sembled for this study. In this first analysis, all measured winds were assigned an anemometer height of 20 m, and stratification was assumed neutral throughout. Stability effects are considered in the analysis presented in section 5. The measureddmated indicator code was taken to classify winds as eitheranemometerbased or Beaufort estimates. In each repion, the files of individual ship reports were reduced to two time series of monthly wind speed anomalies. That is, one of the series which we shall refer to as "reported" was developed from the archived wind speeds, the second referred to as "adjusted" was produced from reports adjusted to 20-meter wind speeds. However, since in this fim analysis, anemometer height is assumed 20 m throughout and stability is not considered, the "adjusted" series consists of

measured winds as reponed and estimated winds adjusted to 20 m using Cardone's scale. The two time series were calculated for each 2degree square within a shipping lane as well as for the shipping lane as a whole. The individual monthly mean wind speed anomalies for both reported and adjusted series were calculated using 20-year ( 1965-84) normals derived from adjusted wind speeds. The time series of individual monthly mean anomalies were smoothed to seasonal mean anomalies, with the winter season defined as the months December, January, February, and so forth. The time series are described below for each region, at first qualitatively, over the entire period 1900-84. For the period 1950-84, linear regression of seasonal wind speed anomaly on time was camed out on both reponed and adjusted series. The slopes are compared for this entire period as well as for shorter periods within this span.

a. South China Sea
n ' e surface wind climate along t i entire &e mute hs is dominated by the Asian monsoon, with steady nonheasterlies in winter, southwesterlies in summer, and transition seasons in between. Mean speeds range from less than 5 m s-l near the equator in spring and fall to greater than IO m s-' in winter in most squares north of IO"N. Figure 3 compares the reponed and adjusted series over the entire period. The main properties of the reponed South China Sea series are very similar to the

JANUARY 1990

CARDONE, GREENWOOD A N D CANE
$6

119

0

20th century portion of the comparable global series developed from COADS "Uhmed" data by Ramage ( 1987). Wind speeds in the fim four decades of this century are systematically lower than in the most recent t r e decades. Also, from about 1950 forward, the pehe riod of most interest to modelers, there is an apparent trend in wind speed anomaly. Between these two broad periods, roughly coinciding with the end ofworld War 11, there is an apparent large anomaly oscillation. The first reports indicated to be measured winds do not appear in this region until June 1947. All earlier reports are presumably Beaufort estimates, yet the adjustment of Beaufort winds leaves an average negative wind speed anomaly for the pre-1950 period relative to the 1965-84 adjusted normals. As discussed also by Ramage ( 1987), this anomaly persistence is probably related to the absence ofa standard Beaufort-force seastate equivalence scale until 1946. Inhomogeneity in source deck may also introduce errors to the time series. Figure 4 shows the contributions 'of each source deck in this trade route over 1940-84. The apparent large anomaly oscillation between 1944 and 1947 appears especially suspect, since almost all ship reports in these years come from two decks containing a small number of U.S. Navy observations. There is also a relative deficit of observations between 1962 and 1964 when the gap between deck 119 and 128 in this period leaves only deck 116. Figure 5a shows the partition found between Beauestimates and measured wind speeds between 1950-84 in the shipping laneasa whole. Thesedbtributions only approximate the true mix. For example, all measured reparts between 1947-55 are contributed by deck 189 (Netherlands Marine Observations), which reports are coded as measured. between 1956-63 all wind R P N are coded aS estimated on the relevant source decks though almost cerrainly some measured winds m u t be present. To determine accurate distributions, if at all possible, would require tedious investigations of the history for each source deck. of the precise observation encoding, punching, data handing, and archive procedures, moa of which are not presently documented. In the next section we present results of our partially successful attempts to identify and filter from the data base, source decks with unreliable measured-estimated indicators. The analysis of the total data base, however, does not contradict the results obtained from the data base sub jected to tighter, albeit tedious, quality control. The alternate time histories of seasonal wind speed anomaly for the shipping lane as a whole are shown in Fig. 6. Superimposed on each figure are the trend o for the indicated lines derived from linear - n historical periods, each ending in 1984. The slopes of the trend Lines are tabulated over a l l squares and periods in Table 2. Especially for periods beginning in 1955 and 1960, there is a significant reduction in slope (by factors of

..

*e
le

I-.
"9

~~

lb

,P IP
10.

I P 8 6
IO,

,b

VP
'I*

::
''
w
IO'

lb

w
'w

::
;
IU
Id

.:
'O

. -l L.

L

SOUlH C H I N A SEA

50

60

62

I? SUVIII C l l l W A

SEf2

FIG.4. Total o u m k of ship rrponr pcr season ( 1940-84) and s- deck in South China Sea shipping h e . The follomng listing o

~ ~ e ~ t o ~ e l o u ~ ~ n u m b c r : ~ ~ 9 ~ a p a n o c ~ p ~ ~ NO. L 184 G I C Britain Marine obwmtions. I 16 US. ~ m h a n t ~ ~ Marins I IO U.S. Navy Marine OLmcmtions. 195 U.S. Navy Ship

~

~~

&

-

~

o

tional Mctmrological Center Global Telecommunications (GTS). 850 German Em GARP Global Expriment (FGGE), 849 FGGE.

:~ i

\

~

ete 926 W M O Foreign Exchange Data 888 Global W a h r Center (GTS). 889 AUTODIN (U.S. Navy)GTS. 927 US.Machant and Navy Manurript Obwrvationr. 555 F Numerical OMnography l m (Monterey) GTS, ,ntemational o~watio~,

3 to 6 ) in the adjusted series. The positive trend in both reported and adjusted trend lines for the 1950-84 period arises mainly because of the temporally coherent negative anomalies in both series in the early 1950s. This anomaly appears in all rubsquares and may be coming from the Same effect seen in reported or adjusted Beaufort estimates between 1900 and 1940. Possibly, this represents a period during which the 1946 standard for Beaufort force Sea-state equivalents became gradually accepted in practice. Table 2 (and plots such as Fig. 6 for individual sub squares, not shown) shows that the reduction in trend overall arises mainly in the near-equatorial sub-. The adjustment procedure has virtually no effect at 2O"N. Here, a much higher proportion of Beaufort force reports are in the range 4-8, for which the W M O 1946 and Cardone 1969 scales of wind speed equiva-

120

J O U R N A L OF

CLIMATE

VOLUME 3

Sea region, except that deck 189 (Netherlands) doe.
not contribute. As a result, there are no apparent mea

sured winds before 1963, after which inclusion of a

m

measured-estimated indicator on GTS reports was standardized (Fig. 5b). Over the whole perid- 190084 the adjusted series remain on average negatively biased before about 1940. The large anomaly oscillation between 1944 and 1947 is not so evident here because the contributing US. Navy source decks are so devoid of observations that for several seasons in the period no reports were found in this region. Bcw the wind speed climate did not vary spatially ea in 1950-84 over this midlatitude region, we show the resultsof the hend analysis i Fig. 7 only for the average n of all selected subsquares which connitute this shipping lane. The adjusted series exhibit much lower trends; the slope i reduced by about a factor O 4 for all periods s f beginning in 1965 and earlier. Note the large increase in the proponion of measured winds in this region in the past few decades, shown in Fig. 5b.

c. North Atlantic Ocean The main difference between the source decks which apply to this region and the other two regions studied is the relatively small contribution of Japanese decks
50

0

55
AIINUAL

60 65 S H I P REPORTS

75
LASONAL A V E R I C E WIND SPELOS

VI

-4

50

60

70

80

FIG. 5. Annual number of ship rqmns (1950-84) roned by reponed mcsrurcd4malcd wind indicator. in shipping lanes sludicd

0

ofSouUl~~~(uppr).NorthP~6c((mrddlc).NonhA~tic ocean (lower).

lenu agnc. Given the high steadiness of the winds along this trade route. the seasonal mean anomalies will therefore show little response to the adjustment procedure where the climatology places the mean wind in the vicinity of thcsc Beaufort forccs. There remains the possibility then of a secular change in wind in subtropical pans of the South China Sea.

50

60

70

BO

b. Nonh Pa& Ocean The mix of source decks which contribute ship reports in this region is much like that of the South China

of reponed and a d j d ship rcpoN in South CXM Sca shipping lane. Anomaly h
FIG. 6. Sasonal wind rpsed anomaly ( 1950-84)

rrlative toZ&yara(I%%) o f a d j d p v i n d w e d %T m d l i n s shown arc derived from linea -00 of anomaly on yesr

in indicalcd periods.

JWARY

1990

CARDONE. GREENWOOD A N D CANE

121

TABLE Slop (rn s-llyr) in lincar -on 2 .

of ocaronrl wind wed anomaly on ycar for 2 d e g a nrtsquara and all q u a m togcthcr

bawd on a d j d .uindr (see F .1. nrtsquarrr BIC numtucd from south t no&). s o
south china sea subrsvarr
pniod

1 .02 I a12 .01 I

2
,024

3

4 ' .029

5 ,029
,023 .019 ,007 ,009
.001 .017 .010 a10

6

7
,022 ,022

8

9 ,016 .020

10

11

.All
.027 .022 ,016 .XI5 .009

195044
1955-84

,023
,018 ,013

,021
.016 ,006 ,013 ,005 ,014 ,011 ,007 ,006

.021
,018 .&I4

.032 .023
,021

,025
OR 2 ,016
,010 .005

,022 .027
,017 ,014

,040 ,040
.03I .023 ,028 .022 .002

196044 . .

1965L84
1970-84

-.002 .013 .002 .Ol I
.007

,002
'

.m
,023 ,005
,015 ,013 ,011 a 1I

,012
.005 ,007

,007
.&I4

,004 .004

-

.008 -.001 ,011 ,008 ,005 .005

,012

,002

.002
,011 .008

,002
,010

,003
.@I

.ooo
-.001 -.001

,009
.009

.m
-.m
,008 .010

,002
,010
.010 . .

,006
.005

,009

,008 ,005 ,004

.00I

-.w2

-.COS

-.om
-037

,002
,013 ,013

.005

,005

..

<

*r

in the North Atlantic, though, of course, the several source decks derived from GTS collections include a varied mix of sources within themselves. The alternate time series for the whole period 1900-84 resemble t e h series derived in the other two regions before 1940 as regards the negative bias. There is evidence from analysis of pressure data that the Nonh Atlantic westerlies were stronger in the early 20th century than after World War 11 (Parker and Folland 1988), further supporting our suspicion that the pre-1950 wind speed deficit is spurious However, after 1950, as shown in Fig. 8, t h m
4,+EASDMAL AVERAGE Y l N O S P E E D S

is relatively little trend in either the reported or measured series. There is a suggestion of a slight negative trend in d periods beginning after 1955. Figure 5c &DWS the &bution of m a u e and &mated wind esrd speeds i this region. The tendency for increased inn cidence of measured winds in this region is much less pronounced than in the Pacific regions.

5. Further adjustments: the quality controlled dataset The analysis of the full dataset suggested that the measured-estimated indicator was, at least in some
4$EASDNAL AVERAGE Y l N O S P E E D S

-4

50

60

70

80

so
4

60

70

80

I

4

.2 s
3
Ln
L n l 0

a A

-

1
2

.$
I

z

V I I 0

io 2.1

C A

io

2.1 . .
v)

3-2
-7

a-2
7 0

0

-3
-4

*-3
-4

so

60

70

80

50

60

ro

80

FIG. 7 As in Fg 6,uccpt for No& . i

padfc

Ocean.

FIG. 8. As in Fg 6 except for No& i .

AUantic Oaan.

122

JOURNAL OF CLIMATE

VOLUME 3

source decks, not a reliable measure of the wind mcasurement type. A few examples were noted above. This problem was confirmed when reports from vessels known to carry anemometers were tracked through time and, in many cases,through merent source decks.
It seems that erroneous measured-estimated codes characterized certain decksand even certain time periods within individual decks. This is perhaps explainable in that some source decks are in fact collections of several heterogeneous sources. Another possible source of error may be related to the way the indicator is coded “blank” for estimated, “zero” for measured, with no allowance for a “missing” indicator. A relatively simple scheme was devised that effectively filters erroneously coded reports. The filtering was applied only to the South China Sea data, since it reduces the amount of data significantly. Even there, several gaps w m in the resultant dataset; that is, there were several seasons within the past two decades left with no reports. The total number of ship reports in the whole shipping lane in the 20-year base period 1965-84 was reduced from 264 806 to 157 358. The procedure requires sorting and tallying the ship reports in the shipping lane by year and month and

indicated measurement type. Between January 1900 and May I947 all repo’rtswere treated to be estimated. &tween June 1947 and May 1955 al repom were l indicated to be estimated, except for reports in deck 189 which were all indicated to be measured, probably erroneously. Reliably coded indicators of measured winds 6rst appear in April 1964.in deck 128. Between 1964 and 1984,it was relatively easy to identify specific time periods within specific decks in which the ratio of measured to estimated winds behaved erratically from month to month or fell outside of reasonable ranges established by most other decks. These periods were excluded. Table 3 shows the admitted decks and time periods. Outside the indicated time periods, the ratio often discontinuously passed into near 0 or 1 within specific decks. The trend analysis was repeated on the filtered dataset. In addition, the measured winds were further adjusted for stratification in the calculation of means, anomalies and trends for the adjusted series. As noted in d o n 3, the calculation of the effktive-neuual20m wind f o measured winds requires knowledge of rm anemometer height and air-sea temperature difference (or more precisely, the air-sea virtual potential tem-

0

TABLE Time p e r i d and mum decks admincd into hltmd ship repon database, bawd uwn reliability 3. o f m e a t u d d m a u d indietor. EnW idmti6a year/month limits ofadmincd dam
Sovrct deck
YC3I
(19CW

I I6
5601

I28

555

849

850

888

890

926

927

56 57 58 59 60 61 62 63

I
I
1

I

I I
6306

64
65 66 67 68 69 70 71 72 73 74 75 76 77 78 79

6404

I
I

I
I
6912 7012 700 I 7101 (7107) 7405 (7412) 7812 7911 7812 7911 7801
1

0
I I
7307
1 1
1

1 1

7601 7711 8005 8001
1
1

80
81 82 83
RA ”.

1 1 1 1 1
1
1

1
1

85

1

86 87

8612

1 1

8712

. ..

..

.

. .--

-

~

..

...

JANUARY 1990

CARDONE, GREENWOOD A N D CANE

123.

!,

prature difference). As in the earlier analysis, too few height matches were found to make the height assignment effective and 20 m was assumed throughout, though there appears to be little error introduced by assignment of a mean height of 20 m to all anemometer based reports. To provide air-sea temperature differences for the adjustment of measured reports, we calculated a mean air-sea temperature for each month and each subsquare in the shipping lane by averaging the reports in the relatively data-rich "decade of the 70s" file of filtered data. This was done because air or sea temperature is often missing in individual reports, which otherwise contain valid wind data. The monthly mean climatological difference, which ranges from +0.2"C to -1.3"C, is almost always negative, indicating unstable stratification. After correction for moisture stratification (a relative humidity of 75% was assumed throughout), the virtual potential temperature difference is offset from the air-sea temperature difference by about 1 "C. and even the areas and months with slightly positive air-sea temperature differencebecome unstably stratified. As a check on the consistency of the adjustments of estimated and measured wind speeds, we compared monthly means computed separately for each type, in each subsquare. and each month between 1965 and 1984. The means were computed only if there were at least 15 ship o b t i o n s of each type in a given square in a month/year. This condition was satisfied in 1066 of the possible 2640,monthly means ( I 1 subsquares X 20 yean X 12 months). In each cue, the measured mean (with or without stability correction) is plotted along the abscissa, and the estimated mean (with or without the scale adjustment) along the ordinate. In all four cases the scatter about a best fit tine is about the same, corresponding to a standard deviation o f a p proximately 0.8 m s-'. However, the departure from the 45 degree line, the locus ofconsistent observations, differs among the four. In Fig. 9a, each axis is just as reported, so that here the two types are compared in the form in standard use. As expected for this range of wind speeds on the basis of the earlier cited studies of the Beaufort scale wind equivalents, the estimated winds are systematically lower than the measured winds: 83% of the points fall below the 45 degree line while the mean difference is 0.75 m s-'. Adjustment of estimated winds only, Fig. 9b; brings the means between the two data types into somewhat closer agreement: the mean difference is reduced to 0.53 m s-' and the ratio of points below the line i s 0.24. However, agreement is still far from satisfactory. A more thorough look at the plot reveals that the two types are now consistent at higher wind speeds (say 914 m s-I), but the estimated winds are systematically higher at low winds speeds. The light winds in the South China Seas would tend to occur in unstable conditions, when mating a 20-m measured wind as if the atmw sphere were neutrally stratified would understate the

-

surface strrss. We further reason that the Beaufort estimate, being based on the appearance of the sea, is more a measure of wind stress at the surface than of wind speed at any panicular level. Hence there is reason to believe that adjusting for stability will further reduce the discrepancy. Note from Fig. 9c that the stability correction alone, without the equivalency scale adjustment. only makes matters worse. Figure 9d shows the desired result: adjustment of estimated winds for Beaufort scale effects and measured winds for stratification provides the greatest consistency between the monthly means. The mean difference is only 0.15 m s-' and the ratio is 0.42, close to the ideal one-half. The results of the comparative trend analysis on the filtered dataset, in which both estimates and measurements are adjusted in the "adjusted" series, are shown in Fig. IO (the whole shipping lane). The series shown runs from 1955 to 1987, as opposed to the series in Fig. 6 based upon the total data set, which covers 195084. The base period used for the calculation of anomalies remains 1965-84, and again the base was computed from the adjusted wind speeds. Trend tines are computed for two periods: 1956-87, and 1965-84. Again, we find that over the past 30 years or so, the trend toward strengthening winds, shown in the reported series, is reduced by an order of magnitude (in slope) in the adjusted series, for the series as a whole, and for lower latitude (0"-IO0N) subsquares in particular (not shown). Over the 20-year period 196584, the redudon in vend is not nearly so large.
6. Discussion

There are a number of trends and interannual oscillations in the climate record that most likely reflect real variations in the earth's climate. Presumably at least some of these are accompanied by changes in the surface wind field over the oceans. In view of the high probability that the ocean participates actively in any low-frequency variations in the climate system and the certainty that the surface winds drive the Ocean circulation, surface mixing, and heat exchange, a knowledge of these wind changes is clearly desirable. Our investigation reinforces the pessimism of earlier studies (Ramage 1987; Peterson and Hasse 1987) for the y e a n before 1950. The lack of standardization of either sea-state description or Beaufort scale means that, in its commonly available form, the variations in the wind data are likely to be due to variations i ren porting procedures. This is not to say that there were no significant real changes in the winds, but it does say that their detection will be difficult. A study of each country's procedures (e.&, as documented in their mariner's manuals) would Seem to be needed to assign a consistent wind speed to each reported Beaufort number. Information about each ship, such as the call sign giving country of origin, must also be available. The apparently consistent bias of the 1900-50 data

. - .. . .

....-- ....... .. ..... .. -.-.

.

.~~ . .

-

.-..

. ,

.

..^_

-

124
IS

J O U R N . A L OF CLIMATE

VOLlME 3

MEAN DIFFERENCE8

0.75

.
VI

-RA7lOi

0.83

..

MEAN DIFFERENCE-

-0.53
0.24

RAlID8

..

E

D

"0

c e
0
W

D

In

r 3

o s W
e

< S
e
VI W

0

In

.
5
0 P
W

0

w 10 P
c c
In
0

e 3

E 5 e <

z e
W

m

0 0

NEASURED WINOS. 5 1 A 8 1 1 1 1 1 CORRECTED I N l S l

5

10

IS

FIG.9.Compsiron of monthly mean wind sped in lwadcgra latitude-longitude q u a m of the SO-vth China Sea shipping h e ( 196584) fmm (a) dmatcd and measured ship winds as rrpoNQ (b) adjusted d m a l d winds and mned meawed uind%( C ) rrpoltcd dmud Gods and adjuacd menrrrd winds, (d) both otimatcd and m c a r v d winds adjuned. The mtan diaaeno and the ratio Of poinu klw the line t total poinu arc d m ar well o

(Vir. Fig. 3) holds out some hope that the above i s overly padminic, but those asserting that a nustworthy climate signal can be extracted from the pre- 1950wind data have the burden of proof. We anticipate that a successful m t e g y will have to make use of data auxiliary to wind observations, such as surface pre-ssure (cf. Wright 1988). The data from the early 1950s is problematic; our best guess is that it is a transition period in which the

international standard promulgated in 1946 w d d y gained adherents. Thereafter the principal problem wt wind datasets is solvable: the widely used 1946 ih WMO Beaufon force wind speed equivalency scale is in error. T i problem w s identified twenty y w ago hs a (Cardone 1969; WMO 1970) and is common knowledge in the wave modeling community. CorE~ted formulas are shown in Fig. 2. A comparison of Fig. 9a and 9b or the two panels of Fig. IO illustratesthe effect

WOOD ~

N CANE D

125

0

55

KO

KS

70

75

80

85

FIG.10. 'Asin Fig. 6, except for qualifycontrolled data set in
pcricd 1955-87. and both estimated and measured windl con'ecled in the adjusted r i ru

of this error, while a comparison of Fig. 9d with 9b shows the importance ofaccounting for stability at low wind speed. It would be advantageous to correct marine surface wind data products such as the COADS monthly averages or the FSU wind field analyses without having to reconstruct them from scratch. Figure 11 is an attempt to meet this need: given the average wind speed as reported ship winds-which is the speed given in most climate data products-and the percentage ofship reports of estimated winds using presumably the I946 Beaufort force-wind speed equivalency. and the airsea temperature difference. applied to correction of all measured winds (assumed at 20-m height), we have calculated the adjusted wind using the Cardone ( 1969) equivalency scale and surface layer wind profile model. The difference between the adjusted and reponed average wind speed is shown in Fig. I I for various percentages of data obtained from Beaufon estimates and three stability categories. Unfortunately, the percentage of estimated winds has considerable spatial and temporal variation. Moreover, it is often reponed incord Y . The data for stability are another source of difficulty. Humidity and air-sea temperature differences are not always reponed and are often unreliable. For example,
I

it is the well-founded practice of the British Meteorlogical Office to disregard all daytime air temperature measurements. Careful treatment and the ability to do some averaging can reduce the problem. Fortunately, stability has the largest relative effect on light winds and unstable conditions, and because a large relative error in a light wind is still a small absolute error, the effect on the Ocean is usually small. The principal exception is the equatorial ocean, which is so well tuned to the wind that even small absolute wind errors are non-negligible. A comction based upon a mean airsea temperature difference should work reasonably well in the tropics, where this field exhibits relatively high spatial and temporal coherence. In contrast,over the middle and high latitude oceans, stability variations, including changes of sign, typically occur on synoptic time and space scales. For monthly mean or more highly averaged wind products, the net effect of such stability variations on the higher winds prevailing there is likely to be negligible. For wind products with higher temporal resolution it may be neceSSary to go back to the individual observations and compute the adjusted mean winds directly. Even with the adjustment procedures described, numerous problems remain which limit the quality ofthe data. The Beaufon estimate has limited precision and is only as good as the observer. The comcted scale was consmcted from a limited data base: the differences of the three "corrected" scales in Fig. 2 are a measure of this uncertainty. We have taken the Beaufon force wind equivalent as a direct estimate of the effective neutral wind, since this is the wind which relates to the near-surface wind profile in an arbitrary svatification and which therefore affects the appearance of the sea. Any attempts by the observer to modi& his Beaufon estimate based upon the "feel of the wind in the face" would tend to violate that assumption. Anemometer-based winds are subject to a number of error sources not accounted for here, such as poor instrument calibration, flow distortion effects, improper averaging intervals, and incorrect conversion from "apparent wind" to true wind (e&, see Dobson I98 I ). The true height of a measured wind is rarely known, though with our reasonable default (20 m rather than 10 m ) the error this causes js typically less than 5% and is unbiased. Uncertainties in height may be compounded by uncertainties in the stability. All of the difficulties mentioned in the previous paragraphs would seem to introduce random errors, which can be expected to average out if the data are not too sparse. The incorrect equivalency scale is more prnicious because it creates spternaric ~ O K &cause . it is biased low at low wind speed and high at high wind sped, it alters patterns as well as overall amplitudes. For example, the wind stress curl will be altered more than the wind stress itself. Our results strongly support Ramage's ( 1987) contention that the strengthening of the trades from the 1950s onward is

... . ..... . .

126
2

J O U R N A L OF C L I M A T E

. . . . . . . . . . . . . . . . . . . .

2

. . . . . . . . . . . . . . . . . . . .
.............

.

.

.

-....

I -

o
w

.
.
.

u

0

a .
Y

u

0
Y

,

0.

v.. 3%

'.:..a. ? ., :

--._ ..
1--

w 3

.
.

9
-I

7

".... . :
".\
..,

. . .

-

..
-

A I R - S E A TEnP DIFF 0 . 1. BEAUFORT REPORTS

- 2 l " " " " " " " " " " '
0
2
I

6 8 IO I2 I4 I6 REPORTED WIND SPEED l M / S l

18

20

I .

.................
...-

. . I .

_....-+_....

i '
. .
I'

. ........ . .-..
..

.

0"
, i s

*-

-........

.

: /'
9 ,
I

*--.-*--.
I -

..... . m., .. ........
..

"..

.... .....y::.. *-,. .ss. * __--*---+__-__I:.
': v

.

FIG. I I . DitTercncc bcomn acjUncd and reponed mean wind
s p e d for indicated pcmntaga
ya

of aimatcd winds, and air minus virtual potential t~mpcraturc diffmncc of (a) - I T (b) 0, (c)

+IT.

-..a ._,

a..
I

I'.
.-+,'
.I

*--*--20

-___e.--\..\ : . .%
I'

I x

L

~ I R - S E A i m p DIFF

\

e--,

I

#'

L BEAUFORT REPORTS 0 --e--

."

0
I8 20

-

0

2

1

2

"

~

-4

"

" ' " ' " ' " ' " ~ 6 8 IO I2 I 4 16 REPORTED Y I N D SPEEO I M l S I

a consequence of the change to measured winds rather than the climate change suggested by Whysall et al. ( 1987). Furthermore, the elimination of this trend in the winds eliminates the apparent inconsistency between wind and Ocean data pointed out by Posmentier et al. (1989). A recalculation of the wind fields with the corrected equivalency scale should be decisive. The corrected scale reduced but did not eliminate the low bias of pre-World War I1 winds relative to the post-war data. (The striking mid-1940s anomalies are .. almost certainly due to exclusive reliance on the U S Navy source deck for this period). Though we have

not proven it, ourexperience with the post-1950s data leads us to conclude that Ramage's ( 1987) explanation is most likely correct: the low winds are an artifact of changing sea state and Beaufort force scales. It is interesting, and possibly instructive, that the problem addressed in this paper was uncovered in the course of attempts to use wind data products to simulate long time series of oceanographic variables. With sufficient confidence in the relevance of the ocean model physics, the discrepancies between simulated and observed sea level and sea surface tempcraturc pointed out shortcomings of the wind datasa. The cor-

....

.............

.

~-. . . . . . . . . . .

. . . . .

- ..-.
JANUARY 1990 C A R D O N E . G R E E N' W O O D A N D C A N E

...

. . - . .. . . . . -. .

..

0 -

I27

d o n of these errors is obviously important for studies of climate. In addition,mmhant ship,rrpom continue to be an essential mouiCe for .the development and calibration of more sophisticated wind product& such a those now produced by operational weather fores casting centers, or those soon to be available from the
spaceborne scatternmeters E R S l and NSCAT.

Acknowledgments. Our thanks to Virginia DiBlasi for help in preparing the manuscript. This work, a contribution to the NSCAT Science Definition Team, was supponed by JPL957647 from the Jet Propulsion L b a oratory.
REFERENCES

Merle. I.. and A. Morlicrc, 1988: Toward an o p c r d t i o d 3 dimrnd o d simulation ofthc tropical AUantic Oaan. Geophys. Res. k r . . 15. 7, 653456. Parker, D. E., and C. K. FoUand. 1988 The Nature of Climatic Variability, Mereor. Mag..117. 201-210. P c t m n . E. W.. and L. Haw. 1987: Did the Bcaufon wale or the wind climate change? 1.Phys. Oceanogr.. 17, 1071-1074. Picaul. I., 1. Smain. P. Leomte. M. Seva. S. Lukas and G. Rougicr. 1985: Climnrfc aa o rhe Tropical Arlanrfc Ocean wind stress lr f andsea nr&ceremperarure 1964-1979. 467 pp.. Univcnitt de
Bmagne Occidcntalc. B m l . Frdncc. . Posmentier. E. S. M.A. Cane and S. Zebiak. 1989 Tropical Pacific climate mndssincc 1960.1. Climare. Z 731-736. Ramagc, C. S.. 1984 Can shipboard mca%umunts reveal rmlar changes in tropical airheal flux. 1. C l i m e Appl. M t o . eer. 23, 187-193. , 1987: Secular changes in wned surface wind spc&,over the OCCM. J. C l i m e Appl. Mereor.. 16, 525-528. Rcvcrdin, G., and Y. Du Pcnhoat, 1987: Modeled surface dynamic height in 1964-1984: An effort 10 arvn how well thc low fitqvenciesinthequatorial AUanticwercramplcdin 1982-1984. 1.Gmphys. Res., 9. 1899-1913. R o q D. B.. V. J. Cardone. I. Overland. R. McPhmon, W. J. P i m o n and S. Y u 1985: Oceanic surface winds. Advances in Gmphysfcs, Vol. 27, Academic P 101-139. ki&x, 1989 Modelling wpical Pacific xa surfacevmperature: R. 1970-1987.1, Phys. Omnogr.. 19,419-434. Smain. J., 1. Picaut and A. Eusalacchi. 1985: lntcrannual and s a s o d a b i l i t y ofthc tropical A h t i c Oeean depicted by sixteen y a m ofxa surface temperdNrc and wind as Coupled OceanArmosphne Models, 16th Liege Collcq. on Ocean Hydrodp Mmics. 1. C. Nihoul. Ed., &vier. NY p. 211-237. - M. Sen, 1987:On itlationships benmn tropical AUantic ,and xa surfact lcmpcraturc, windnrru. and e o d precipitation indicer: 1964-1984. OceonAir Inremaion. 1. 183-190. Vcrploegh. G.. 1956: The quivalcnt velocities for the BcaufonCIt i m a i a of the wind force at sea. N c d d a n d r Meteomlogis&-h lnnitut Mededceingcn cn Vcrhandcinpn. Series A, No. 66.38 pp. [available from Royal Netherlands Mcmrological Innitute. D Bill. the Netherlands]. c Whysall, K. D. 8.. N. S. Cwpcr and G. R. Bigg 1987: Long term changes in the tropical Pacific surface wind field. Nature. 327. 116-2 19. W o o d ~ K S. D.. R. J. Slutr. R. L. Jcnnc and P. M. Stcurer. 1987: , A comprchcnsive mn-atmosphere data set. E d / . Am. .?o. &e, Sm.. 68, 1239-1250. World MeteorologicalOrganization. 1970 The Bcaufon S d c ofwind force. WMO Commission for Maritime Meteorology, M r n aie S i n AKain Rcpon NO..^. WMO, 22 pp. cr m - 1976: lntdnational list of rcleclcd, supplementary and.avriliay , ships. WMOIOMM-No. 47 [available from Sccmana ofthe 1 World Mctmrological Organization. Geneva. Switzerland]. Wright, P. B.. 1988: On the rcality ofclimatic changes in wind over the Pacific. J. Climnrol.. 8, 521-527.

-

-':"-a -.

Barnm. T.. N. Graham, M. Cane. S. Zcbiak. S. Dolan. J. OBrien and D. L aa 1988 On the prediction ofthe El Niao of 1986el, 1987. Science, 241, 192-196. Busalaahi. A. J., and J. 1. OBrien, 1981: tntnannual variabilityof the quatorial Pacific in the 196or 1. Geophys. Res.:86, IO. 901-907. Tekcuchi a n d 1.J. O'Er'ep! 1983 lnvrdnnval variability of the quat0mJ Panfic-Renutcd. 1.Geophys. Res., 88,75517567 .

0 .

Cane. M. A.. S. Zcbiak and S. Dolan. 1986: Experimental Forecasts ofEl N i h . Naflue,322,827-832. Cardone. V. I., 1969 Spcciiicauon ofthc wind distribution in the marine boundary layer for wave forecasting Rcpon TR69-I. NewYorkUnivcnify.NmYork.N.Y., 131 pp.INTISAD702 490 I. - J. Bmceoli. C. V. G m d and J. A. Grccnwwd. I980 ,A. Ermr c m of exuatropicalltorm wind fieldsspecified 'CI f o historical data 1. ofPerrol. Techno/.. 32,872-880. rm Dobson, F. W., 1981: Reofrefmncc height for and avc&ng time of surface wind mcasumncnD at sea Rcpon No. 3. Marine Meteorology and Related Ocranographic Activities. World Memrologid Organization. Available from Samariat of the WMO.Geneva, Switzerland. Goldenbcrg S. E., and J. J. O'Brien, 1981: Time and rpaa variability Mon. Wen. Rev..109. 1190-1207. a f v a o i d panfic wind Kavfeld L.. 1981: The dcvelopmcnt of a new &aufok equivalent d e . Mereor. Rundcch.. 34. 17-23. Latif. M., 1987: Tropical ocean circulation exprimenu. J. Phys. Oceanogr.. 17.2, 246-263. Large. W. G.. and S. Pond, 198 I: Open ocean momcntum flux measuremenu in modcrdv to m n g winds. 1.Phys. Ocmnogr.. 11. 324-336. Mrnaa, A.. 1987: Progmss toward an oarational orran model of the tropical P a c k at NMC/CAC, in Funherpwess in €SUOrorid Oceam#mphy, € J. Katr and J. M. Wine. Ed%. 439-450, Nova Univcnity Rar
~~

.

.
... .

. .-,

i

AN

INTERACTIVE OBJECTIVE KINEMATIC AWL\’SJIS SYSTEM
A.T. COS’, J.A Greenwood’. V.J. Cardone’ anr * R swall’

I

’OceanweatherInc. Cos Cob, Cl ‘Environment Canada Downsvlew. Ontano
1. INTRODUCTION

0

The need for highquality wind fields for ocean response models arises in hindcast studies of operational and eweme climate, in coastal and offshore structure design, and in forecasting for Ocean platform operation and ships. Ocean response models such as the third generation (3G) wave model (WAM) and the Oceanweather’s 3G wave model have shown great skill in producing nearly perfect hindcasu of signiricant wave height and peak period in severe tropical and ematropical systems when driven by high quality wind fields. The Surface Wave Dynamics (SWADE) study special Intense observational Period (IOP) of the October 1990 US East coan event put several uind fields using both objectively derived and handdrawn man-intensive wind fields through a common wave model (WAM 3 ) The results show (Cardone et.al., G. 1995) that the suite of hindcasu produced by very sophisticated purely objective analysis schemes was clearly beaten by hand4rawn kinematic analysis (Figure 1). Udonunately. lhis manintensive. tediously produced analysis took approximately 100 man-hours to produce a IO day hindcast which is a time frame clearly inapplicable to long term hindcast studies and forecasting applications. The Interactive Objective Analysis (IOKA) system w s developed at Oceanweather to combine the a advantages of manual analysis bolh shown during SWADE study and emphasized by Sanders (1990) and Ucccllini et al. (1992). with the speed of a purely objective analysis scheme in deriving high quality marine surface winds. Using the SWADE winds as a control. Oceanweather first dmloped the objective analysis algorithm. Seidel. for the exprcss purpose of analyzing wind fields. The interactive part of IOKA consisted of manual editingldeleting of wind inputs in ASCII format. This procedure worked well in SEAMOS

(Southeast . & Meteorological i and - ~ Oceanographc 3indcaS Study) where ships, typhoon mo& .urput winds and a background climatology WIN fields were combined using Seide1 10 achict high q d i r y wind fields for some 200 typhoom d molwons. While the procedure xnsiderably faacr that manualkinematic amws md yielded better ~ s u l t than s running pur: nphoon winds by including observations. ‘ & system nceded a final component: at. .aezictiv~graphical workstation. The Wind w,+iStation was dwloped to allow the u diqGp and manipulate the wind inputs , 10 kidel, p s wurk nation is already used operationally YI cktm-ther’s global 7 h ~ wind/wve f-sting servia and has been used in several h i d - S l studies. the mOSt recCnt k i n g the addition N‘ 10 YOrmS to the Canadian climate Center (CCQ EY1 Coast Storm Study (CCC. 1991; alw S \ d d. al., 1995). This paper will p e n t tsr’ steps involved in the IOKA prand ,+.&be the development and use of a graphical W a d WorkStation.
2. I ~ R A C I I \ ’ E OBJECTIVE KIIEMATIC ANALYSIS
2.1 Overview ’& hean of ~ h c O U system is the graphical I interface knoirll as the Wind WorkStation (wws). WWS is an analyst-friendly M S Windows b a d program (version 3.1, Windows 95 or Windo\il; NT) which allows the analyst to v i m a& manildate wind inputs for the objective analysis algorithm. The display is very’ flexible and allows the user lo both scroll and use a me zoom capabilit!’ (the wind barbs are redrawn to the kn possi~~tz resolution) to display any region of the basin. The analyst may also customize the wind inputs displayed by the WWS 10 plot o p t i o ~ linforoniwtion such as Significant. Wave. HeighL Peak Pen& Surface Pressure and

.a

I09

StatiodCall Sign Identification and may display any or none of the wind inputs (useful for a final check of the analyzed wind field). A selectable latituddongitude grid may be displayed with the data. and the final objective analysis wind field can be displayed from every barb to every 4th barb according to chc uscr's prefcrence. The program also suppons printing on a true Mercator projection with a fine resolution digitized coastline. The WWS can be s up very easily in any basin. a a d s u p p o n s any latituddongitude grid which is a submultiple of 2.5 degras down to .25 degrees. The latitude and longitude grid spacing need not be identical. which is very useful in northern latitudes where less resolution in longitude is desirable for computational spcod considerations. Cumntly. Ihe objective anal.ysis algorithm. Seidel. suppons up to 200 by 200 parallel grid (a .30 30 degree latituddongiludc a m With a .25 by degree resolution, 300 by 300 degree area at 2.5 degree molution) although this limit can be easily increased should the need cvcr arise. Typically. grids baween 60 and 70 parallels square are used as a m d e o f f betwv.cn resolution and compu~ationalspeed. The basic objective analysis method follows the approach of Ooyama (1987) by fitting quadratic forms to the velocity components and wind speed separately. minimizing the differenas bet\veen the analysis and the observations in the least-squared sense:

can g m w rather large) and more impoflantlg imposes no mtriaions on the time difference betwv.cn maps. -For instance. maps can be analyzed every 12 hours for a spin-up period, every 6 hours during the initial stages of a storm, then every 3 hours during the intense period: The resulting wind fields can then be time-interpolated to the desired time step for input into a wave model. This flexibility greatly decreases the time 0 the analyst needs 1 spend on spin-up periods and grcatly enhances hislher ability to do a fine time step analysis during the storm peaks. .This is also very uscful for long term operatio~l climate audies where long periods of inactivity can be hindcast with a larger time step and imponant storm events can use a shorter time aep.
2.2 Meteoroloeical Invuu

,

b!

'

:.

where q is the weight assigned to the inputs of class k Fk is a measurement of class k , is the .F analysis value at the location of the measurement. and B is a scale fanor which is used to achieve the desired lye1 of smoothing. The fitted velocity components are used to mover the wind direction only. the wind spud is directly analyzed (Cardone, et. al. 1993. see also Cardone and Grant. 1994). The WWS ws a flesible storm database file to contain all wind inputs and output (objectively analyed) winds. This provides a single source file for a particular aorm/hindcan period and is very convenient for archiving purposes. The WWS &CS no assumptions as to the IengIh of a particular h i n d m (though the storm database file

..
1111,:
i:;[#

....

The firn nage in the JOKA system is the preproassing of mettorological inputs. Typically wind observations from buoys, ships, off-shore platforms. coasral manned nations (CMANS). cloud uack winds. well apostd land stations and satellitederived scatteromcter winds are used in the analysis of the marine wind field. The WWS places no reminions on the number of types of data. or the inclusion of other trpcs of dam. Typically a pressurederived background wind field is also used, although this is optional if the data density is significantly fine (grid spacing dependent). The inclusion of other wind fields such as typhwn model output for tropical locations is also commonly done. All data to be brought into the WWS is first adjusted for stability and brought to a common reference level. tpically 20 meters. following the methodology developed by Cardone ( I %9; see also Cardone et. al.. 1990). Standard buoy wind measurements (usually 5 to 10 minute averages) are temporally smoothed to effective hourly averages. Averaging is done on meridional andzonal wind components of the wind to calculate the wind direction, and on scalar wind speed to recover the average wind speed. Buoy wind spceds derived by the "vector averaging" method are inflated to e 6 d v e "scalar-averaged" using the empirical relationship Asynoptic described by Gilhousen (1987). observations can bc optionally repositioned to onhour locations via moving centers relocation, which is essentially similar to the procedure that relocates aimaft flight lcvel winds to a moving v o n a . Asynoptic obsenalions can also be included without moving antcrs by giving them a

a

110

lower weight in the objective analysis scheme. and signifymg to the analyst that it is an asynoptic observation and should be given exua scrutiny to determine its representativeness in the wind field. All wind inputs are put into the W W S input format, so-called 'uvw' file. and brought in the Workstation storm database. Weights can be assigned to each type of wind input: common wind inputs such a5 buoys. ships. scattemmer winds. CMAN stations. typhoon model input and background prcssurederived winds can be assigned default weights in the objective analysis scheme which were determined by Oceanweather to be representative of the wiad's reliability. Typically, buoys gel a very high weight. while ships get lower weight in the objective analysis scheme. The analyst can also over-ride these standard default weights. if they are deemed inappropriate for a cenain dam type. Typo of winds are also assigned standard colon (although these can be customized for individual preference and display types), which is very useful for the analyst when all the data is plotted on the screen. 2.3 InteractiviN with the Wind WorkStation Once the wind data is incorporated into the WorkStation, it is displayed as colorsoded wind barbs (by type) over a coastline map on an .y plot projection. The wind field can be viewed as a full basin, or zoomed and scrolled to display any section. The analyst can 'point and click' on any wind ObseNaIiOn to bring up a test bos which displays the latitude. longitude. wind specd, wind direction and station identification of the wind observation and its neighbors. The analya has the ability to delete individual wind obun2tions: deleted data. displayed in a light blue color. can be undeleted if the analyst changes hislher mind. Usually quality control of & wind inputs is &nr at this step. although automatic quality control can be performed in the prepmasing step before bringing the winds into the WWS. The analyst typically uses the background wind field, handdrawn pressure cham, continuity analysis and other sources to determine the quality and reliability of each p i a of data. The most imponant feature of the WWS is the ability to add highly weighted Kinematic Control Points (KCP) to the wind analysis. This is the analyst's m s powerful tool in shaping the ot resulting wind field. With the KCP, the analyst can input and define the fine-s.de frontal

features. and add and maintain jet streaks and other features which have proven to be very important in exweme storm seas (ESS).and are oRen missed by purely objective methods. The analyst can use KCPs to define data-sparse areas using continuity analysis, satellite interpretation, climatolog)' of developing systems and other analysis tools. Winds can be run (put through the w i v e anaJysis) on an individual map for instant feedback to the analyst. or run for the entire length of the storm. When the w i n k are run interactjvefy (one map at a time) the analyst has the ability to add KCP points. run the winds. analyze the changes reflected in the final winds, and either make more changes or acccpt the winds as final. This.inieractjvity greatly enhances the analyst's ability to make changes to the wind field and boosts hislher confidence in the final wind product. 2.4 E m r t and Interwlation of the Wind Field Once the final wind field is run through the objective analysis scheme and accepted by the rm analyst, the final winds can then be exported f o the stormfife database. If the output of the WWS is not at a regular time step, or if a finer time step is required, a general time interpolation program, TIME-INTERP, is wed. This program can produce time interpolated wind fields on any time step. and can be optionally used with a file of moving centers to help preserve featurcs in the interpolated maps. Output of the time interpolator can be sent directly into a matching grid wave model. or put though a separate spatial interpolation program. WINDZWAVEGRID. which can place the winds onto any target wave model grid.
3. APPLICATJON IN THE CCC EAST COAST

STORM UPDATE STUDY
The IOKA system is currently being implemented in the-addition of IO m c n i norms IO the CCC Ean Coast norm population. The previous 68 storms were hindcast using the same handdrawn kinematic analysis technique chat was proven to give high quality winds in the SWADE study. In lhjs update study, the WWS was sct up on a area from 22.5' N to 77S0N and 82.5OW to O O E . Grid spacing was seluxed to be 1.25O in longitude and ,8333' in latitude. resulting in a 4489 grid point wind grid Figwe 4). A three-hour time a e p was

111

selected to do the wind analysis: this is also the time step of the wave model. Winds wcrc spatially interpolated to the CSOWM (Canadian Specral O a a n Wave Model) wave grid (Khandekar et. al.. 1991) using the WIND2WAVEGRID utility. Wind inputs for the 10 update storms include US and Canadian buoys, ships and CMAN stations. All data inputs are adjusted for height and stability to 20 meters neutral. The buoy observations are temporally smoothed to efiective hourly averages. Asynoptic data are not currently being used in this study. The background field used for this study is the ECMWF wind analysis for storms through 1994. and Oceanweather's wind analysis from its real-time global forecast for the February and April 1995 storms. Both background wind fields are on 2.5 by 2.5 degree grids, and both have had real-time observations already blended into the Hind fields. However. Occanweather's global winds have gone through the IOKA p and have had some analyst interaction in a forecast mode. Initial work on the April 1995 event has shown

the efficiency with which an analyst can produce a final wind field. and the analyst's confidence in the final wind fields delivered to the wave model. Additional tools such' as the general lime interpolation and spatial interpolation routines have allowed Ihe analyst to use flexible intervak . between maps, and easily port the wind output to any target grid. Development of and improvements to the Wind Workstation continue almost on a daily basis. owing to the number of current hindcasting and forecasting studies the system is being used on. As the system is applied 10 different basins, both tropical and extratropical, the need for new tools arises and most users' requests have already been implemented into the current system. Areas of future development include: the addition of a manipulative movinganten fable in the WWS which can used for repositioning of aspoptic data as well as in the time interpolation of wind fields: addition of continuity tools which would better allow the user to uack and smooth such weather features as fronts. uoughs. ridges, and jet nrraks, looping of final wind fields in a movie sequence for final check of the continuity of the wind fields: and contouring of the final wind fields.
5. REFERENCES

0

the WWS to be a time-saving tool in the analysis
of the winds. The analyst was able to complete the analysis of the wind field in less time. due lo the ability to view all the input and output winds together on one display, and the ability lo run winds interactively to achieve a final wind product. Funher significant time savings were also achieved by not having to manually grid and enter a kinematic winds fields by hand. which had been done in previous hindcasts. While some kinematic sketches were done on printouts of the wind field. most work was done dinctly on the WWS. Time histories (Figure 5 ) at two Canadian buoys (44138 and 44141) shou, good agreement between the measured significant wave height and the hindcast wave heights using the CSOWM 3G shallow wave model. These wave time histories are equivalent IO those c x p x e d uith handdrawn kinematic analysis.
4. SUMMARY AND FUTURE DEVELOPMENT

Canadian Climate Center. 1991. Wind wave hindcasl extremes for the east wasl of Canada. Volume 1. Prepared under conuact no. KMI69-66678 by Macbren Plansearch Ltd., Halifax. N.S., and Occanweather Inc.. Cos Cob. CT 109p. plus appendices. Cardone, V.J.. 1969. Specification of the wind distribution in the marine boundary layer for wave forecasting. Report TR49-01, Goophys. Sciences Lab.. New York University. 137pp. (Available from NTIS:Order AD 702 4-90), Cardone, V.J.. J.A. Greenwood, and M.A. Cane, 1990. On trends in historical marine wind data. 1 , of-, 3, 113-127. Cardone, V.J., LA. Greenwood, 1993. Final Repon: Development and Initial Test Againsl SWADE of an Interactive Objective Kinematic Analysis Program for Marine Surface Wind Field Analysis. Submitted to U.S. Army Engineer Waterways Experiment Station.

0

IOKA Vstem has proven to be an effective and time-saving tool for the analysis of marine surface winds. It successfully blends the man-intensive kinematic analysis with the speed of a purely objective analysis. The development of the graphical Wind WorkStation has increased both
112

Cardone, V.J.. H.C. Graber. RE. Jenscn. S. Hassclmann. and M.J..Caruso. 1995. In Search of the True Surface Wind Field in SWADE IOP-I: Ocean Wave Modeling Perspcctive. Acccpted July IO; 1995 to Global Atmo. and Ocean Svs. Cardone. V.J., and C.K. Grant. 1994. Southeast Asia Meteorological and Oceanographic Hindcan Sludy. Offshore Southeast Asia 10th Conference and Eshibition. 6-9 Detember. Singapore. Cardone. V. J.. and V.R Swail 1995. UnceMiny in Prediction of Esveme Storm Seas (ESS). Fourth International Workshop on Wave Hindcasting and Forecasting. mober 16-20. 1995. B a d . AlbeM. Canada. Khandekar, M.L.; R Lalbeharry and . V.J. Cardone. 1994. The performance of the Canadian spectral Ocean wave model (CSOWM) during the Grand Banks ERS-I S A R wave spectra validation esperiment. AlrnosDhere&ean. 32.3 1-60,

Gilhouxn. D. B.. 1987. A field evaluation of NDBC moored buoy winds. J. Atmos. Ocean.

Tech.. 4. 94-104.
Sanders. F., 1990. Surface weather analysis over the oceans - Searching for the s uuth. Wea. a Forecasting, 5. 596-612. Swail. V.R.. M.J. Parsons, B.T. Callahan and V.J. Cardone. 1995. A &sed enreme wave climatology for the east mast of Canada. Fourth International Workshop on Wave Hindcasting and Banff. Forecasting. October 16-20. 1995. Alberta. Canada. Uccellini. L.W., S.F. Corftdi, N.W. Junker, P.J. lo, Kccin, and D.A. O s n 1992. Repon on the Surface Analysis Workshop held at the National Metcornlogical Center. 25-28 March. 1991. Bull. Amer. Meteor. h, 495412. 73.

< ;
'5

--

113

Buoy 41001

- Owl Winds

SUOV 4iW1

-

NMC Winds

BUOY 41001

-- ECMWFWinds

Buoy 41001

- NASA Winds

BUOV 41001

- FNOC Winds

Figure I . Comparison of Wave Heights derived from five objective analysis winds (NMC, ECMWF, NASA, MOC, UKMO). and Oceanweather’s (OWI) hand-drawn kinematic andysis winds during

SWADE lOP2.

I14
I
..

.

.

1ntektive.Kinernatic Analysis Flow Chart
Prepantion of Wind Inputs: All Data Adjusted for Height B Stlbillty, Buoys temporary smoothed. Appropriate Weights Aasigned

*

I 4
Background Field

1

Graphical Wind WorkStation

Pualky Control d W n d Inputs

4ddWon of Klnematlc Control Polnts

Run Objective Analysis

Typhoon or other Model Output

Wind Time Interpolation

Wind Spdlal Interpolation

Wave Model

Figure 2.

I15

Figure 3. Wind WorkStation sample display.

0

0

a

Figure 4. Final Wind Barbs in the April 1995 CCC Storm. Wore: Winds are allowed to fall off in the Baffin Bay since the basin is enclosed by ice.)

117

.=

...~ ~.. .. - .

-

.

.

.

:

. . .. .

. ..

~

. . . .. . . .

..

Comparision of Significant Wave Height Hindcast (3G Shallow CSOWM) During April 1995 Storm

- Measured
0-0 Hindcast

E
v

l5

Buoy: 44138 - CSWOM Grid Point: 2788

E

0
L I

L

I

I

I

I

m

Apr 5.1995

Apr 6,1995

l5

1 Buoy: 44141

____

Apr 7,1995

______

Apr 8.1995

_-

0 '

I

,

,

I

I

I

I

Apr 5.1995

Apr 6.1995

Apr 7.1995

Apr 8,1995

D

Figure 5. Wave Height C o m a " During April 1995 Slorm.

e

The Global A'rmosphere curd Ocean System, Vol. j ( l - 3 ) . pp 107-150

IN SEARCH OF THE TRUE SURFACE WIND FIELD
IN SWADE IOP-1: OCEAN WAVE MODELLING PERSPECTIVE
Viricent J. Cardone. H n C. Graber, Robert E. Jensen, as Susanne H&ehann and Michael J. Caruso
'?

Gordon and Breach Publishers

-

.. .

.

- -. . .

....

I SEARCH OF THE TRUE SURFACE h
WIND FIELD I SWADE IOP-1: OCEAN N WAVE MODELLING PERSPECTIVE*
VINCENT J. CARDONE', HANS C. GRABER', ROBERT E. JENSEN', SUSANNE UASSELMANN" and MICHAEL J. CARUSO' 'Oceunweother, Inc. Cos Cob, Connecricur, USA. 'Rosenstiel School of Murine a d Atmospheric Science, Unioersiry of Miami. Miami. Florida. USA. USAE Waterways Experiment Srarion. Cocutal Engineering Reseurch Center, Vicksburg. Mississippi, USA. 'Mux-Planck lnstirur f i Mereorologie. Hamburg. G m n y . 'Woods Hole Oceanographic Institution Woods Hole, Mllssachuserts, U S A

'

An imponant expcrimcntnl objeniV. of the Surface Wave Dynamia Experiment (SWADE) was to measure the surlaa meteomlogy with sufadent accuracy to vinually eliminate the mntribution of errors in wind fields to simulations carried out with the latot numerid spsnral ocean surface wave models T i study seeks such a wind M d through the evaluation against the cxtcnsirc SWADE measured wave hs dam bas.= of wave hindcam of the fim Intensive Observational Period (IOP-I) made with Ihc WAM model using six alternative wind ful& b of ths wind kids .um.takcn fmm ths standard d h e objerrvc analysis pmdvm of numerid wther pmjicrion Brwq centen (ECMWF. FNOC UKMOI, a fovnh W M O and IiM (NASA) fmm rpedal high-molution objcaive rranalM of d time and n o n d rimc mcreodoEid dam while Ihs d (OW/AES) was &rind by manual kimnanc rranalysis udng all mnventional and rpsdal SWADE mncOrc. logid dat% The high rsolution reanalysis pmducu (OW/AES, NASA) wett found to fit thc buoy wind okrvations uwd m rhc analysis rather d d y (typicauy man diflncnu m wind rpad of about Urn-'. I MdiRcrrna of about 1 M - I and rrancr id of about 12% 1 % However. diRemuu bwm p'dded na 7b fields and buoy winds werc r y p i d y rwo u) thtu dmcr lpcatcrin Ihc NWP center wind fi& ~n though the buoy dam ~ 1 pramably ufilbcd in ths o b w v e analysis pmocn. The manual analyses wrt drarly s superior u) the objativc analywr m maintaining mnLinuiIy ofthe U m e - w a evolution of Ihe major kinematic fea~res Iht fry0 major storm s m which a o s d the array during IOP-I. of y m The WAM (Cydc4) deep-water physics model hindcast l o d by the manual kinematic (OW/AESI wind fields provided distinctly p a l e r skill in specification of time Wries ol significant wave height (SWH) at the buoys moored in deep and intermediate water depths in the SWADE area. than hindcase driven by any of the o b j d v e wind 6cl& The mean difference of 0 2 m and scatter index of 14% found are about half those normally reported in.prsvious studio ol verification of continuous wave time wries h i n b t with WAM. Thas rcsulu indicate that the OWIAES wind field of SWADE IOP-I is accurate cnough to allow toting and evaluation ol alternative model source terms,

-

KEY WORDS 0 -

wave

model. WAM. operational and dncmatic wind-fields, SWADE

1. INTRODUCTION

The primary scientific goals of the Surface Wave D n m c Experiment (SWADE) yais arc to improve understanding of the dynamics o the evolution of the surface wave f
.Contribution to Surface Wave Dywnisr Exprimcnt
107

/

108

V. J. CARDONE ct a ! .

field in the open ocean, and determine the effects of waves on transfer processes at the air-sea interface, on mixing processes in the upper ocean, and remotely sensed . microwave signatures. A detailed description of the scientific objectives and the - experimental strategy and objectives of the field phase of the experiment is given by f Weller et.al. (1991). The field phase was camcd out o f the mid-Atlantic coast oier a &month period beginning October 1, 1990. Numerical spectral ocean wave modelling plays a dual role in SWADE. On the one hand, though greatly improved models have been developed within the past decade (e.g. WAMDI-Group, 1988). uncertainties remain in the underlying physical processes modelled, &penally those which transfer energy from the atmosphere to surface waves, and which dissipate energy within the wave field. Therefore, such models will benefit from improved understanding of fundamental physical processes gained from the analysis of the SWADE field data. in the form o more accurate f parameterizations of model source terms. On the other hand, a wave model may serve as a powerful tool in the analysis of wave measurements from the array of moored buoys in the inner SWADE array (so-called FINE domain, set Figure 3, for example by providing a space-time context to the site measurements, by resolving the external basin-wide forcing on the wave field observed within the SWADE array, and by serving as a test bed for proposed new parameterizations developed from new knowledge of wave dynamics resulting from SWADE. It is well established that uncertainties in the wind forcing are the largest source of errors in model generated wave fields in operational implementations of contunporary wave models (e.g. Janssen et al., 1984; Cardone and Szabo, 1985; Bauer et al, 1992), and are usually so large as to mask errors associated solely with deficiencies in the wave models physics or numerics. In recognition o the need to prescribe the f wind forcing for the wave modelling component of SWADE as accurately as possible, SWADE, unlike previous field experiments of this type (e.g. JONSWAP, ARSLOE, LEWEX), included a major meteorologica1,component to augment the conventional network of buoys from the National Data Buoy Center (NDBC) with additional moored buoys which acquired surface wind measurements and at some locations, surface fluxes. In addition, buoys traditionally used to measure only directional wave spectra were equipped with meteorological sensor packages as well (Weller et al., 1991). The special SWADE meteorological array provided not only an enhanced synoptic data network, but also a basis for evaluation of conventional surface wind fields, and the potential to provide the surface wind forcing at much higher resolution than possible from conventional marine data, especially in the immediate vicinity of the wave and flux measurements. or so-called SWADE FINE domain (the three domains of SWADE: FINE, REGIONAL, BASIN will be described below. see also Figure 5). Most of the additional SWADE surface marine meteorological data were transmitted over the Global Telecommunications System (GTS) in real time and therefore were more or less available for assimilation into objective analyses prepued at NWP centers. SWADE also included t h m separate post-analyses of surface wind fields for special Intensive Observational Periods (IOP). The first IOP. bctwetn 20-31 October 1990 was in many ways the most interesting of the t h m IOPs defined in SWADE. The several storms which formed in the waters of the US East

0

0

0

M SEARCH OF THE b U E WIND FIELD

Lo9

.-. I ,.

1:
.

Coast (so-called SWADE REGIONAL domain) and moved across the SWADE FINE domain during this IOP provide a wide dynamic -range of meteorological forcing within a relatively quiescent BASIN&IC meteorological regime. he most which occurred on 26 October, was energetic enough severe storm during this IOP, to sink the SWADE central spar 'Spike" buoy moored east of Chesapeake Bay at' the peak of the storm, and to cause significant damage along the middle East Coast (Morris, 19914. This study investigates the d e g m to which the objective of SWADE to provide accurate wind forcing was achieved through the evaluation against the extensive SWADE measured wave data base, of wave hindcasts of IOP-1 made with the latest version (that is. Cycle4; see Janssen, 1992) of the-WAM model using six alternative wind fields. This basic approach of using simulations provided by a skillful wave model evaluated against highquality wave measurements to assess wind field quality has also been followed recently, for example, by Guillaume and Mognard (1992) and Khandekar et aL (1994). T h m of the wind fields were taken from the standard real-time objective analysis products of numerical weather prediction centers including the European C a t e r for Medium Range Weather Fo-ting (ECMWF), the U S Navy Fleet Numerical Oceanography Center (MOC), the United Kingdom .. Meteorological Otflce (UKMO). Two wind fields were produced at the NOAA National Meteorological Center (NMC) and at the Goddard Space Flight Center (GSFC) of the National Aeronautic and Space Administration (NASA) from a special high-resolution objective reanalysis of real time and non-real time meteorological data The sixth wind field was derived by intensive manual kinematic reanalysis (produced by Oceanweather Inc and supported by the Atmospheric Envi-.. ronment S e M a of Canada (OW/AES)) using all conventional and special SWADE meteorological data The wind fields were compared to the measured winds in the SWADE array o f the'middle-Atlantic Easi Coasr, and to each other. while the f alternative wave hindcasts were evaluated against the extensive wave measurements acquired by the conventional NDBC buoys moored in intermediate and deep water o f the US East Coasr, and the SpeEial SWADE buoys moored in the SWADE FINE f domain. We show that the wind field derived by kinematic analysis (OW/AES) is clearly superior to the others and of sufficient accuracy to allow the further refinement of the parameterization of the physics and numerics of wave models. This paper is organized as follows. In Section 2 we give a concise description of the meteorological regime of IOP-1with emphasis on the storms of 24 and 26 October. Section 3 describes the alternative wind fields, including sources, rnethodology and characteristics. Section 4 gives the strategy of the modelling experiment; that is, we describe the particular setup and configuration of the wave model used. the way the alternative REGIONAL scale wind fields were combined with one BASIN scale wind field to provide boundary conditions for the wave run$ how the runs were made and what parameters were utilized in the evaluation. Section 5 gives the main results of the study, which is an assessment of the aciuracy of the altemative wind fields in terms of the accuracy of the resulting wave hindcasts. using mainly the results of the hindcast error analysis against buoy measurements. Finally, in Section 6 we discuss the main results and assess their significance in advancing the objectives of SWADE

I IO

V.

I. C A R W N E ct nl

2. SWADE IOP-I (20-31 OCTOBER 1990) . 2.1. M e w e d Data The SWADE measurement array, shown in Figure 1, consists of both the existing network of buoys and C-MAN stations maintained by the NDBC and the special buoys deployed for SWADE. The latter include four 3-meter discus buoys (44001 [D-NorthTJ,44015[D-Enst], 44014[D-CERq and IMET). four MiniMet meteorological buoys (MET-I, MET-2, MET-3, MET-4). the Brookhaven spar buoy (SPAR). Unfortunately, the spar buoy sank after the peak of the storm of 26 October. It was recovered in June 1992 but no useable data were retrieved. Three of the four MiniMet buoys (MET-I, MET-2. MET-4) began to drift at different times during the experiment as a result of the severe storms (Caruso et al., 1993). Eventually, two of the MET buoys (MET-2, MET-4) were captured by the Gulf Stream and drifted out of the SWADE area. MET-1 reanchored twice in the steady southwestward drift but still remained within the SWADE area. IMET was a special 3-m discus buoy provided by Woods Hole Oceanographic Institute. It also succumbed to the storm

42'

yrm

.

P

.
8

I

40'

_/-

__-/

38

36

34

-76

.74

-72

-70

Fl(iURE I

L w t i o n map OlSWADEsrpsrirnmlal domain.

IN SEARCH O F T H E TRUE WIND FIELD

111

L

of 26 Octobcr, sunaining considerable damage. The Coastal Enginrring Research Center (CERCJof the US Army Corps of Enginan deployed the 2-m discus buoy at 44014. The sensor and systcm.capabilities of the NDBC C-MAN stations and buoys are well documented (e.g. Steele et al.. 1992). All such buoys transmit once per hour, values of wind speed and direction (two sensors), atmospheric pressure, air and water temperature, significant wave height, average and dominant wave period. and at some sites, frequency or directional wave spectra. The special 3-meter discus buoys deployed specically for SWADE included, in addition to the NDBC payload. a special motion sensor package and K-Gill anemometers which sampled continuously at 1 HZ and stored data internaly on five 120 MB laser discs. Anemometer heights, sampling interval and period varied among the sites. These are summarized in Table 1. The typical averaging interval for the hourly wind speed and direction is 5-8 minutes: however, at 41M)1,44004,44008,44011 and DSLN7, the wind was sampled continuously and reduced to continuoti series of 10-minute averages before transmission. Directional wave data were acquired by buoys 44001,44015 and 44014. Nondirectional wave series were measured from buoys 41001.44004.44008, 44009.44011,44012,DSLN'I. CHLVZ, BUZM3. Data from the MiniMet buoys were not utilized in the wind field reanalyses because the locations of the drifting buoys were not known at the time those reanalyses were carried out. Of course, in addition to the core data base noted above, a comprehensive base of conventional data and analyses has bem assembled and is archived in the official SWADE data base maintained at NASA Wallops Right Facility (Oberbolaer and Donelan, 1995). including NOAA surface weather analyses, satellite cloud and AVHRR imagery and SST analyses, conventional ship report and landbased synop tic data. There were no data acquired from research aircraft during IOP-1, though considerable airborne data were acquired in later SWADE ~ P s .
TABLE I Summary of marine windlwavs mCaSUmmt rites in SWADE a m y during IOP-I.

Station

Hull

Payload

Position LAT LONG 73.0W 13.6W 70.6W 68.6 W 69.5 W 74.6W 66.6W 74.6W 74.8 W 73.6 W 71.0 W 75.7 W ' 753 W 80.7W

Anrmo- Sample meter intmol
height

Sample
priod

Wave E(0

Measuremcnt

F(f.0)
N
Y

0
-

41001 44001 Moo4

44005

44008 44009
44011

44012 44014 44015 BUZM3 CHLVZ DSLN7 SVLSl Hull: b N

* den to six 10-minute samples

-

bN NDBC 3-D SWADE bN NDBC NDBC bN NDBC LNB LNB NDBC NDBC bN LNB NDBC CERC 3-D 3D SWADE C-MAN NDBC C-MAN NDBC C-MAN NDBC C-MAN NDBC

6 m NOMAD. 3-D

-

34.9N 38.4N 385N 427 N 40.5N 3&5N 41.1 N 38.8N 36.6 N 37.1 N 41.4N 36.9N 352N

5.0m 3Jm 5Dm
5.Om 13.8m 13.8m

I ys I ys I s a I ys I yc
I ys

I hP
8min

I h? I hr.
1hP

Y Y Y

Y
Y Y

3.8 20

3k Dkcus, LNB

-

5.0m -13.8m 3.5 m 3.5m 43.3111 43.3m 46.6m 329m

I
1 1

yc
ys

I I I
1

ys ys
yc

I hP I h P -. I hP 8 min
Bmin I h?

Y Y
Y

N N N N N N

.

YE
ys

I

ys

I hP I hP I h?

Y Y Y Y
Y

Y Y
N N N N

Large Navigational Buoy

I12

V. I.

CARDONE e! d.

2.2.

The Synoptic Weather Conditions of IOP-I

0

In this section we provide a general description of the principal weather systems which -.affected the SWADE arca during IOP-1 in terms of both conventional sea-level p,ressure systems and surface wind and wave measurements in the SWADE array;Our description of the time and space evolution of the pressure and wind systems is based mainly upon the OW/AES analyss, which as will be shown below, in esscnce correctly link the &ace meteorology with the observed wave response in the SWADE area. Three main systems affected the SWADE area directly during the period. In general, because storms in the open North Atlantic were not well developed in this period, wave energy from distant systems in the form of swell was generally slight but not negligible. One source of swell. especially for the southern pans of the SWADE area, was tropical storm N a m centered near 30 ON. 65 W on 20 October. This system moved slowly in a clockwise loop and weakened in the following days but was probably strong enough to serve as a source of swell which the buoys detected in the array at the beginning of IOP-1 with height of 1.5-2m. and peak period of about 9 seconds. Figure 2 shows the isobar analyses at 12-hourly intervals between oo00 UT 23 October and 1200 UT 27 October, the period during which the two most energetic cyclonic systems of IOP-1 crossed the measurement area. These high resolution analyses were drawn by hand as pan of the OW/AES wind field analysis process to be described below. Only ,over the area shown in this figure does the data coverage justify the I-mb isobar andyses. The first of the three systems to directly affect the SWADE area formed near the South Carolina Coast early on 23 October in .the base of a large trough of low pressure which approached the US East Coast..Thc center of this low moved northeastward at an average spced of 25 knots over the next 24 hours as its central pressure decreased slowly from 1005mb to 997mb. Figure 3 shows the continuity of the center and the associated frontal system. A narrow band of strong southerlies diveloped in advance of this front. Maximum winds of 2Oms-' were measured at DSLN7 (1-hour average at 47m) and 17ms-' at 44015(8-minute average at 3.5m). As the intense pressure gradient east of the front propagated over the cooler shelf waters north of the G l Stream. an area of stable thermal boundary layer stratificauf tion, measured maximum surface winds were much lower: 12ms-' at both 44001 and 44004(8 minute average at 5m). Maximum observed s i w c a n t wave heights (SWH) generally vaned between 3.5 m and 4.5m in the SWADE array in this event. The second and most severe of the IOP-1 systems began to evolve on 25 October near the South Carolina c o s t at the trailing.edge of a frontal zone/shear axis left i n the wake of the previous system which, by then, had moved northeast of the Cana&an Maritima. This shear axis was well defined in conventional weather maps add most objective analyses. However, the details of the cyclonic development off the North Carolina coast between oo00 UT 25 October and oo00 UT 27 October were not well defined on conventional analyses and were illuminated only after considerable post-analysis (as pan of the OW/AES analysis) to define the formation and continuity of the multiple cyclonic circulation anters which propagated along the shear axis.

0

'0

/'

I SEARCH OF THE TRUE WIND FIELD N

I13

The conventional anal$- of the cyclone center and frontal S~NCIUR mainly followed the classical Norwegian frontal cyclone life-cycle model but this system appeared to possess features of the revision of the Norwegian model described by Shapiro and Keyser (1990). In particular the shear axis seemed IO develop the

.. .

FIGURE 2

I I4

V. 1 CARDONE e1 PL .

FIGURE 2 Sca level pressure analyses (isobars are d r a m at I millibar intervals) a1 t2-hourly inIeyYais from U-27Onobcr. 1990.

.

"bent-back warm front" characteristic of the new conceptual model with multiple cyclone centers developing along this front. The post-analysis of this system resolved two principal centers which propagated along the shear axis as shown in Figure 4 (labeled "LOW 1" and "LOW 2"). The two centers appeared to deepen at about the same rate on the 26th (as also observed in another case of East Coast cyclogenesis during the ERICA projccq Cardone. 1992) though on the 27th the trailing center attained the lower central pressure and was associated with the stronger pressure gradients. The central pressure in that system lowered from 1005 mb at 1200 UT 25 October to 986mb at I200 UT 26 October and 982mb at 1800 UT 26 October. The maximum deepening rate therefore is slightly less than the arbitrary threshold of 24mblday of 'explosive" development or "bomb" cyclones (Sanders and Gyakum. 1980).

IN S E A R C H OF T H E TRUE W l N D F I E L D

115

. -

FIGURE 3 High-resolution s level p ~ u analysis (isobars drawn at I mb intervals) at 00 UT 24 u m October 1990 with time-space continuity ofslonn csntcn and frontal boundaries for23-25 October.

0
.

Unlike the quasi-circular flow about a single idealized extratropical cyclone. the bent-back warm front and the preexisting elongated shear axis led to the development of highly linear and energetic surface wind flows. North of the shear axis. northcast flow (direction 040 degrees 10 degrees) extended from just offshore Nova Scotia to the North Carolina Coast. South of the shear axis the flow was persistently west-southwesterly. Wind speeds in the northeasterly flow o f the midf Atlantic coast in the SWADE array increased rapidly f o less than 7ms-' at 1200 rm UT 25 October to 20-25 ms-l (measured at 5 m height) at most buoys by 1200 L ! T 26 October. The maximum measured SWH in this flow was &Om associated with spectrai p k k period (TP) 12.5s at 1800 uf 26 October at 44015. Buoy 44014. in of shallower water, recorded its maximum SWH of 6.0m (spectral peak period of 11.1 s) at the same time. Of the measurement sites shown in Figure 1, only 41001 experienced the flow south of'the shear axis and cyclone tracks The maximum wind speed measured by 41001 in that Row was 15ms-' at 1300 UT 26 October (I-hour average at 5 m) with associated SWH of 4m.

.

.

.
f

..

.,

I16

V. I. CARDONE e1 d.

m

sw

FIGURE4 High-rrrolution sea lsvcl p m u r c analysis at Kl UT 24 October 1990 and position or frontal boundary and the space-umc evolution of two low p-urc system3 from 25-21 October over the SWADEdornain.

It was not until about 1200 UT 26 October that the shear axis was "breached" when the strong northeast flow accelerated southward then southeastward as it came under the influence of the intense tightly curved cyclonic pressure gradient trailing the center of LOW 2. The surge-like character of this flow was revealed in the continuous time series of wind measurements at 41001 on 26 October, when surface winds (1-hour average at Sm) backed from 210 degrees at 13ms-I at 1700 UT to 280 degrees at 19ms-' at 1800 UT and 300 degrees at 24ms-' at 1900 UT; SWH peaked at about 9.0m at this buoy several hours later. It was at about the same time that the spar buoy sank. According to the captain of the tanker MS Tropic Sun, which passed by the SPAR at 1600 EDT inbound to Delaware Bay, the buoy was floating upright and with all its sensors intact. ARGOS transmission stopped at about 2000 EDT (C.Flagg. personal communication). The SPAR was located on the sea bottom by the RJV Oceanus on 3 October. Data f o an 0 rm acoustic survey indicated that the resting place of the SPAR was in 301.9m water depth about 3 km southwest of the original deployment position in 2OOm depth

IN SEARCH OF THE TRUE WIND FIELD

117

( r b r and Boutin, 1991). In mid-June 1992 the USNS GRASP recovered the Gac
SPAR and evaluations are underway to determine the cause of its failure. The third organized system to affect the SWADE array (not shown in Figure 2) in IOP-1 was associated with a wcak cyclone and associated frontal system which moved rapidly east-northeastward off the New England coast on 29 October. The main effect of this system was to induce first a fresh southwesterly wind f o then a lw fresh to strong northwesterly wind field over the SWADE ana. in which there was still running a swell left over from the second system. This third system, therefore, created a regime which allowed the detection and study of wave generation in a shifting wind field in the presence of background swell. The time histories of measured wind speed and SWH at the SWADE buoys generally peak at between IO and 131x1s-' and between 3 and 4x. respectively, in response to this weather system. 11

3. SWADE IOP-1 WIND FIELDS

Six different wind fields were utilized in this study. Thrn of these wind fields are products of the operational NWP centers at ECMWF, FNOC and UKMO. T e e hs wind fields (except for UKMO)and preliminary objective reanalyses of the SWADE IOP-Iwind field produced by NMC and GSFC/NASA were compared in a prelimirbr nary study reponed by G a c et al. (1591) which also included alternative wave hindcasts with Cycle-3 of WAM. Those preliminary results revcaled large difierenm between the wind fields and as a mult intolerably large (for the objectives o f SWADE) difierencs betwen hindcast and measured wave height time s e n e in the central part of the SWADE area. Following that study, a sixth wind field analysis effort was commissioned, resulting in a wind field (OW/AES) based upon a blend of manual and objective analysis techniques. TaaMe 2 gives a summary of the properties of the six wind fields. The wind fields are specified on various domains and grids. The thrn specific domains defined for SWADE modelling studies are shown in Figure 5. Only the
TABLE 2 Wind fields used 10 fora WAM Cycle4 in SWADE IOP-I.
Saura

Method

Variables spatial

Rcrolution

Reference

-Pod
I-hourly
&hourly
'

!
Cardone et al
(1980)

OWIAES ECMWF

moc
UKMO NASAIGSFC NOAA/NMC

Kinematic analysis Optimum intsrpolation Optimum interpolation Optimum interpolation
Suamive

u...
U. .
7

05 dcg

."

I.fZS deg

Shaw ct al. 11987)
-.

1.25 deg

&hourly

Barker (1992)

UW.,
ut0

1.5 dcg La1 ?.-hourly 1.875 dcg Long 025 deg 3-hourly
'

Bell and Dickinron (1986) Barker et al.
(1984)

codon Conditional relaxation

U,,

0.3 dcg L t

&hourly

Gnnmill(1991)

05 dcg Long

I18

V. 1 CARWNE et a. . l

ECMWF wind field was utiliad on the SWADE BASIN domain, and was wd to drive thc BASIN wave modd for BASIN spin-up Wore IOP-1and to provide boundary conditions for the REGIONAL wave model domain during the IOP-1 period. . (0) OW/RES: This wind field was developed using the procedures routinely applied by Oceanweather. Inc (OWI) for specifying wind fields in series of historical events for use in hindcasting the extreme wave climate. This approach, first described by .Cardone et al. (1980)attempts to blend the advantages of classical manual analysis and objective techniques.' The advantages o manual analysis for f marine surface weather analyses have been recently emphasized by Sanders (1990) and Uccellini et a (1992).and then is indecd a trend at NWP centers to maintain !

SWADE Model Grid Nesting

FIGURE 5

Hierarchy of gtid nesting implnnmted lor SWADE wave simulations

IN SEARCH OF THE TRUE WIND FIELD

I19

and facilitate the use of hand analysis through the use of analyst-friendly graphical interfaces implemented on powerful workstations (e.g Kocin et al., 1991; Swail eial., 1992). The OW/AES system consists of the following basic steps:
1. Data P~processing:The preprocessing of the stream of coded synoptic observations consists of basically three parts: (a) filtering of mislocated. miscoded or otherwise grossly erroneous data; (b) quality control of ship reports of sea level pressure; (c) adjustment of all measurements and observations of the marine surface wind to a common averaging interval, reference height and to "equivalent" neutral boundary layer-strati6cation. The data stream filtering was n d e d in part to avoid the use of the MET buoy data which found their way onto the data stream in real time, mislocated, and in some instances with the wind spced units indicator incorrectly coded. These data may have contaminated the analyses of the ECMWF (A. Guillaume, pcrsonal communication) and possibly other mal time objective analyses. Following procedures discussed by Sanders (1992). ship reports of sea level pressure were crossed checked against buoy data wherever possible and shipspcdfic barometer biases were determined. The intent o the OW/= analysis is to m l v e the 3ynOpUc xale" wind fjeld at f three-hourly intervals on a grid of spacing 0.5 deg in latitude and longitude covering the SWADE REGIONAL domain. To minimiz the &ea of 'mesoscale" variations in the buoy measurements, hourly measurcments of the 8.5 minute average wind an smoothed to crTeaive hourly averages by averaging t h m succoSive measurements with weights 1/4,1/L 1/4. For the buoys which obtained continuous ICMninute averages, an hourly average wind is obtained directly by averaging 6 su&ve 10- minute averages. The averaging is done on meridional and zonal wind components of the wind to derive average wind dirrniop and on the scalar wind spad to compute average wind speed. The adjustment of measured average wind speed to a common reference level is carried out following the procedure originally suggested by Cardone (1969; see also Cardone et. al.. 1990). The procedure uses stabilitydependent surface wind profile laws and an assumed dependence of the roughness parameter on wind speed, to calculate an "effectiveneutral" wind s p e d at a reference height. The effective neutral wind speed, U, is that "virtual" wind speed which at height z, imparts, in neutral thermal surface boundary-layer stratification and for the assumed drag law, the same stress as imparted by the measured, U,, measured at height I, in a thermally stratified boundary-layer.

where L is the Monin-Obukov length and Y is the "profile' stability function (for a more detailed description of the iterative procedure used to calculate U, from the t h r n known quantities: measured wind speed. measurement height. and air-sea temperature difference, see Cardone, 1969; or Ross et al., 1985). OW1 conveniently use the effective height of 2Om. so that buoy winds can be compared in an analysis to winds reponed by ships. Cardone et al. (1990) describe a consistent way to refer ship reports to a common level and find 20m as the most logical level for ships because Beaufort estimates are most directly converted to equivalent wind at 20m.

i m

V. J. CARWNE CI d.

and the average height of anemometers on ships sc-equippcd is about 2Om. Before the OW/AES wind fields were used to drive the WAM model they were' reduced to .~ 10m level by multiplying the wind speed by 0.93. the Thete are two additional subtle and related assumptions implied in the adjustment procedure. The first is that the roughness parameter does not depend on wave age, but a number of recent studies (Maat etal., 1991) suggest that it does. To account fop this effect would have required coupling the wind analysis and wave hindcast models. This was not done for the alternative wind fields produced at the NWP centers and to.maintain consistency was not carried out at this step for this analysis. However. as described below, weak coupling is assumed in the WAM wave model, following the approach of Janssen (1991). The second assumption is that normalizing the effect of thermal stratification through the friction velocity. as i implied in the adopted procedure, also accounts for s the effect of stabdity on the wave generation p r o m itself. However, as suggested by Kahma and Calkocn (1992) scaling of wave growth with fetch in terms of nondimensional variables scaled with u as opposed to U appcars to be insufficient to , , fully explain the observed effects of thermal stratification of the surface boundarylayer on wave development with fetch. As the first system of IOP-I crossed the SWADE are& the boundary layer was unstably stratified in the southern part of the array and stably stratified in the northern part of array. This case therefore may indicate the sensitivity of the wave generation procss to unmodelled stability effects.

2 Boundary Layer-Model Surface Winds An initial estimate of the 3-hourly wind field in the REGIONAL domain is made using a marine planetary boundarylayer model (MPBL) applied. to grid-point specific gradients of sea level pressure and air temperature, and the air-sea temperature difference. Since its development (Cardone, 1969) this particular model has been w d at many national centers for routine analysis and forecasting and in many studies (e.& Overland and Gemmill 1978; Cardone 1991). For this application, the required MPBL variables were derived from carefully hand analyzed and digitized maps (e.g see Figure 2) drawn to carefully maintain continuity of centers of action and fronts. Despite the presumably more accurate resolution of the MPBL input data allowed by the SWADE data and hand-analyses in this application, the resulting MPBL winds, when compared to the measured buoy winds, were found to be specified no more accurately than in the cited prior studies. For example, Table 3 compares the results of this application of this MPBL in SWADE and the composite of published comparisons reported by Cardone (1991). These results suggest that a scatter of about 3ms-' in wind speed and 30 degrees in wind direction is about the intrinsic limit of skill in specification of synoptic scale surface wind fields over the ocean using simple analytical steady-state MPBL models. To reduce this error in a surface wind analysis scheme requires the assimilation of measured surface wind data.
-.

3. Kinematic Analysis: MPBL winds were retained over the entire domain only for the first 48-hours of IOP-1. Thereafter, and over a more limited domain, MPBL winds are simply superseded by winds derived through a kinematic analysis. Streamlines and isotachs were manually drawn and subsequently handdigitized at 0.5 degree latitudeflongitude and 3-hourly intervals in the sub-domain shown in

IN SEARCH OF THE TRUE WIND FIELD TABLE 3 Summary of wind m n derived from prrviow numerical cxpcrimcnu in wave hindcasting Invclitgator/Expcrimcnr N Ob% Overland & Grmmill(l977) Grmmill er d (1988) Cardonc(1992).ERICA IOP-2 SWADE IOP-I (MPBL)
NA.

121

.

Wind Sped
Mcar DIE.
~~~

Wind Direction
Scattcr(rmr) Meas. Dif Scaitcr(m)
~ ~~

-0.50

20 4

N.A.
351

-0.10

,290
3.10 267

539

-

0.22 0.4 I

N.A. N.A. 7.3'
11.1'

N.A.

NA.
38.2328'

_,..."

. .. -,. j_..,.

.

Figure 6, which shows examples of some of the maps for 26 October, the peak of the major IOP-1 storm. This sequence shows the strengthening and migration of a distinct wind maximum in the northeasterly flow from a position offshore DELMARVA at 0600 UT(maximum ~,,,of19ms-'), to a position cast of the North Carolina .coast at 1200 UT(maximum winds of 24m-') then to the area just west of buoy 41001 at 1800 UT(maximum wind 29ms-I). The imposition of the detailed spacetime continuity of major kinematic properties of the wind field is possible becaw the subjective analysis p r o w utilizs information contained in the full time histories of hourly winds from buoys and coastal C-MAN stations more effectively than most objective analysis procedures, which assimilate mainly observations at synoptic analysis time only and furthermore mume all measured winds refer to an assumed fixed height (usually 1Om). Of course, real-time analysis, as opposed to hindcasting, cannot use the full history of systems.to develop continuity, but can refer only to past history. The three-hourly wind fields are interpolated to hourly intervals using an objective temporal interpolation algorithm designed to minimize smearing of translating kinematic features such as shear zones and centers of circulation.
(6) ECMWF The wind fields provided by the ECMWF are the 10-meter level u and D component fields produced operationally at &hourly intervals. These fields are from the so-called ECMWF-TOGA advanced operational surface and diagnostic fields data set. The analyses are a product of the global three-dimensional optimal interpolation objective analysis method based on all available data (Lorenc, 1981) followed by non-linear normal-mode initialization to remove gravity wave oscillation. The analysis is carried out on constant u-surfam (u is a vertical coordinate in pressure, p, defined as the ratio p/p, where p, is surface pressure). Low level winds (buoys, ships, land stations etc.) are assimilated at the first u-level = pJ to update the analysis on u-coordinates. The wind field on the first u-level is taken over the sea to represent the 30m level and a stability PBL parameterization is used to link the lowest level analysis, the 10m level and the surface stress. Relatively little direct verification of ECMWF marine surface wind fields ag&nst highquality measurements has been reported. Since 1987. the ECMWF data have been used to drive a global WAM model on a 3 d e g m grid. Zambrcsky (1989) compared one year of hindcasts (December 1987-November 1988) against buoy data in various basins. For buoys off the East Coast of the US,the mean wind speed difference is 0.07ms-' and scatter index.is 30% (scatter index is the m difference normalized by mean measured wind sped of verification sample).

i

122

V. J. CARWNE et aL

..

FIGURE 6 Sequence of aunosphcric flaw from 09-21 UT 26 October 1990 dunng SWADE IOP-1 in tcrms oiarumlmnc and irorach, (knolt).

(e) FNOC: The particular wind field file obtained from the US.Navy FNOC is the file-of wind stress resolved at spatial resolution of 1.25 degms in latitude and longitude and temporal resolution of 6 hours. These fields arc derived basically in a two-step process. which form part of the FNOC Navy Global Analysis and Prcdiction System (NOGAPS; Barker 1992). First, a single-level non-linear variational analysis is used to assimilate low level winds into a first guess-field. This step utilizes measured winds from ships, buoys, island and coastal stations and low-level cloud wind estimates. All observations are assumed to be at the 19.5m level. The first guess field is a blend of the previous 6-hour wind forecast (in the tropics) provided by the FNOC NOGAPS forecast model (Hogan and Rosmond, 1991) and the

IN SEARCH OF THE TRUE WIND FIELD

I23

geostrophic wind computed from the &hour NOGAPS sea level pr&sure forecast (outside the tropia). In the second part of the pr, the analyzed fields arc passed through the NOGAPS boundary-layer parameterization to calculate stability dependent surface stresses. Using the surface s t r m (as friction velocity) a logarithmic profile is used to extrapolate winds upward to 19.Sm to provide the effective neutral wind speeds.

(4 NASA: The Goddard Space Right Center (GSFC) hindcast the wind fields using a high-resolution version of its global objective analysis system. This system has been used in global climate simulations, a large number of observing system simulation experiments, in studies designed to assess the impact of, various remote sensing systems (e.g. Baker et al.. 1984) and in studies (e.g. Duffy and Atlas, 1986; Bauer et al, 1992) of the impact of SEASAT scatterometer derived surface winds on surface marine analysis and forecasts. Thanalysis system is based upon the Successive Correction Method (SCM), in this case applied on multi-level constant n pressure surfaces. The vertical increment i pressure between layers is SO millibars. Surface data arc assimilated on the 1000-mb surface. For SWADE IOP-I, a first g u m field was taken from the standard NOAA Global Data Analysis System (GDAS) which has a grid spacing of 25 x 2 5 degrees. These analyzed fields arc interpolated to a grid of SO km spacing for the NASA analysis. The SWADE buoy data were reduced to hourly averages of effective neutral wind (as in OW/AES) and then assimilated on the analysis grid via the SCM in seven successive scans, during which the area of influence of each observation was successively reduced from 500 km to I20 km. The time interval between wind fields is 3 hours.
(e) NMC: The method used to provide surface winds at NMC is described by Gemmill (1991). The starting point is the standard 2.5 x 2.5 d e w wind field analysis provided by N O M S Global Data Assimilation System (GDAS) at the lowest model u-layer wind (about 45m height). These winds arc reduced to 10m logarithmically. Statistical verification of the operational 10m wind fields at NMC (Gemmill, 1991) indicate a negative bias of 0.6ms-' and nns difference of 2.1 ms-'. The NMC SWADE analyses were camed out on a grid of spacing 0.3 latitude x 0.5 longitude. The fitting method is called the Conditional Relaxation Analysis Method (CRAM). In CRAM, which is applied to each wind component separately. the wind field is considered to.obey the solution of Poisson's equation with the Laplacian of the first guess field as the forcing function, and observations as internal boundary conditions. The analysis considered buoy and ship data but with ship observations given lower weight This method has also been explored by Cardone (1969). The fit of these analyzed winds and the buoy winds a n descxibcd below. Gemmill (1991) has attempted to determine the error of the NMC wind fields using wind data which was not used in the analysis, or effectively independent da& He reported mean and differences for one selected IOPyl map of -1.1 ms-' and 2.4ms-l, respectively. between the analysis and IO ship reports withheld one at a 0 time in 1 repetitions of the analysis.

( J ) UKMO: Thesc wind fields arc computed from the analysis cycle of the coarse-mesh, 11-layer general circulation model in operational use at the UKMO

a :

.

. .

.

124

V. J CARDONE n PL .

(Bell and Dickinson, 1986). The model grid covers the globe with a mesh with grid lengths of 1.5 in latitude and 1.875 in longitude. UKMO also runs a limited-area .fine mesh version of the model which produces hourly wind fields on a grid with half
O

'the mesh size of the global grid. T h a e were not used in this study since over the North Allantic O c a n the UKMO fine grid extends only as far south as 50 degrees. which is not sufficient to cover the entire SWADE REGIONAL domain. The UKMO analysis is basically an optimum interpolation assimilation of obwrvations on a &hour cycle into a first guess field generated by the most r e n t forecast The analysis is carried out on constant u-layers. There are 15 such layers. Surface marine wind observations are assimilated on the lowest a-layer surface only, the mid-point of which is only about 30 meters above the surface. The wind specified for this level is used to drive a boundary layer parameterization to provide surface fluxes. This parameterization assumes a constant roughness parameter for the sea of 10-'m (this corresponds to a 10m drag coefficient of 1.2 x lo-'). The surface stress is used to extrapolate winds to 19.5 meters following a logarithmic profile.

4. THE NUMERICAL OCEAN WAVE MODELLING EXPERIMENTS
4.1. The WAM Model

A brief summary of the WAM model is given here. A more detailed description of the basic WAM model is given in WAMDI-Group (1988). The version of the model implemented here is the so-called Cycle4 version of WAM. or WAM-4, in which the atmospheric boundary-layer is coupled to the wave model following Janssen (1991). In WAM, the evolution of the directional wave spectrum F(f.S; 4.2, t ) as a function of frequency,f, and direction, 8, in spherical coordinates defined by latitude, 6, and longitude, A, is determined from the integration of the energy balance equation
-+(cOS~)-'-(c,cos4F)~-(c,f)+-(c,F)=S,+S,, at ab ai a8

aF

a

a

a

+s,

(2)

Here c,. c,, and c, are the appropriate deep water group velocities along a great circle path. The thrce source terms consist of S an empirical wind input function , based on the results of Snyder et al. (1981). S, the nonlinear energy transfer integral ,, S following Hasselmann et al. (1983, and , the dissipation due to white-capping waves from Komen et al. (1984). WAM-4 incorporates the source terms dcscribed by Janssen (1990). Namely, the input is quadratic in the ratio of friction velocity to wave celerity, uJc( f),and the dissipation is proportional to the fourth power of the frequency. The wind input is given at standard height, usually 10 meters, and the surface stress is calculated internally within the wave model as a function of both wind speed at height and stage of wave development The present implementation has 25 frequency bins logarithmically spaced from 0.042Hz to 0.41 Hz at intervals of Af/f=O.l and 24 directional bins 15 degrees apart.

M SMRCH OF THE TRUE WIND FIELD

IZS

Deep water physics only was considered in the propagation and the source terms.

Thee rsults therefore should not be compared to measurements from SWADE
buoys moored in water depths which can be considered to be shallow for the wave periods excited.
42. The Hindcast Rwrs

,. . .

- 0 ’:

WAM4 was implemented on a twc-net grid system to represent the BASIN and REGIONAL. SWADE domains (scc Figure 5). The BASIN grid covers the entire North and South Atlantic Oceans with a grid of 1 d e g m spaclng T h e BASIN grid hindcast with WAM-4 was run first and only once, using the ECMWF &hourly wind fields as input This run was started at oo00 UT 15 October and run to oo00 UT 20 October to provide overall spinup of the domain and a set of directional spectra to i i i l z the interior of the REGIONAL grid at oo00 UT 20 October. ntaie The BASIN run was continued through October 31 to provide directional spectra along the boundary of the REGIONAL model throughout its run. T h e REGIONAL model covers the domain shown in Figure 5 with a grid of 0 2 5 degree spacing in latitude and longitude. Directional wave spectra are supplied from the BASIN grid to the w t e r n and southern rows of points of the REGIONAL grid at 1200 second intervals. T h e spectral components arc interpolated bi-linearly in space from the BASIN grid points to the REGIONAL. grid points, except of course for coinciding points. Wave spectra on the boundary of the REGIONAL grid are temporally interpolated to the REGIONAL model time step ( A t = 240s). The REGIONAL run was spun up at oo00 UT 15 October and run to oo00 UT 20 October with input generated from the BASIN run. The REGIONAL wind fields were kept constant during the wave model integration over the same time intervals ( i s 1.3 or 6 hours) at which they w m supplied (see Table 2) S x diBercnt simulations werc made of the REGIONAL model each driven by a i different wind field and spanning from 20 to 31 October. Model directional spectra were saved at hourly intervals at the four grid points surrounding each measurement site. For comparisons of model output with measured data at buoys, model derived parameters were simply averaged over the four grid points, spanng 0.25 degree. surrounding the buoy. The archived hindcast hourly directional wave spectra were reduced to the following integrated propertis of the spectrum:
HI,, = 4[

rl

W.B)dfdB]L’*

(3)

126

V. 1 CARDONE CI d. .

8 = parabolic fit to ,

ae

=O
(Yamartino. 1984)

(7)

e

uo= sin-'(&)[l +0.1547&']

where
E

= Ji - [ m ( ( y

+ a)q

An additional output variable of the model is the friction velocity, . and the . 1 10-meter drag coefficient, C,,, both of which are affected internally in W.AM-4 by the source term integration. From thesc output variables, the 10-meter wind speed may be derived simply as

u,, = -

A

(9)

The U,,derived this way does not neccssarily coincide with the U,, provided to WAM by the alternative wind field sources. However, the differences weie found to be extremely small. Figure 7 makes this comparison for the OW/AES wind fields at the four buoys for which the buoy, the nearest WAM model f d e g m grid point and the wind grid +-degree grid point are virtually collocated. In view of such small differences, for consistency and convenience all comparisons of model and buoy wind speeds used the wave model output derived hourly 025degrrc winds.
5. EVALUATION O F ALTERNATIVE HINDCASTS

.

A direct evaluation of the accuracy of the wind fields is of course hampered by the absence of high quality wind data which have not already been assimilated into the wind fields as part of the analysis process. The buoy wave measurements, however, were not used in the wave hindcast process and therefore may be used to objectively evaluate the accuracy of the alternative wave hindcasts. The relative skill in the wave hindcasts may then be considered to be an independent measure of the skill in the forcing wind fields, since there is no reason to believe that the WAM model favors wind fields produced by any particular wind analysis system. Nevertheless. it is useful to compare the traditional statistical measures of difference between the buoy winds and wind fields interpolated to the buoy locations,as well as the differences between hindcast and measured wave parameters. Therefore. at each of the buoy and C-MAN sites listed in Table 1 the following statistics were computed on IO-m wind speed and direction, significant wave heighht. peak spectral and mean wave period:

IN SEARCH OF THE,TRUE WIND FELD

127

Buoy 41001

Buoy44004

5

5
0
10 20 OWIAES Wind Input [ d s ] Buoy 44015 25

-

.(, -.

OW/AES Wind Input [ d s ]
Buoy 44001

5

5

3

5
10 20 OW/AES Wind input [ d s ]

5

5
10 20 OWIAES Wind Input [ d s ]

FIGURE 7 Comparison or IO-m wind r p d derived from W A M 4 u. and input OWIAES 10-m wind rpctdatsollosatsdgrid painun~buoy~41M)1.44014.44015and44M)l.

Mean measured (Buoy)and hindcast (Model) Bias (hindcastmeasured) Mean absolute differma Root-mean-square error (RMSE) Scatter index Correlation coeffcicnt Slope and intercept of linear regression
where(

B = (E).&? = ( M )
(M-B) (IM - BI) ( ( M B)’)’’’ RMSE

B

-

M = a + b+B

...) refers to arithmetic mean.

128

V J. C A R W N E ci al .

For wind direction the following measures were computed:
Mean angular difference . .
Standard deviation of angular difference where
E

e

u8= sin-'(e)[l +0.1547~']

= 4 1 - [(sin A

sin A 8 = (sin (, e
COS A

- 0,)

+ (cos A e)q

e = (cos(e, -e,)

The peak and mean .wave periods of the measured buoy wave spectra were determined in the same way as the corresponding model parameters (4 and S. For ) consistency in comparing the mean wave period, we added a parametric high-frequency tail, as done by the WAM model, to the last frequency bin of NDBC spectral wave data such that

EV) =

E(A,)L$]-'

for

/>A,.

Time histories of the same variables were also plotted over the period oo00 UT 22 October to oo00 UT 31 October, and the comparisons were displayed as scatter diagrams for each variable, these plots showing also the linear regressions. Figure 8 shows the scatter diagrams for eight deep water buoys and for just the OWjAES wind fields. The skill in the WAM wave hindcasts driven by OWjAES winds as shown in this compatison is exceptional. In the remainder of this section we will restrict the comparative analysis to the eight buoys which lie in relatively deep water within the SWADE REGIONAL domain, since the deep water version of WAM was used for these hindcasts. Hindcasts with the shallow water physics included in WAM are reponed in a separate paper. The buoys included here are, in order basically from south to nonh: 41OO1,~l4,44015,44OO1,44004,44008,44011.44005

Figure 9a compares the alternative hindcast-measured time histories of wind
s p e d and direction at buoy 41001. This buoy was located well east of the track of the first IOP-1 cyclonic system on the 24th and south of the tracks of the multiple centers ir'the main IOP-1 event on the 26th. Despite the fact that the observations from this buoy were available to all analysis systems, there are remarkably large differenas in analyzed and measured histories of wind speed accompanied by very small differcnccs in general between analyzed and measured wind diwrions. OWjAES and NASA track the 10-m winds closely while ECMWF and N M C underestimate wind speeds by up to Sms-* during the main storm on the 26th. FNOC, ECMWF and UKMO winds are also lower than measured by u p to 5 ms-I during the event on the 24a.

IN SEARCH OF THE TRUE W l N D H E L D

129

P - 30 Cctober, 1990

I

."

22 30 October, 1990

-

I

O :

5 10 15 20 Measured U10 [&I

9 1 0

? i

I

I

P 30 Ocmber, 1990
1, 0
12,

-

200 300 Measured Theta [dq]
100

22 -30Oadmr. 1990

1

OO'

5 Measured lis [ml

10

Ak 4

I
6 6

1 0

12

MoaYRed Tm[s]

FIGURE 8 S a t t c r plou of OWlAES hindan and m ~ u %ind(adjurred to l&m ncuull).W o n d diKmnoc wave height and mean wave pried at 8 ofLshorcbuoys

The alternative wave hindcasts a n compared at 41001 in Figures 9c and 9d. The ECMWF. NMC and UKMO hindcasts are generally biased low in SWH during the energetic events. FNOC also underspecifies the peak SWH in the main storm but is much closer to it than ECMWF, NASA, UKMO or NMC, though the hindcast peak occurs about 12 hours later than the measured storm peak. The NASA hindcast simulates the first event on the 24th and the weak system on the 30th rather well but underestimates wave heights during the main storm on the 26th. The OW/AES hindcast follows the data closely throughout, except for a slightly too rapid decay of SWH following the first storm peak on the 24th. Figure 9d also indicates a tendency for hindcast TP to be lower than measured. Where the SWH of a windsea is underpredicted this 4 to be expetvd but this tendency is sea at almost all other sites evaluated and even during periods when SWH is well hindcash suggesting a wave-model based source of t i effect. The sucass of the OW/AES hs hindcast relative to those made by the other wind fields. coupled with the remarkable slrill of the OW/AES hindcast in an absolute sense, provides strong evidence that errors in the alternative wind fields arc the overwhelming source of errors io the alternative hindcasts of the S W H fields in the vicinity of this buoy, and, as will be shown below, throughout the SWADE area as well.

..

130

V. J. CARDONE el d. Buoy 41001 OW1 Winds

-

Bwv 41001 - NMC Winds

25

,

24 26 October 1990

28

Buoy 41001

- ECMWFWinds ..
U

BuOv 41001

- NASA Winds

U

a
() I

15

E 10
5

.
82 --

.i

-

- . . ..
28 24 26 Ocmber 1990 28

24 26 Onober 1990

Buoy 41001
25

- FNOC Winds

Buoy 41001 - UKMO Winds

() I

2 ='

24 26 Onober 1990

28

24

26

28

October 1990
FIGURE 9a

IN SEARCH OF THE TRUE WIND FIELD
Buoy 41001

151

- OW1 Winds

Buoy 4 0 1 - NMC Winds 1 0

October 1990

Buoy 41001

- ECMWF Winds

4 K4
Buoy 41001

Ocmber 1990

- NASA hnds

- .h b e r l!
I

28

28

IO

Bwv 41001

- FNOC Winds
3300

Buoy 4 0 1 - UKMO Winds 1 0

..

9

.z 200
0

C

0 :

5

z

p 100

5

.

. a

24

26

28

Ocmber 1990

FIGURE 9b

132
BUOY 41001

V. J. CARDONE e! SI.

- Owl Winds

~~~

z 6

s
5
0

- 4

F m
Buoy 41001

- NMC Winds
* :

.> 6 m

s
9 24 26 Cdober 1990 28
0

- 4

s
C

9 2 C m

2 $2 24

E

E

m

.

0

%2

26

28

October 1990

Buoy41001

-1 €
-10

- ECMWF Winds

-

Buoy 41001

- NASA Winds

1
k

-10

E

c 6
- 4 c
C

..
i-9 .

s
3 E
C

e 6

s E
E

- 4 E

2

m

" L
28

2

.-.

ca

* :
- UKMO Winds

0

24 26 October 1990

24 26 October 1990

28

BUOY 001 41

- FNCC Winds

Buoy 41001

m E
October 1990 October 1990

FIGURE%

/-

IN S U R C H OF THE TRUE WIND FIELD
Buoy 41001

133

- NMC Winds
5

3

E 1 0..-. $ -e. >

0 . 9

1

. ......m

-...

s P
2
I

5

24

.-26 28

24

26 October 1990

28

k -

Oaober 1940

Buoy 41001

- ECMWF Winds

Buoy 41001

- NASA Winds

0
P

m

24

26

28

24

26

28

October 1990

Oaober 1990

Buoy 41001

-

FNOC Winds

BUOY 41W1- UKMO W&

. u %
24

2 6

28

24

26

28

Odober 1990

Ocmber 1990

( h t or

FIGURE9 a Comparison of model hindart (solid line) and buoy m u s u m m u (+) at buoy 41001 Hati-) in IOP for IOm wind sped. b Plots of wind direction. c Plou of significant wave height d Plou of spectral p a k ways prid

134

V. J. CARDONE et al.

.

Figure 10 shows time history comparisons for buoy 44015. This .buoy experienced the strong southerlies east of the first system and the strong northeasterlies north ofthe tracks of the main storms on the 25-26th. A I L d the special SWADE discus buoys were eqposcd similarly to these systems, and the comparisons for 44015 to be described here are very similar to those at 44014. and 44001. Figure 10b shows that. again, wind direction is well specified by all sources during periods of significant forcing. The OW/AES winds arc faithful to the wind speed measurements throughouS all other wind fields more or less underspecify winds during the storm on the 24th. with smaller although systematic departures during the storm on the'26th. with a tendency for the analyzed wind speed peak to lag the measured peak by 6- 12 hours. The NASA wind field appears to have ignored the buoy data in the event on the 24th; it then tracks the buoy closely during the buildup of the storm on the 26ti-i only to underestimate the storm peak. The wave height and period comparisons in Figure 1Oc and 1Od show that all but the'OW/AES hindcast greatly underspecify the SWH associated with the first storm; SWH is underpredicted in the second storm by the ECMWF, NASA, NMC and UKMO hindcasts but reasonably well hindcast, albeit with lag, by FNOC. The OW/AES hindcast is uniformly skillful, and within f I m of the buoy SWH throughout. The tendency for all hindcasts to underspecify TP seen at 41001 is also seen at 44015. Site differences reveal but do not .necessarily define the spatial distribution of errors. For example, any type of objective analysis system for wind analysis may be tuned to fit observations closely at nearby grid points by admitting resolution of very small spatial scales in the analysis. In effect the error in the analyzed fields is minimized near the observations at the expense of increased errors in regions devoid of observations. Such a strategy, however, would be ill advised if wind fields are to be used to drive a wave model, which responds sensitively to the entire space time evolution of the wind field. Mere inspection of alternative fields i not much more s indicative of spatial error patterns. except that in the case of the OW/AES, the errors in the time histories are so low that errors in the spatial fields within at least the SWADE FINE domain must be comparably low. To some extent therefore, the alternative fields may be evaluated against the OW/AES wind fields and the wave hindcasts derived therefrom. For example, Figure 11 displays the alternative wind fields at 2100 UT 26 October. The differences in the spatial fields between the alternative winds are remarkable, panicularly between OW/AES and NASA, which place the most energetic surface winds in completely different parts of the circulation. In the NASA wind fields maximum wind spetds are placed in the broad south-westerly flow east of the storm center. In the OW/AES wind field the maximum winds are placed southwest of the center, consistent with the time-space continuity analyses of 'the kinematic feature. Most other wind fields also place the maximum winds southwest of the storm center but with weaker wind spnds. The wave height fields for the same time are shown in Figure 12 The OW/AES run resolves a single SWH maximum southwest of the storm anter. In general the other runs elongate and weaken the SWH field west of the storm. Even larger differences between the runs arc seen east of the elongated shear axis with NASA

0

'

0

..
IN SEARCH OF THE TRUE WIND FIELD Buoy M o l 5
135

- owl

Wlnds
I

Buoy 44015

- NMC

Winds

..

24 26 Ocmber 1990

28

.
*__

Buoy 44015

- ECMWF

Winds
251

Buoy 44015

- NASA Winds

I

u
'1

m

10

c

. .4

24

26

28

Oaober 1990

Buoy 44015

- FNCC Winds

Buoy M o l 5

- UKMO Winds

25

I
.. . ..
24

251

I

e

.
26 28

.. ..
22
24 26 ,28

October 1990

October 1990 FIGURE IOa

0 :

/,

.

..

136.

V. I. CARDONE et al.

Buov 44015 OW1 Winds
9300

-

Buoy 44015

- NMC Winds

5

E

p 100
24 26 28 24 26 28

October 1990

October 1990

Buoy 44015
9300

- ECMWF Winds

BUOY 44015 NASA Winds

-

I

Y

0

m

5

P 100

&-

I
- 24 =

26

28

6

26

28

October 1990

October 1990

BUOY 44015

- FNOC WIMS

Buoy 44015

- UKMO Winds

C

. .

5
U

e

24

26

28

0 1

22

*

October 1990
FIGURE IOb

- I 24 26 Ocmber 1990

I
28

0
.Buoy 44015

IN SEARCH OF THE TRUE WIND FIELD

137

F o lJ I

- OW1 Winds

Buoy 44015
-10

- NMC Winds -

c

- 4

E 2
m

E 2 C
m

24 26 Oaober 1990

28

24

26

28

October 1990

..
Buov44015

- ECMWFWinds
m 6
C

E
Oaober 1990

..
' t
4 -

- 4

5 2

8 &

*.

24

26

28

Oaober1990

Buoy 44015

- UKMO Winds

FIGURE IOc

. . :..

138

V.1 CARDONE et al. . Buoy 44015 OW1 Winds

-

Buoy 44015 NMC Winds

-

.

1 5

m

a
24 26 Oclober 1990 28 24 26 OClOber 1990 28

Buoy44015- ECMWFWinds

Buoy 44015 NASA Winds

-

m

24 26 October 1990

28

24 26 October 1990

28

Buov 44015

- FNCC Winds

Buoy Mol5 - UKMO Winds

m

24 26 Cember 1990

28

RGURE 10 a Companion 01 model hindat lsohd I m l and buoy measurements I-) a1 buoy 44015 (SWADE D l w u €aril 8" IOP l r IOm wnd speed. b Plots 01 wnd drrection. c Plolr olugnoficanl wave o hnghr d Plots olspenral peak wave pcnod

I SEARCH OF THE TRUE W N D FIELD N

lj9

. m

n

.o I

FIGURE I I

Cornpariron ofalternauve wind fields lor 2100 UT. 26 Oaobcr 1990.

and ECMWF placing the highest SWH in this region. Unfortunately there are no high quality observations of sea state in these regions to help evaluate these altemative solutions. It has become standard practice in recent studies of this type to evaluate wave hindcasts in terms of standard site-specific mean and NdiRerences ktwetn model derived wave series interpolated to buoy locations and the measurements

140

V. J. CARDONE et al.

.

.

......... ......... ........... ...........

10

10

u
a
1 ,

. ,
a
3,

m

m

......... .........

........... ...........

Y

-. FIGURE I2 Comparison of alternative SWH fields for 2IM) UT. 26 October IYW. Vectors repfcxnt magnitude of SWH and mean wave direction.

themselves. For standard integrated properties of the wave spectrum, such as SWH and TP. the measurements are not of course error free. For a typical 20-minute sampling interval, statistical sampling variability alone imparts a scatter of about k 12% in SWH and about k 9% in TP (Donelan and Pierson, 1984). Additional errors, specially in heavy seas, may arise in calibration errors and mooring

.-...

IN SEARCH OF THE TRUE WIND FIELD

141

-

interference on buoy vertical 'displacement, though for the SWADE buoys these latter effects appear to be negligible. Table 4 gives a concise summary of hindcast-measured difference statistics for wind spe=d and direction, significant wave height and peak spectral period. The table gives the mean difference, root-mean-square difference and scatter index (based on m)for scalar variables and mean angular difference, standard deviation of angular difference and percentage of standard deviation with respect to 3-54 ",calculated over the 8day period of IOP-I at each of the eight "deep water" buoys located within the SWADE FINE domain. At most buoys these statistics are based upon about 200 hourly buoy measuremmts. The "measured" wind speeds are the -effective neutral" winds at 10-m derived from the buoy measured average wind spccd and air-sea temperature difference, smoothed over three successive observations as described in Section 2. The measured SWH and 'IT are similarly smoothed in an attempt to reduce sampling effects. The aggregate mean wind speed differences given in Table 4 are negative for all wind fields except OW/AES. The ECMWF and UKMO have the largest negative speed bias and also the largest negative bias in SWH at 0.6 m and 0.8 rn. respectively. The m difference in wind speed is lowest for OW/AES (at l.Zms-') and less than half that of the real-time N W P products. The NASA wind fields fit the buoy winds much more closely than the N W P products but not as closely as OW/AES. The aggregate mean difference in wind direction varies from + 4 "in UKMO to - 1.3 "in NMC, while the standard deviation in wind direction varies from 22 for OW/AES to 39 O for FNOC. The aggregate mean difference in SWH is within i 0.4m for four of the wind fields, except for ECMWF and UKMO. NASA exhibits the smallest aggregate mean difference in TP and is the only wind field source to produce a small positive mean difference in TP. However, even for NASA, TP is biased slightly negatively in the SWADE area, and positively north and east of the SWADE area. The general skill in the wave hindcasts of'the principal properties of the spectrum, SWH and TP, are shown in the scatter indices (SI) given in Table 4. The OW/AES winds provide the lowest SI for SWH at all sites; at most sites, the SI i near the s lowest (about 15%) which is to be expected even of *perfect" hindcasts evaluated against this type of data. This agreement in SWH, of course, does not preclude .larger systematic and random erron in more detailed properties of the spectrum. The OW/AES wind speed SI is also generally lowest though it is matched at some sites by other wind fields. For example. at 41001.44014,44001 and 44005,the NASA wind speed SI is comparable to OW/AES. As noted above, site-specific wind statistics are not necessarily indicative of spatially coherent errors, and this seems to be the case for NASA winds at least, which exhibit comparable wind speed SI as OW/AES, whereas the NASA SWH-SI are generally larger than OW/AES SW.H-SI by about a factor of two. Differences between model runs narrow considerably for TP-SI. This suggests that some additional factor. besides wind errors is strongly influencing this property. Possibilities include wave model error, or importation of error from the basin scale domain, where all hindcasts were driven by one wind field (ECMWF). Figure 13 shows the scatter index as bar graphs for the main IOP-1 storm period only and just for four buoys, two (41001.44015) located within the wind data rich

SWADE verificnlian o 0 winds. SWH and TP a1 8 deep wale1 silcs lor 22-31 Oclober. 1990. Scnller index (SI) i s based an mi dillerenm lor scalar variables. f S lor wind dirslion is d e h h ai rinndard dsvidnn alllie mptilar dikcrcnce divided by 30. I 6'
SOllrcc

TAOI.E4

P

Variable
BIAS

41001

44014

44015

RMS
23 0.17 0.14 2.99 31 0.91 1.56 3.57 19 0.97 1.36 1.43 21 0.68 1.62 2.34 10 0.74 1.18 3.19 34 0.92 1.57 44MI

S.I.
11% 6% 14%

CORR

BIAS

RMS
1.02 I5 0.38 1.12 2.20 24 0.82 1.52 2.41 11% 0.69 1.62 1.27 I7 0.56 1.24 IS9 20 0.45 0.91 2.88 22.7

S.I.
12% 4% 16% 13% 25% 7% 35% 17% 21% 29% .18% 14% 5% 24% 14% 18% 6% 19% IO% 33% 6% 47% 19%

CORR 0.98 0.97 0.87 0.92 0.96 0.82 0.89 -3.3 0.89 0.76 0.97 0.92 081 0.97 0.97 0.92 0.94 0.91 0.73

OlAS

RMS

S.I.

. CORR
0.99 0.98 0.88
0.90

OWIAES

WI

U["lr-']

SWll[ml

.ECMWF

FNOC

0.11 2.9 -0.12 -0.22 1.93 2.3 -0.68 -0.54 -0.83 -0.6 -0.18 -0.10 -0.58

1.M

0.98 0.98 0.90. 0.87 0.94 0.82 0.69 -5.5 0.82 0.85 0.96
0.90 0.77 0.90

-

13%
33% 9% 34% 18% 4nv. 11% 37% 16% 16% 6% 26% 18% 26% 8% 28% 16% 35% 10% 35% 18%

-3.0
NASA
-0.01

NMC

UKMO
..

0.28 -0.26 -2.6 -0.39 -0.44 -2.05 6.0 -0.67

0.94 0.86 0.85 0.92 0.79

-0.33

-0.09 -0.22 -0.92 - 1.4 -0.59 -0.85 -0.46 18 -0.09 -0.32 -0.34 5.8 . -0.05 -0.04 -0.98 -4.1 -0.25 -0.28 -2.01 1.7 -0.79 -0.70

-0.21 -1.6 -0.15 -0.55 -2.23 -2.1 -0.79 -1.20 -1.14 39 -0.14 -0.58 -2.36 -1.8 -0.32 -0.08 1.68 -0.4 -0.50 -0.63 -2.27 2.8 -0.88 -1.10

0.95. 22 0.37 1.16 3.25 1.16 1.79 3.23 0.77 1.51 3.54 37 0.75 1.20 2.90 31
0.80
30

9% 6% 14% 13% 32% 8% 44% 20% 32% 29% 17% 35% 28% 11% !9% 9%
10%

0.92 0.79 0.84 0.90 0.80 0.87 0.92 0.83
0.90

It%

. <
L

2 m
0 0

-

z

rn

I
'y

1.10
1.69 44004

1.30 3.27 29 1.21 1.83
44008

30% 14%

32% 8% 46% 20%

0.95 0.85 0.90 II!J2 0.75

OW
3.1 -025
0.54

-1.72 5.3 -0.76 1.10 0.13 1.8

-

1.11 23 049 1.71 2.66 31 1.10 2.06 2.03 43

13% 7% 21% 19%

0.97 0.96 0.66 0.92 0.93 0.63 0.92

0.05
0.8

1.01 22

47% 23% 24% 12%

31% 9%

-048 0.57 -1.88 5.0 -1.10 -1.13 -0.20 6.6

60
1.04 1.19 28 1.24 1.68 2.49 32

-

11% 6% 20% 11% 33% 8% 41% 18% 26% 9%

0.98 0.98 0.90 0 86 093 079 0.89

0.88 -0.9 -0.11 -0.39 030 09 -035 -087 1.21 3.8

0.57 25 0.39 0.93 1.95 I2 058 141 2.37 36

19% 7% 16%
IO%

11.98
0 97

23% 9 Y. 24% 16% 28% 10%

0.84 0 94
1 95 1 074 0.94

'

FNOC

Tl'(S1 U[ms-'l NASA

swiirln1

-0.1 I -0.18 -0.45 1.9 -0.24 -0.14

11.71
1.68

12%
19%

1.11 21 0.58 1.19 205

in

11 6'1

15% 6% 25% 15% 24% 8% 10%
19%

O.WI 0.69 0.97
0.91 0.76 0.92 0.9s 0.66 0.91 0.87 0.61

-n..s
-0.14 -4.6

-0.62

0.78 1.29

1.s~
19 0.51 1.02 1.70 29 0.78 1.10 2.74 2U 1.12 1.76 44ms

26~. 14% 16% 17% IIY. 18% 8% 26% 12% 29% 8% 18% 19%
5%

0.91 0.86 0.9s 0.96 0.86 0.97

1.74

u[m-']
UKMO

%lI (",I
T P.. I ~S

-1.20 5.7 -0.97 -0.8 I

4.n6
29 1.31 1.81 44oJ I

48% 8% 57% 2n.h

-0.22 -0.09 -0.96 1.8 -0.64 -0.u 1.60 9.0 -0.91 -1.14

-

0.98 0.90 0.90
091 0.75

n.14 11.72 -0.291.29 0.39 1.71 -1.2 20 0.05 0.51 n.11 1.15 0.29 1.57 0.4 12 -0.0711.41 -0.~o0.87 -0.~72.1s 2.2 27 -067 0.84 -091 1.54
A*.. e, .e n .--.----

JOY.
15%
~~

m%

6% 21% 11-A 19% 9% 17v.

IVI~ 0.71 0.96
0.94

.:

.

0.67 0.96 0.97

in-h

1182
095 094 066

26~. 7% 14% 17%

U[mr-'J
OW/AES

0.16 1.2 -0.41

u rntS-q
ECMWF

%b[mJ
Tl'[r]

u[m~-']
FNOC 1 l"] 1
U~IBIS-~

-0.14 -0.14 1.6 -0.62 70.59 0.49 8.8 -0.27
-0.11

NASA

SW

TPbJ u[ms-'J
NhlC

IlmJ

SWlllm]
TI'CS]

U[

v rmr- '1
UKMO 1) [ swlirm]

Tl'[sl

0.27 1.9 -0.10 ,1159 0.11 1.01 -11.21 2.18 1.8 21 -0.40 0.58 -001 1.14 -081 2.14 4.11 26 -0.87 I.OU -0.91 1.64

-

1.46 23 0.57 1.05 2.56 27 0.81 1.14 2.57 16 061 1.12 1.79 211

16% 6% 20% 12% 2'1% 7% 29% 14% 29% 10% 21% 15% 20% 6% 21% 12% 25% 6% 21% 11% 24% 7% 18% 19%

0.97 0.98 0.81 0.91 0.96 0.79

-0.09 -0.16

0.59 6.2

129 19 0.11 1.21

15%

0.97 0.97 0.70 11.91 0.91 0.49 0.91 0.94 0.41 0.~7 0.91 066 n.95 0.95 0.56 0.95
11.89.

16%

5%

15%
6%

0.27 2.2 -0.21 -037
1.08 0.6

1.20 22 0.45 1.16 2.62 28 0.92 1.59 2.69 39 0.74
1.48

11% 6% 18% 11%
29% 8% 16% 18% 10%
11%

0.w
0.9s 0.72 0.96 .

0.07 A0 -0.20 -0.22 0.61 4.5 0.06 0.19 0.44

0.95 0.84 0.91
0.97 0.76
0.95

0.4 0.14 0.47 -0.17 -1.6 -0.20
-0.10

11.94 0.65

-0.17 1.2 -0.54 -0.64

1.72 21 n.51 1.56 '2.27 48 0.51 1.76 1.21 I4 0.48 1.42 1.55 21 0.46 1.45 1.51 24 0.78 1.68

20% 2s~. 19% 26% 11% 24% 21% 14% 4% 21% 17% IU% 6% 22%
18%

-

rn

-0.64 -0.81 -0.02 2. I -0.12

C

;I ;
5
0

-n.n

18% 17% 20%
1%

0.46

-0.15 I.D -0.9 21 -0.09 1I.59 11.11 1.27 -0.752.n3 -1.1 27 -0.16 061 -0171.24 -1.61 2.77 4.11 29 -0.79 1.05 -11.8s 1.69

29% 17% 21% 6% 21% 14% 2s~. U% 25%
14% 11% 8%

m F

3
rn

6

41%
19%

I44

V.1 CARDONE et al. .

AEYOW

SWADE lOPl Period of 25-28 Ccrober, 7 990

E EumF ?
E NASA
I N K

I F

m

Comparison of Scatter Index For Offshore Dee? Water Buoys

m M r m 0

"

41001

44015

44004

4401 1

E
I
v)

50 40

0

30

c
L

0
(u

(u

8

- 10 iii
0

20

- 50 5
I -

41001

44015

44004

44011

Peak Period

: 30
41001 44015 44004 44011

a 40

-.

FIGURE 13 Comparison of scatter index (based on mrr) for six wind fields at four deep water buoys and lor the SWADE storm period only.

area, and two (44004,4401.1) bordering the sparse data area and probably more typical of the open N o n h Atlantic Ocean. Note how the margin of improvement for the OW/AES SWH-SI, about a factor of two in the data rich area'narrows

:.. .

.

/

145

'

'

.

Palcu01mmer25.28.1939
scarier 01y * l r m wan WK.

wd a-

"

E "

I

Y

I

Y '

/'/

w
u /

"
u

'
./ ,/

/'

' '
..

.
30
40

10

20

50

Scatter Index of Ws (%)
FIGURE 14 Comparisonof wind s p e d SI versus SWH SI for SWADE IOP-I.

:.

considerably offshore Hudson Canyon and Gorges Bank. This provides further evidence that the exceptionally high level of skill exhibited in the OW/AES series is attainable only from the enhanced data base in the antral SWADE area, and that only slightly further east of this area, wind errors increase to levels characteristic of the open-oaan sparse data regime. The premise of this study is that wind errors dominate wave hindcast errors for a model as skillful as WAM4. Therefore, we should expect a high correlation ktwecn wind errors and wave hindcast errors. Figure 14 shows the correlation between the "site" wind errors and the "site" wave errors. There i reasonably high cornlation s between these CXTOR overall, with the notable exaption of NASA, which as noted above nearly matches the very low wind errors of OW/AES but for which wave height errors are generally higher.

. .

.

..

146

V. J. CARDONE et aI.

6. ‘DISCUSSION
Large errors in surface marine wind fields specified from conventional historical or real-time data remain a stubborn obstacle to further research in many areas of air-sea interaction. This limitation affects problems in which high frequency wind forcing is important. such as specification of air-sea fluxes for numerical weather prediction, and ocean wave and winddriven current prediction, as well as longer time scale problems, such as modelling the seasonal cycle of ocean current systems. and climate prediction (e.g. ENSO). It appears that truly global specification of accurate synoptic scale surface-marine winds must await the operational implementation of satellite-borne radar scatterometers and multi-frequency microwave radiometers specifically developed for remote sensing of surface-marine wind and wind stres. With such a system not expected for at least several more years, there remains a critical need for high quality surface-marine wind data, especially to further develop and test contemporary ocean response models. In this study, the WAM numerical spectral octan wave model is applied to hindcast SWADE IOP-1. The WAM model is already so refined that when applied to real cases, effects arising in remaining deficiencies in physics or numerics are usually totally masked by errors in the forcing wind fields. SWADE included an enhancement of the conventional meteorological buoy network off the East Coast, with all such data transmitted in real time so that it would be available to NWP centers for assimilation into their routine global scale analysis-forecast cycles. Initially. it was expected by the SWADE team of investigators that among the products generated by three such centers (ECMWF, FNOC. UKMO) and a fourth and fifth wind field produced by a special higher resolution objective analysis (NMC and NASA) would be found oncor more forcing wind fields possessing acceptably small wind errors. In this study, each of these wind fields was used to drive a deep-water high resolution adaptation of WAM for SWADE IOP-I, and unfortunately errors in hindcast SWH (significant wave height) and TP (peak spectral period) were found to be nearly as large as found in routine real-time wave analyses. That is, the scatter index (ratio of rms difference to the mean of the measurements) in specification of the time series of the SWH at a buoy moored in the SWADE array varied between about 20% and 55%. The aggregate SI over the 8 buoys used for this evaluation varied between 23% and 41%. There was also a systematic underprediction of wave heights during the two most energetic cyclonic systems which traversed the SWADE array during this IOP. Hindcasts from every objective wind field underpredicted the peak SWH in at least one of thee events by 30-50%. The scatter in specification of the peak spectral period, typically about 1.5s (N) varied less over the different simulations than the scatter in SWH. with a tendency for all runi.to underprcdict TP. A sixth wind field (OW/AES), derived by tedious manual kinematic analysis of’all conventional and SWADE wind data, was found to provide exceptional and unprecedented skill in specification of the time series of SWH at buoys moored in intermediate and deep water depths in the SWADE region. The scatter index in SWH varied between 14% and 20% or half of the SI found for the alternative wind fields. For the main IOP-1 storm the SI in SWH averaged 12% in the SWADE array, which is statistically indistinguishable from the skill of “perfect” hindcasts

.

/

0

.

.

IN SEARCH OF THE TRUE WIND FIELD

147

.. .

after considering sampling variability in the wave measurements. Surprisingly. however, the skill in TP was not commensurately reduced. suggesting strongly a source of error other than local wind error, such as model physics or numerics or perhaps importation of "emor" from the BASIN domain where wind errors in all of the alternative wind fields are larger. Our evaluation of the alternative wind fields establishes the OW/AES wind fields as clearly superior to the alternative wind fields, at least within the SWADE area, but has not dwelled on precisely why there is such a large differential behavior in the wave hindcasts between OW/AES and all other wind fields. One might logically suspect the large range in nominal temporal and spatial resolution in the sources (see Table 2). However, the OW/AES resolution of 0.5 degrees is coarser than NASA/GSFC and comparable to NOAA/NMC. The intrinsic temporal resolution of OW/AES is 3 hours (later interpolated to 1-hour), which is comparable to those of UKMO and NASA/GSFC. Therefore, we do not believe resolution effects are the dominant source of difference although the OW/AES winds provide an ideal opportunity to quantify the effects of temporal and spatial resolution of wind forcing on wave model skill, and such a study is underway. We surmise rather, that the S U C ~ S of OW/AES r a t s mainly on the following characteristics: (1) the more precise specification of the space-time continuity in the evolution, absolute magnitude and translation of the wind speed maxima associated with the migratory cyclonic and frontal systems which affected the SWADE FINE area during IOP-I; (2) the precise n definition of the wind speed variable analyzed, i chis case the effective neutral wind speed at a height of 10 meten, and the a.reful adjustment of all wind measurements to this definition before assimilation into the analysis. Objective analysis systems have great difficulty maintaining continuity because the analysis is usually carried separately at 6-hourly intervals using mainly 'analysis time" observations, with relatively slight communication betwcen.neigbboring analyses (real-time systems, of course, can have no knowledge of future bhourly analysis states). Furthermore. objective analysis systems usually assimilate 'surface" data on the NWP model lowest constant pressure or constant a-layer. making a precise analysis of 10 meter or 20 meter information difficult. Indetd, most operational systems fail to recognize the large raage of anemometer heights of "surface" marine wind measurements, typically 3.5 m to 80 m, with hardly any data acquired at 10 meters! The main conclusion of this Siudy is that the OW/AES Winds posscss errom low enough in IOP-1to further the development and testing of third-generation wave models. That evaluation involves the analysis of model simulated and measured directional wave spectra at selected SWADE sites. investigations of alternative source term formulations and, using appropriate measurements, of shallow water effects and wave current interaction. These additional studies are underway and will be reponed in future papers.
Acknowledgmenrs

The authors gratefully acknowledge funding suppon by the Office of Naval Research (ONR) under Grant Nos. N00014-90-J-1464, N00014-92-J-1546, and N00014-88-J-1025 and the National Aeronautics and Space Administration under

I48

V. J. CARDONE et al.

NAGW-2672 and Contract 957652 (NSCAT Science Working Team). R. Jensen conducted his work at the Coastal Engineering Research Center of USAEWES under the Coastal Research and Development Program, "Upgrading of Discrete Spectral Wave Modelling Work Unit". Permission was granted by Headquarters, US. Army Corps of Engineers, IO publish this information. The authors extend special thanks to the following Models Division of the Fleet Numerical Oceanography Center, National Meteorological Center of the National Weather Service (William H. Gemmill), the Global Modeling and Simulation Branch, Goddard S p a n Right Center/National Aeronautics and S p a n Administration (GSFCWASA) (Dean D u e ) , the Europcan Center for Medium Range Weather Forecasts, the United Kingdom Meteorological Ofice. the Atmospheric Environment SeMcc of Environment Canada (Madhav Khandekar. who funded the OWIAES wind field analyses), and the National Data Buoy Center, especially Kenneth Stccle. The authors are also grateful to Ram Vakkayil (RSMAS), Renate Brokopf (MPI), Andrew Cox ( O wand Michael Parsons (OW) for assisting in data proassing, wind field analyses and the numerous model runs. Jean Carpenter (RSMAS) expertly draw the figum depicting the p m u n and streamline fields.
g&IK

References
Baker. W.L A W R Kalnay. E. Halcm. M, Woiarhp. P.M. Pevhcryds S. Eddmann D. (1984) . L a r g s l a l e annlysis and forecast cxpcrimcnu with wind data fmm the SEASAT A scattemmeter. 1. t m p h y . Re% 89(D3),4927-4936. Bauer. . Hauslmann, S Hrrwlmans K. and G m k . H.C (1992) Validation and Assimilation of E SEASAT Altimeter Wave Hcighu Using the WAM Wave Model 1. Geophyr R q 97. I 671-12 2 681 Barker. EH. (1992) Design of the Navy's multivariate optimum interpolation analysis rysum. Wea and Porecas~. 7.27.0-231. Bell. RS. and Dickiwn. A. (1986) The Metsomlogial Offia operational numerical weather prtdidon MeteoroL 0s No. 41. Her Majufy's Stationary Offifia,London. 61pp. system ScL PQ~, Cardons V J . (1969) Spsdfication of the wind distribution in the marine boundary layer for wave lorecming. Rupn T - 9 0 . &ph* R6.1 Scicnm Lab, New York University. 137pp. (Available from N7IS: Order AD 702 490). Cardons VJ- Bmccoli AJ, Greenwood. C.V. and Granwood, J.A. (1980) Error charanexistics of extralropical storm wind fields rpsdficd from historical data J . Petrol. T e c h 32 872-880. Cardon%VJ. G m w v w d JA. and Cans. M.A. (1990) On trends in historical marine wind data J . of Climnre.3. 113-127. Cardons VJ. (1991) The LEWEX wind fields and barcline hindarr In: Direcriod Ouan Wow Spectra Ed Robcn C. Beal. The Johns Hopkins Umvenity PU , 136-146. IS Cardons V J . (1992) On the strunum of the &ne surfn~c wind field in cxtratropicpl st~nns. Im. 3rd Workhop on Wme Hindcorting and Forecorting, M o n t d , Q u e k May 19-22 (1992), 54-66. Pub1 -. Envimnmmt Canada Doamrvinv. Ontario. Cardone, VJ. and Srabo. D. (1985) lmpnn of uncmoinly in specification ofoKshorc wind on assurpcy of wave hindsarb and lorscaru P r a . Int. Workhop on Offshore W f d and Icing, Nova Scotia. 7-11 Onobcr. 1985, h b l . Envimnmcnt C a d DounrVicw, Ontario., Canuo. MJ. G m k r H.C, Jensen, R.E and Dondan. MA. (1993) Observations and rnodclling of win& and w a v a dqring the surface wave dynamics cxpcrimcnr Intensive Observation Period IOP-I: 20-31 Onobcr 1990. Tech Rep.. CERC-93-6, US Army C o w of Engineers. Washington. D C 203pp. Donclan, MA. and Pi& WJ. (1983) The rampling variability of crtimatcr of SPeClm of wind-pncrarcd w a v a J. Geophyr Ru. 88, 43814392 'DuRy. D.G. and Adas, R (1986) The impan of Ssprar-A scancromner data on the numerical prsdidon of ths Quecn Elizabclhll norm. J . Geophyr R q 91(U). 2241-2248.

a
IN SEARCH O F THE TRUE WIND FIELD

149

a

Gemmill. W.H.. Yu T.-W. and F C ~ L D.M. (1988) A statktid comparison of methods of determining ocean surface winds. Wen. ond Forecast. 3. 153-160. Gcmmill. W.H. (1991) High-resolution regional ocean surfaa wind fields. 9th Con! on Nwn Wca. Pied.. oslober 14-18. 1991. Denver. Colorado. Amcr. MeL SOC, Boston, MA. 190-192 G m k . H.C. a d Boutin P.R (1991)SWADE cruix r p n for voyages 224 and 226 on the R / v Ocemus. eo ONR Contract Repon. 21 pp. Graber, H C Camso MJ. and J n u m R.E (1991) Surface wave simulations during the Oerober storm in S W M L Pmc M 7 S PI Con& Marine Technology sodery, New O l a s LA. lS9-164. ren. Guillaums A and Momad. N.M. (1992) A ucw method for the validation ofdtimcterderivcd s u state paramelm wt d t s from wind and wave modetr J. Geophys. RU, 9XC6). 9705-9718. ih Has&mann. S. and Hasxlmann, K. (1989 Computations and pnrameterirPtions or the nonlinear energy transfer m gravity wave spcttum. P a n I: A new method lor efficient computations of the exact nonlinear transfer inrsgraL J. Phys. Ocemogr- 15411). 1369-1377. Hogan. T.F. and Rosmond. T.E (1991) The description or Navy operational global nunospheric p d i c tion system's Ipnral forrcart m o d d Mon Wea Rm, 119. 1786-1815. Januns P A E M . (1991) Quui-linear thcory for wind-wave generation applied to wave forecasting J . Phys Ocemwgr- 21, 1631-1642 Januse P A E M , Komcn. GJ. and de V w s s WLP. (1984). An operational coupkd hybrid wave prediction model. 1 Gmphys Rrr WC3). 3635-3654 . Janucn PAEM. Lionello, P, Reistad. M.and Hollingswonh. A (1989) Hindcasts and data mimilation studies with the WAM modcl during the Scaut period J. Gmphys. Rs. 9yC1). 973-993. e. Knhrm. ILK. and Calkom. U. (1992) Reconciling diocrcpMda in the o b r s m d p w I h of wind-gmra d wave% J. Phys Occafqr- P 1389- 1405. Khandelur. M.L Lalbchany. R and Cardons VJ. (1994) The performance of the Canadian spcnral acem wave model (CSOWM) during the Grand Bank E R S l SAR wave spsnra validation e@ xmcnL Amsphrc-Occmu. % 31-60. Kodn, PJ. Olwn. D A Wick, AF. and H m e r . RD. (1991) Surface weather analysis at the National Mets~mlogid Centsr: Current p d u m and fvtvre plans. Wen. Forecmting. 6.289-298. Komcn GJ- Huwimanh 5. and H~uclmann,K. (1984) On the a i s t c n a ora fully developed wind-ra rpsvum 1 Phys. Occmwgr. 14 1271-1285. . Lorens. AC (1981) A global t h r e c d i m i o n d multivariate s t a t i s t i d interpolation schcmr Mon. Wea Rm, 109.701-721. Maat, N. Kraan. C and OMS W A (1991) The mughnm of wind waves B o d - L a y e r Metmr, 54. 89- 103. Morris, V.F. (1991) The Bonncr Bridge storm. Mminrn W e a h Lag. 3 No. 1 4-9. , Oberholm. D. and Dondan. M A (1995) SWADE dam guidr NASA Tech. M n n xx. Goddard Space flight On=. Gmbelt. MD. 8Opp. Omland, I.€ and Gemmill, W.H. (1977) Prediction of marinc winds in the New York bight. Moa Wea Rlo. lM.1003-1008. Rou. D.B. Cardons VJ, Overland J.E. McPhmon. R, Pierron. WJ. and Yu, T.W. (198% Oaanic surfaa winds. Ada in Geophyr, 27. 101-139. Sanderr F. (1990) Surface weather analysis over the oaans Searching for sea truth Wen Fooncmting. 5, 596-611 Sand- F. and GyakurpJ.R (1980) Synopticdynamic climatology or the "bomb". Mon Wen Rm. 108. IS89-1606. Shapim. M A and K w . D. (1990) F m n u jet s t and the tmpopnuoc &suatmpud Cyclonex The E * Polrmn Mmrinl Volwnr. N o n o n C.W. and Halopainns EO. E. Anmr. Memr. SOL, d % 167-191. Shaw. D.B. Lonnbeg. P, HoUinpwonh A and Undcs P. (1987) The 1984/85 revisions of the ECMWF I U and wind a d M n + Q a J. Roy. Mec. Sc 113. S33-566. un o. ~. Snyder. R L Dobson, F.W. ElioL J.A and Long RB. (1981) A m y mearurements of aunorpheric -e r Bunvations above surface graviry waves 1 Fluid M e c h 101 1-59. . Stcecs*. L , E Tcng C.C. andWang D.W.C. (1992) Wave dirmion mcasumcnts using pitch-roll buoys. O c m Engng. 19(4). 349-375. Swail, V.R. de Lorcnds B Do+ C. and Bigio. R (1992) An i n t d v e graphics approach to wave , analysis and fo-tinp 3rd Im. Wkshop om Wave Hinacmting and F o r c c n n ~Montreal, Quebec , May 19-221992 Publ. Environment Canada. Downsview. Ontario. 333-338. U d i n i . LW. Corfidi S.F, Junker, N.W. Kocin, PJ. and Olwn, D.A. (1992) Repon on the Surfnu A -,-- d d Workshoo held PI the National Mctmmlomeal Center. 2S-28 March. 1991. Bull. Amrr. Meteor. S O ~73.4%-472 ,

-

a

V. J. C A R D O N E et d.
31 Group: HaJselmana S .
yr.. 8(12).1775- 1810.

. L. J.A. (1988)P. Guillaume.modelC-rad o n s V.C. Greenwood J.A. Reirtad. M. model J. Phys. Lionello. A. a hmbrcsky L. mg The WAM third generation ocean wave prediction

Haselmann. K. bauer. E. Janucn. P.A.E.M. Komm. GJ.

. Donelan. MA.,

Briwoz M.G. and Huang N.E. (19911 Riding the crest: A d e o two wave f iens. AMS Bullrein ?Z No. 1163-183. 1984) A comparison of several 's~ngle-w.- s l i m a t o n ol the standard deviation of wind n. J. Climate Appl. Mcrmr.. 23. 1362-1366. L 11989) A verification study of the global WAM model k m b c r 1 9 8 7 - N o m k r 1988. F Tech. Rep. 6 . European Cenirc for Medium-Range WY Fo-ring 3 Reading England

The Performance of the Canadian Spectral Ocean Wave Model (CSOWM)Durina the Grand Banks ERS-1 SAR Wave Spectra Validation Experiment
u

ML. Khandckar and R. Lalbehany
Amtospheric Envimnnum Service D o w m i m , Ontario

and

V. Cardone
Oceanweather Inc. Cos Cob, Connecricut. U S A .

[Original IMrmyTipc rcecivcd 21 April 1993; i rsrircd form 10 Nowmber 19931 n

~~

..

~~~~

~

~~

ABSTRUI CSOWM ir trvimnmem C d ' s operational wave mode11haI L opemed daily nr the Canadian Mereoio&gicd Cenm (CMC) Monrrinl and u driven by surface windr in generaled by the regional weatkrpredim'on model o CMC. For tk dum'on o the Gmnd f f Banks ERS-I SaR wave sperm validation jeld uperinum. ww versiom of QOWM a fimr genemwn (IG) version and a rhird gcnem'on (3G) wrswn - were used i a hn c r n idw .mode t gencme wave model products. I n &ifion, ww sers o wind fuldr. one obminable o f fmm r k CMC weather prediction model and the o r k r obraincd rhmugh a h a n d d y s u o 3-hoM surface pressure c h r s and an appticarion of a marine planetary bo*f layer model were used w generate m e model pmdunr. The Various wave d l pmduus were eduared a g h r wind and wave measuremc~sd e owr the mrdy sire and &o agoinrr auailable buoy &ca from the nemrk o Canadian buoys on t Gmnd Banks and f k

-

0

i

rhc Scorinn Shelf rqionr o rhe Canadian Arlonric. f The m M o n suggests I k both versiom of CSOWM pe$ormed reasonably well and genemed e r m r stazutics rhaI were quire cowprabk wirh rhase generated by some o f r k opemrionnl wave models presemly implenwued in ~ t i wearher sewices o many ~ d f cowurirc. Funher. the wave model products genemed using windr obminabk f m m handanalysed M a c e pressure cham had the closest overall ogreemem with buoy m e a s u m ~ ~ ~ and thu i rum led IO an improwmcnt in the ermr sratinics and in model specma n The res& presemed here h e provided a degree o c o n t n c e in r k use o a o w M f f pmducu for f i m k r &sU on SAR &a assimikuion in wave~modek.

I

..,

32 I M L Wandekm, R Lalbeharry and V. Cardone
k S venU de su*C g i d r t s par le d l c r k 8 b a l de pdvision mCrCo du CMC. Pour la durir de 1'crpCriCncr ERS-I SAR de miidmwn des specrres des ~agues le terrain. on sur a mi4 deux versions de CSOWM (premihre gCnCmrion. IG: rmisi2mc g6M'MriOn. 3G)en is mode de rkrmspecrion pour ginircr des pmduiu de moddcs des yo~urr. .En plus. deux ensembles de champs de VCN. l'un g d d d par le modhle de prdvuwn mirko du CMC et l ' o u p p m v e m d'un analyse m ~ v r l l e h pression de surface oyx m i s heurci er de d'unr applicmon d'un modhle de la couche limire marine planCloire onr irk urilisCs pour gCnCrer ces d m c s pmduiu. On a comporC ks divers pmduirs de mod& des vogues d de5 obscrwrions du veN er des Lagues sur le rennin et aussi oyx donndcs de bou'cs disponibles du rcscOu de boukes canadien sur les Gmnde Bmcs cr la piare-forme S c o r i a L conparaison indque que lcs deux versions du CSOWM perforrm111 geniralemcnr bien a et gb~hrenrdes statisripues d'crrcun qui so111 fncikmcN compambks d cclles gCnCrCcs par d'aurres modhles opcmriOnnels des vogues udisis par les services mtrkomlogiqurs krrmgen. De plus. k s pmduiu de modhlcs des vogues urilism les VCNS des analyses manuriles de la presswn de swface s'accorde111 gbbalemru de plus pr*s a w c ks mcsures des bodes, cc qui en retour o conduit d des d l i o m i o n s dam its sra~isriqucs d'emeurs er dam les specrm des modhles. L s dsulrau ci-dessous nous donne un degrd de confionce dam l'urilisarion des pmduirs du CSOWMpour de plus mplcs analyses de lhrsimilnrion des & d e s SAR dam les modhles

a

des Vngucs.

1 Introduction Since J a n u q 1991 the Atmospheric Environment Service (AES) of Environment

Canada has implemented in its forecasting system an operational specaal ocean wave model called OOWM that is run twice daily at the Canadian Meteorological Centre (CMC) in Monaiai and is driven by the 10-m level winds generated by the regional finite element (WE) weather prediction model of CMC. The wave model OOWM is designed to operate over two separate o c d c regions, the northwest Atlantic and the nonheast Pacific. Over each of the= two regions. a coarse grid and a nested fine grid a designed for model calculations. Figure 1 shows the n coarse and the fine grids of n o w covering the northwest Atlantic region from about 24 to 75"N and f o 0 to 81"W. rm The diamond-shaped area in Fig. I represents the CTOSSOVCT node for the ascending and descending passes of the EM-I during the field experiment. The crossover node formed the principal validation site where in situ wind and wave measurements were made using meteorological and directional wave buoys. The primary aim of the field experiment (see Dobson and Vachon, 1994. this issue) was to provide in situ, aircraft. and numerical wave forecast model validation of the Synthetic Apemre Radar (SAR) specna aboard the satellite ERS-I: accordingly. the meteorological and directional wave buoys were deployed at locations coinciding wt selected grid point locations of CSOWM. These wind and wave meaSuRmenLS ih formed the principal dataset for our model validation. In addition, wc also used available buoy data fromthe network of Canadian buoys maintained by the Marine Environmental Data Service (MEDS), Ottawq on the Grand Banks and the Scotian Shelf area. to validate our wave model.

CSOWM Performance During the Grand Banks ERS-I S A R Experiment 1 33

-a
grid The c o ~ grid has 2081 grid pinu M d the ncurd fim grid. e fig. 1 The M M ~ C for Q. 2419 poinn. The diamond-sha&d arra rcprcwnn Ibc arns0v-x node for Ihc uccnding and dewcnding paue of the sntrllirc ERS-I during h e field upcrimcnt.

0

Wind is the primary driving force for all ocean surface wave models. and hence an ocean wave model can only be as good as the wind field that drives the model. An accurately specified wind field will help generate wave model products that arc close to the “sea” truth, assuming that the wave physics is adequately represented by the source terms of the wave model. Thex wave model products (generated in a hindcast mode) can be validated against in situ wave measuRments to determine the error srmcmre of the hindcast wave fields. Having defined the error srmcNre, the bindcast wave field could then be used to evaluate and calibrate the remotely sensed wave informadon provided by the ERS-1 aidmem and the Synthetic Apcrmn Radar (SAR). An accur;ucly specified wind field cari also be used to evaluate CIIUK in conventional surface wind analysis and to serve as a useful rrfeknce for the rcmotely sensed winds from the scanemmeter aboard the satellite ERS-1.

3 1 M . Khandekar, R Lalbeharry and V. Cardone . L In order t fulfil the principal aim of the ERS-I SAR wave spectra validation o
experiment, the wave model CSOWM was used in a hindcast mode to generate wave model products in and around the field experiment area. Two sets of wind fields. one identified as CMC winds and the other as MMM (Man-Machine-Mix) winds were used to generate wave model products. Recent studies (e.g. Sanders, 1990 Cardone. 1992) on the SUUCNE of marine surface winds in extratropical storms have dcmonsuated the uwfulness of a careful (and ,subjective) analysis of surfact weather charts in generating accurate surface wind fields over an oceanic region. In view of these studies, wave model products using two sets of wind fields (CMC and MMM) were generated and evaluated against buoy measurements. In Section 2, details of the wave model arc briefly presented. In Section 3. brief details of the rwo wind fields are presented followed by an evaluation of the various model products against buoy measurements. Further. the model-generated wave spectra are evaluated against buoy-measured spectra by analysing the evolution of the important wave systems during the entire experimental period and also during selected time periods; these results are funher discussed in terms of wind field differences. The final section summarizes the important findings of this study.

a

2 The wave model h ga ) The wave model CSOWM is a version of t e ODGP (Qcean Data Gathering m r m model originally developed by Cardone et al. (1976). The CSOWM grid for the nonhwest Atlantic is laid on a transverse mercator projection w t assumed equator ih at 5 " and a grid spacing of 1.084' of longitude on the assumed equator. The 1W aia. spacing varies from about 108 km south of Florida to about 119 kin near H l f x Nova Scotia. A nested fine grid covering the Canadian shelf region with a grid spacing of about 0.361" of 1ongiNde is designed 85 shown in fig. 1. The grid spacing for the fine grid varies from about 40 ktn at the east edge to about 38 h at the west edge. The basic framework for all specmi wave models is the energy balance equation, which is written in standard notation as

a

a€ - CV -(C,€) = s
at

where E = E(f,e,x,r) the spectral energy density. which is a function of is frequency f, direction of wave propagation 8, position vector x and time r. C, is the group velocity. which can be expressed in terms of phase speed C.wavenumber k and water depth h; in deep water, the group velocity is a function of frequency only. S = Su, I) is the net source function representing the physical processes 9,x, that -fer energy to and f o the wave s p e m m . The source t r S can be rm em symbolically written as

s = Si" + Sd + SdI

CSOWM Performance

During

the Grand Banks ERS-1 SAR Experiment

t 35

In ( 2 ) Si. the energy input to the wave field from the amosphere. Sd represents is
the uansfer of energy associated with the non-linear wave-wave interactions. and
Sb represents the energy dissipation. which includes dissipation in deep as well

..
i

as in shallow waters. In this study we have used only the deepwater version of the'model. since the impact of bathymetry on wave measurements during the field experiment was considered to be minimal. Full details k d definitions of various terms in (1) and (2) are given in the Appendix. However, brief explanations of important terms are provided below. In discretized form the two-dimensional spectrum E ( f , 8)represents the specual energy density (m' Hz-' rad-') in a frequency-direction bin, from which wave parameten such as significant wave height, mean wave period and mean wave direction arc deduced. In terms of the discrete specmm. the toral energy (wind sea and swell) is given by
(3)
j=I i1 n

.. >

where N = 23 is the number of frequency bands, M = 24 is the number of directional bands, A0 = ~ K / M the directional bandwidth. and Af is the frequency is bandwidth. The frequencies range from 0.039 to 0.32 Hz inmasing in geomcmc progression with a constant ratio of l . I W applicable to both the coarse and the fine grids. The total energy specmm is made up of a wind s a s e t u and a swell spece pcrm m To parddon the total energy specmm into wind sea and swell components . the well known one-dimensional Pienon-Moskowia spectrum (hereinafter refemd to as the PM specmun) for a fully developed wind sea (Pierson and Moskowiu 1964) is applied The PM spemrn is given by

wheref, = 0.14g/U19~is the frequency corrcsponding to the peak in the specpal energy of the PM s e t u and UIg.3 is the local wind speed (m I-') at 19.5 m pcrm above the ocean surface. The toral specmm is partitioned into wind sea and swell components as detailed in the Appendix. The propagation of wave energy in the model is effected through the advection term V . (C,R i (1). The csow uses a table of pmpragation coefficients to n compute the change in specaal energy due to advection at a given point by a p p m priately weighting the specua at surrounding grid points. For a given grid layout (uansverse memtor projection for csow) and for a preselected set of wave frequencies and directions the propagation coefficients are functions of IatiNde o d y for a dupwarer model. The propagarion time step mist be small enough,so that the specaal component with the fastest group velocity will not navel more than

36 I

M L Khandekar, R Lalbeharry and V. Cardone ..

one grid distance. More derails of the propagauon algorithm and' the numerical integration time step can be found in Greenwood ei al. (1985) and Khandekar and Lalbehany (1990). The various source terms in the wave model are expressed symbolically in (2). several important theoretical and experimental studies of the last 35 years have helped develop suitable expmsions for these source t r s as a function of input em parameters. In the fint genemion (1G) wave models. which were developed in the mid-sixties, the source urn expressions were based on two imponant studies, one by Phillips (1957) and the other by Miles (1957). Later theoretical and experimental studies have yielded source term expressions that M believed to provide an improved representation of wave physics and have now become the standard expressions for most third-generation (3G) wave models. The 1G and 3G versions Of C S O w use different source term expressions as detailed in the Appendix.

3 R s l s and diseussiom eut As mentioned earlier. two versions (1G and 3G)of the CSOWM were used to generate wave model products for the duration of the field experiment. The model was driven by two sets of wind fields (CMC and MMM) each prescribed at 3-hourly intervals for the experiment's duration. The CMC winds were obtained from the CMC archives at Mondal and the MMM winds were obtained through a 3-hourly subjective analysis of surface pressure cham and an application of a marine plane~ boundary-layer model. The model was staned with a flat sea (zero energy everywhere). and a spin-up time of 48 h from oo00 VTC on 8 November to oo00 m on 10 November 1991 was used before the wave model products were generated for the duration of the field experiment. The.wave model products were evaluated at four grid locations inside the cmssovcr node and a five grid locations elsewhere in the Canadian Atlantic t where MEDS buoys an located. The various buoy locations in the Scotian Shelf and the Grand Banks regions of the Canadian Atlantic are shown in Fig. 2. The boundary conditions on the coarse grid require that the land points along the boundary be set to zero wave energy before each propagation time step: water points are allowed to grow and propagate energy. Therefore, specwl wave energy that propagates to the boundacy' at any given time step is completely absorbed. The boundary conditions also allow the spectral growth to evolve as a function of fetch for offshore winds. However, since the distance between the first interior point and its neighbouring land point differs fium the distance between the fint intenor point and the actual coastline. the fetch i not precisely defined. This boundary en?r is s usually important within about two grid distances of the coast. Since the cenm of action in this study is away from the boundary, the influence of the boundary points on the interior grid points is minimal. For the nested fine grid. the fine grid points receive information before each time s u p from 44 coincident coarse grid poinrs that lie on the boundary of the fine grid. Thc energy values M then !inearly interpolated to the 0 h points on the boundary of the fine grid that arc not poins t-

0

0

CSOWM Performance During the G a d Baniis Em-1 SAR Experiment I 3 7 rn

Fig. 2 Map rhowing buoy loadonr in rtr c.nadim Allanric and mC grid poinu c10141 u) rtr buoy I o a d O n r w h a t wave model pmducrr mxc e n l d Iluide rtr diamand-rhapd area rtr buoy IOCationr ainSidc d y With dx model gidpohl locadonr.

on the coarse grid and then 10 the grid points that arc one row in f o the boundary rm of the fine grid.

The Two Wind Fields 1 m c m s The staning point for CMC surface winds is a f i ~ guess. also referred to as a t mal or background field. which is usually a 6-h forecast produced by the global forecast model using a (defined as the ratio p / p , . p being the pressure at the level in question and p , the surface pressure) as the vertical coordinate on a global Gaussian grid. The CMC winds used in this study w e n generated by the RFE m d l which UKS as a trial field the 6-h forecast generated by the global forecast oe. model. Io the CMC objective analysis cycle. the optimum interpolation technique is applied to produce analysed residuals of winds. t e r n p k m , geopotential height and dew-point d e p w i o n at grid points of the model a-levels. The til 'field is ra
a

38 I M L Khaudekar, R Lalbehary and V. Cardone
added to the analysed miduals and this comctcd field is initialized and then used as input to the weather prediction model for generating forecasts. More details of

._

the CMC global data assimilation system can be found in Mitchell et al. (1993). Ship and buoy winds are mated exactly the same way in CMC objective analysis. The shiplbuoy winds arc assigned an error variance of about 13 mz s - ~ , and ' As a result, the optimum ins rawinsonde winds, a variance of about 4 mz . terpolation scheme gives less weight to the ship/buoy winds than the rawinsonde winds in generating the analysed residuals for the u- and vsomponents of the wind field at the model's lowest active level, namely, u = 0.99. Using a boundary-layer model @elage. 1985. 1988). the winds at the lowest active u-level arc extrapolated funha to obtain the surface winds at u = 1.0. assumed to be nominally the IO-m level. Thus. the surface winds generated by the RFE model may not reflect the full impact of the shiplbuoy winds because of the larger variance assigned to these

winds.
The above procedure is applied at two synoptic times. namely.
M)oo

and 1200

m, and diagnostic surface winds applicable at the IO-m level arc generated every 12 h. Since the wave model csow requires wind input every 3 h, a set of 3-. 6'
and 9-h forecast winds obtainable from the CMC we3ther prediction model were used in conjunction with the diagnostic winds valid at oo00 and 1200 m to create a sequence of 3-hourly wind fields for the'duration of the field experiment
2
MMMWMDS

The man-machine-mix (MMM) winds, as the name suggests. rue the winds obtained through a subjective analysis of the surface pressure cham and an application of a marine planetary boundary-layer model. The starting point for generating MMM winds is the assembling of bhourly surface analysis cham of the METOC (Meteorology and Qpanography) Centre, Halifax. 12-hourly final surface analysis cham f o NOM (National Oceanic and Annospheric Administration). Washington. rm D.C., 6hourly ship repom, U.S. and Canadian buoy EpOrrC and C-MAN (Coastal Marine Automated Network) data including, where available. cowcutive l&min average winds. The scatteromcur data from the ERS-1 were not used since the technical problem associated with the interpretation of data was not molved at the time of this analysis. Having assembled the various somes of data. the following steps arc followed
1) Reduce the measured winds to effective neutral 20-m level hourly averaged s winds; for ship repom. a skbilicy adjustment based on the anemometer height i carried out following Cardone et al. (1990). 2) On computer-plotted base maps, carry out a derailed hand analysis of surface prcssurc and surface a r temperature at 3-hourly intervals and a x a surface i -. lcmperarurc analysis once per day. 3) Digitizc each analyxd map and calculate gridded pressures and tempuaturcs
. .

CSOWM Performance During the Grand Banks ERS-I SAR Experiment I 39
fitting a paraboloid to seven or more digitized poino located within a small circle about each point. 4) Calculate effecrive neutral wind fields over the entire domain of analysis from the a d d e d pressures and tempera&res using the marine planetary boundary-layer model of Cardone (1969). The winds (speed and direction) obtained by application of the boundary-layer model arc called PRESTO winds. 5). Over a more limited domain. carry out a subjective kinematic analysis of the adjusted wind observations using the winds generated by the planetary boundarylayer model as a backgmund field 6) Digitize the wind speed and direction at grid points and replace the PRES~U winds,with kinematic winds over the domain where kinematic analysis is performed. The combined PRESTO and kinematic winds arc' identified as M M M winds. Since kinematic analysis is exmmely labour intensive. it was performed over a Limited domain of ( s o w grid where additional shipbuoy dara w r available and ee for a shorrer time period extending from 1200 urc on 13 November to 1200 urc ,,..? -. 21 November 1991. For the field experiment, wind information was available .on from a total of eight buoys over the Scodan Shelf and the Grand Banks region . Fig. 2). Since the majority of the buoys arc located to the south and the ,t of the crossover node, the domain of the kinematic analysis was choxn to ..... .. . cover the Scotian Shelf region southwest of the cmssover node; An example of hand-drawn kinematic analysis for two map k e s is given in Fig. 3. which shows smamlines and isotachs drawn subjectively, taking into account available wind data and applying the principle of continuity of wind field. From these kinematic analyses. wind speed and direction can be digitized at various grid-point locations.

!

;.:&

b Evalm'on of Baric Wr and Wave Modrl Products ud Before we discuss the evaluation of the basic model parameters such as wave height
and peak period. it is insuuctive to examine the variation of wind speed at selected locations. In Fig. 4 arc shown time plots of wind spced at two buoy locations, 44137 and 44140, which arc compared against the h e plots of CMC- and MMMgenerated winds at grid poino closest to thew buoy locations. A close examinadon of the rime plots reveals that, in general, CMC winds significantly underestimated the observed winds at 44137 (located in the westernmost region of the Scodan Shelf), which experienced three (or more) cold outbreaks during the two-week period of the field experiment The MMM procedure adjusts the wind speed based on the'aanospheric stabiliry as represented by the air-sea ( . T,) tempcranu'c T '%rencc and hence generates wind sped with a much smaller bias than the mponding CMC winds. For the buoy 44140 (located in the castern region of the Scotian Shelf), CMC winds ovmstimarcd the observed winds, especially during periods of high wind speed, whereas the MMM winds showed a closer a m m c n t wt the buoy winds for the two-week period. A quantitative comparison of CMC ih

I

*

-

40 I

M L Khandekar, R Ldbeharry and V. Cardone ..

Fig. 3 A m p l c kincmaric Malysis at two mnp timU on 14 November 1991 for the Sconm Shelf nnd the southern Grand B ~ k area The analysis shows the sIlwmlincs Md isofack ( e w n Md s labelled P 5-knot imervds) fmm which wind speed and direction M diginzed to prrparr the M M M wind field Top OD00 m. Bonm o600 m

a

CSOWM Performance During the Grand Banks ERS-1 SAR Experiment I 4 1

-0
.... .,.

*

n

Y

I.

O w . in N o v e m M IUD1

"

lull

m

n

2 .

I.

fig. 4 'fariadon of wind spcsd as m s a t d by ~m) xlectcd buoys. c o m p d with the wind rpced V l obtained fmm CMC d MMM wind field Mllyaw a the nearest model pid-poim (GP) a w t locadom. Top Buoy 44137 and model grid p i n t 3691. Bonom Buoy 44140 and model grid
p i m 2308.

a

with MMM winds in terms of ME (Mean Error). SD (Standard Deviation) and S I @carter Index) is given in Table 1. The Table shows thar the MMh4 pmedure has reduced the overall bias in the wind speed t almost zm, and the seater index o has been reduced by 4%. There is ais0 an overall reduction in bias for the wind direction. The impact of MMM winds on wave height simulation can be clearly seen in Fig. 5 , which shows time plots of wave heights generated by CMC as well as by MMM winds and comparisons with measured wave hiights at four buoy locations. The wave height simulation is obtained using two versions (1G and 3G) of the wave model CSOWM. In general. the MMM winds have generated wave heights that are in closer agreement with the buoy wave heights during the two-week period; in particular, the wave heights simulated by the MMh4 winds during three active periods (when wave heights between 4 and 6 m were measured) have the closest match with buoy wave heighu. The 3G version o QOWM has produced a f slight underprediction everywhere in comparison with the 1G version and ti may hs be amibured to a higher value of the dissipation coefficient as discussed in the Appendix. A quantitative evaluation of the model in tcrms of various ermr statistics is given

a

42 I ML. Khndekar, R Lalbehary m d a
TABLE . I
A ampariron o

V. Cardone

f a c wvm m ' w i n d s m nvaiiable buoy ~ o ~ a d oiriuti?i c ~ ~ d ~il a n r i c ii l ~ n

CMC Wtnds
S p c d (m 1-9
~ i r r ~ d (*In o

-

.

MMMWmds

Spcd (m

I-')

Didon
ME

C)

Bwy
44131 Ut37 44138 44139 44140 44141 44143
AN Buoys

ME
-0.7 -3.2 -1.2

SD SI
2.6 2.2 3.4

ME

SD
24.1

SI
=

N

ME

SD S I

SD SI

0 4.8 -1.3 0
-0.8

2.8 3.2 2.6 2.6
3.0

26 +2.4 17 -39.4 -5.1 31 29 -6.6 42 r8.0 26 -1.5 27 +2.8

303
37.2 26.7 35.5 733 21.9 41.9

* * *

*

.
*

146 4 . 3 2.2 2.9 146 -1.3 146 -03 2.3 146 4.1 2.1 146 4 . 6 3.3 146 -0.5 23 ,146 4 . 7 2.2 1022 0
2.6

22 +8.2 28.1 22 -18.5 29.1
+2.8 29.4 +7.0 27.4 44 +23.4 45.9 25 -16.7 63.3 23 +11.3 26.2 21

2?

30

-5.7

26

+2.4

40.5

* * *

* *

N
146 146 146

146
146

146
146

1022

ME Mesn Ermc SD Swdard Devidon: SI Scancr I d .N n u 'Nor meaningful.

Number of dam poinn

in Table 2. which shows the verification of two basic wave parameters. namely. wave height and peak period using buoy data located inside as well as outride the crossover node. The first pan of the table shows verification based on the en& period of the field experiment. and the second pan shows verification covering a shorter period when kinematic analysis was performed over a selected region of the Canadian Atlantic. Without the kinematic analysis the MMM winds arc the PRESTO winds as discussed earlier. Several interesting aspects of the verification can be summarized from Table 2:
1) The bias in significant wave height as well as in peak period has been reduced when MMM winds arc used to drive the wave model. 2) The W E and the scaner index values show a definite decrrase with MMM winds wt respcct to wave heights but only a marginal decrease with respect to ih peak period 3) The linear correlation coefficient between model and observed d u e s shows a significant increase wt MMM winds with respect to wave height. but a significant ih and unexpected decrease with respect to peak period.

A closer examination of the verification smtistics reveals that the 1G version of c s o w has slightly outperformed the 3G version for the duration of the field experiment. It may be noted that the 1G version is a well tuned version for the No& Atlantic ( E d and Cardone, 1987); the 3G version of csow is being tuned at present through adjustment of the wind input and the dissipation source t e r n . The ermr staristics of Table 2 arc quite comparable with the error statistics generated by opemtional wave models presently implemented in national weather services of many countries in Europe and Nonh America (see Khandekar. 1989). A state-of-the-an global wave model called WAM (see-The WAMOI Group. 1988) has been running operationally since early 1992 at the E (European Cenm for

CSOWM Performance During the Grand Banks E R S l S A R Experiment I 4 3
Medium Range Weather Forecasts, U K ) the WAM model products are validated ..; against available buoy data over different regions of world oceans. For the northwest Atlantic the WAM model products are evaluated against wind and wave data from about five NOAA buoys. For most of 1992. the significant wave height generated by the WAM model had an RMSE ranging from 0.4 to 0.8 m. a s c a m index ranging from IS to 30% , and a correlation coefficient @eween model and observed values) ran& from 0.70 to 0.90 (Anne Guillaumc. ECMWF p n commun.). The a o w e. error statistics pmented in Table 1 compare quite favourably with those for the WAM model.

..

-

.

c Evaluation o Model and Observed Spectra f
The spectral data, model and observed. arc analysed in terms of wind sea and swell components. wave vectors and two-dimensional (2-D) spectra. The spectral data compared are f o the MEDS Wavec buoy and from the 1G and the 3G versions rm of CSOWM a rhe fine-mesh grid-point numtcr. 2519, whose location is coincident t wt the Wavec buoy locadon. Comparisons arc made at model output times closest ih to four EI1s-1 S A R overpass times, namely, 1500 urc on 14 November. 1500 wrc on 20 November, oo00 urc on 21 November. and oo00 urc on 24 November. and for the period 1500 urc 21 November to 0900 urc 22 November during which there was relatively high wave activity. Figure 6 gives an overview of the observed and the model sea states at the Wavec buoy location and the model grid point 2519 at which the model spectra werc partitioned into wind sea swell components. Figurc 6 was obtained using the 3 version of the model driven by both CMC and MMM winds. In Fig. 6a the G SWH (significanr wave height) time plots show fcsturcs similar to those indicated in Fig. 5 with the S W H for MMM winds showing.somewhat beacr agreement with the observed SWH. As shown in Figs 6b and c the sea s ~ f e more mixed at the was ERS-1 overpass time at 1500 urc on 14 November, was more wind sea dominated for a 24-h period centred at oo00 urc on 19 November. and was smngly swell dominated at the ERS-1 overpass time at oo00 urc on 24 November. However, in Fig. 6c the MMM winds gemmed stronger wind sea and weaker swell components compared with the corresponding components generated by the CMC winds in Fig. 6b. Using the format convention and display of Gerling (1991). Fig. 7 prrsenrs the time evolution of the imporrant wave systems - e bd in terms of wave vectors (magnitude and direction) of significant specnal peak wave energy measured by the Wavec buoy (Fig. 7a) and modelled by the 1G and 3G versions of CSOWM for CMC winds (Figs 7b and c) as a function of period/wavelength and time. The length of the wave vector is proportional to the SWH of the independent wave system and. therefore. gives a measure of the wave intensity; the armw indicates the direction of wave propagation. The waves axe assumed to be deepwater gravity waves so that the dispersion relation L SJ 1.56T2 can be applied to obrain the wavelength L given the wave period T. The buoy wave systems labelled A to H

. ...
I

.. . ..

. .

44 I M L Khandekar,

R hibeharry and V. Cardone

. .
0'

Culuo"

w i i a ----. YM.I~O e.. .. ..

-

War. haiphti with CMC wind8
(a1
BUOY ;uu3

0
GP

: 2474

.A

. h

.Y

1 .

Oavi in Norombar 1001 (Ul7

I.

10

n

a 4

la

I

io
auov uod.iia

li

l i

-

Oiya in Noramber lS9: tUl7
IC)

is

ti

io

n

i .
BUOY :U U O GP : 2508

ii

04
IO
D Y
38

I.

IO

n
BUOY

24

II

Oavi in Nowmbar 1091 IUl7

:44137 GP : 3601

0

10

h

Y

1 .

9 .

m
1091 (UTI

P

Ir

. I

I
1 1

Day. in N0r.rnb.r

with Fig. 5 Vnriadon of significant wave height wnu lim P four selected buoy lomionr corn@ wave heights genmted by I and 3G verrioiu of the w c model. at arid points (GP) 1 ~ ~ G y 1 the four buoy locadow. Lq? Modcl waw heights obtained wing the CMC wind field. Right Modci wave heights obtaind using Ihe MMM wind field.

in Fig. 7a have approximate model counterpan wave systems in Figs 7b and c and
are similarly labelled. Ihe systems labelled MI to M5 have no buoy counterpans. and the systems labelled B1 to B2 have no model countqmts. The figure includes

0

wave vectors of the most dominant spectral modes within individual wave systems. with adjustments to the periods of the chosen spectral modes to form wave paItCmS

i

CSOWM Performance During the Grand Bankc ERS-1S A R Experiment I 4 5

0

Wava haiphts with MMM winds

O

10

U

Y

, I

1s

10

a?
Buoy : 44131 GP : 2680

I
2 .

..

Me"
YMSilO

----.

Oaya in N O V O ~ ~ ~ I 1881 (UT)

D

10

U

Y

I*

1 .

Oavs in Noramber 1991 IUll

ao

aa

24

2s

I

0

m

=

Y

..

1 .

I.

10

n

24

n

-_ .

Oavs in November 1881 IUD

0

3 0

U

Y

1 .

I S

10

P

14

1s

I

Oavs in Nowmbor 1001 IUll

Fig. 5 IConclUdrdl.
. .

that are smwthly w i n g with recognizable features. The wave patterns in Fig. 7 . therefore. represent a "filtered" version of the raw data and show persistence and temporal correlation w t i an individual wave system. ihn The wave systems generated using the MMM winds (not shown here) were similar to those i Fig. 7 generated using CMC winds, the only exception being n the absence of the M5 wave system in both the 1G and 3G versions of the model. The close agreement in general between the wave trains generated using &e two

VvlfICdtiOnperiod OD00 utc 10 November Lo oo00 utc 26 November 1991 EWS inside the clolfm*l M& ME 4 3 -0.5 0 -0.4 -1.9 -. 18 -1.5 RMSE 06 . 08 . 05 . 07 . 27 . 29 . 27 . 25 17 St(% 1 20 2 1 26 28 26 r 0.82 07 .7 0.85 08 .3 04 .9 02 .2 0.2s N 387 387 387 387 258 258 258 B w s ourride Ik cmrsover d e ME’ -. 01 -0.5 02 . -0.3 -0.9 -1.1 -. 04 RMSE 07 . 07 . 0.6 06 . 21 . 2.1 19 . 27 22 2 1 23 23 21 St(% ) 23 I 07 .7 0.80 08 .3 08 .5 03 .5 0.36 03 .4 N 645 645 645 645 645 645 645

-2.0 31 . 30 01 .1 258 -0.8 19 . 21
0.43

645

N

1032

1032

1032
Lo

1032

903

903

W3

903

VuiBcntion period: I200 urc U Nmcmber
All buoys

1200 VIC 21 November 1991

ME RMSE

SI(%
r

)

N
~

-0.2 07 . 2 1 0.74 648
~

-0.6 08 . 26 07 .5

0.2

648

06 . 18 0.83 648
~_____

-. 10 -1.2 -0.7 -. 10 06 . 20 . 22 . 2.1 22 . 1 9 20 23 21 23 0.83 0.48 0.27 02 .2 0.24 648 567 567 567 567
~ ~~

-03

~_______

ME Mean enor = (IINY (model - ob-.

lndu = RMSE x IW/obred mean: I obwmd value: N Nwnber of dam poinu.

RMSE Rmt-Man-Sqm Ewx SI Scauer Linear cornladon coefficient b c ~ model .nd n

different seu of wind fields is an indication that CMC.winds were overall of good quality for the duration of the field experiment. The directional spectra an examined in more detail in Fig. 8 for CMC winds and in Fig. 9 for MMhl winds at four ERS-1 overpass times. Each 2-Dspecfllm is normalized wt q e c t to iu pcak energy value in frequency-direction space. ih The individually normalized spectra an plotted in polar coordinates in a linear hquency format with concenmc circles at 0.075. 0.150. 0.23 and 0.300 Hz inrm creasing f o the cenm. The direction of wave propagationis the compass direction with n o d toward the top. The normalized wave energy is shown in the direction to which waves arc propagating in relative uniu for values beween 0.05 and 0.95.

CSOWM Performance During the Grand Banks ERS-I SAR Experiment I 47

E
0

Buoy :WAVEC GP : 2618

:

IO

11

I 4

Day. in Normbar I 8 (UTI 81

(I

II

10

za

u

I
IS

IO

U

I*

IS

IO

u

Davr in Novmbrr IEEl (UTI

f i g 6 Tm sen- of model Md observed w a n paranelm g e n d by the 3G m i o n of o o w i using both CMC and MMM wndi P gnd point (GP) 2519 corncrdcnt the W a r n buoy !O. -n (a) Model Md observed s1gnIfisnn1 wave herghu: wth wmd Sea and S W I I w a n hclghu usmg (b) CMC d. (c) MMM wnds. and

The top rows of Figs 8 and 9 give the buoy specma. the middle rows the model 1G specaa. and the bottom rows the model 3G spectra. At I500 UTC on 14 November the sea state was more swell dominated for CMC winds but more mixed for MMM winds (see Fig. 6). The buoy specaal s i g n a m indicates the presence of thrte wave systems. namely, a weak northward propagating swell. a weak west-northwesfward propagating wind sea, and a more intense southsouthwesfward propagating swell. Thesewave systems can be identified in the

2 0

000

i 0

D

10

11

I 2

Ia

u

le

le

n

18

8s

ro

a1

P

a$

2 .

ae

ae

z r"

d
0

" 2

E

UP

.1: -

f
D 0

z
IO
11

>

la

la

u

is

18

Day8

ao 21 in NOVembar lee1 IVn
n
18 IO

P

13

a,

am

ae

*"

fig. 7 lime evolution of (nw s y s u m s given in unm of wave MW(S during lhc field cxperimcnt at. the locarion of csow grid point 2519.as (a) obwned by the Wnvcc buoy, and PI gmerarcd by the (b) I and (c) 3G wnions of sow. 7he vmon Rprexnt lhe magniNde and dirccdon of G significantp&s lhc directional wave c n c g l S ~ Z C I ~ . in Syspms B1 and 82 W+R WI g e m by the model. and sysums MI to M "cm M I ohwMd by the buoy. S

CSOWM Performance During the G a d Banks ERS-1 SAR Experiment I 4 9 rn
IS00 UT I 4 No. IS00 VI 20 Nor
d)

0000 UT 21 Nor

0000 UT 24 No7

@.
k
..

!

wave analysis chan of Fig. loa valid at 1200 m on 14 November 1991. The southward moving swell. identified as system A in Fig. 7 appeared as a . unimodal system in the modcl'spectrabut as a multimodal system in the buoy specuum. The 3G version of the model spenra gave a b m representation of the c thrce wave systems than the 1G. In Fig. 8c for CMC winds the wave systems have relative intensities comparable with those measured by the buoy; however. in Fig. 9c for MMM winds the west-northwestward propagating wind sea appeared more intense in rrspow to a stronger MMM southeast wind. At the overpass times at 1500 m on 20 November and oo00 m on 21 November (about 9 h apart), the wind sea component 'bas more dominant for MMM winds than for CMC winds (see Figs 6b and c). Figurcs 8d-i indicate three major wave trains by the buoy, nvo by the 1G. and four by the 3G and am identified as wave systems D. E. F and M5 in Fig. 7. The south-southwarward travelling wave train (system D in Fig. 7) was the swell system located just south of Greenland in the
METOC

I
. ,

50 /
IS00 UT I4 Nor

M L Khandekar, R Lalbeharry and V. Cardone ..
IS00 UT 20 Nov (d 0000 UT P I Nor,

(e)

0000

UT,24

Nov

0

Fig. 9 Same as Fig. 8 but using MMM win&.

term analysis in Fig. 10b valid at oo00 m on 20 November 1991. This swell with

a period near 12 s arrived at the buoy location some 15 h later. As s e n in Fig 7b and c. it was absent altogether in the 1G after oo00 UTC on 20 November but showed persistence in the 3G.The eastward propagating wind sea and the southsoutheastward propagating swell (systems E and F in Fig. 7) were also evident in the manalysis in Fig. 1Oc valid at 1200 m on 20 November 1991. The north-oonhwesnvard moving swell (system M5 in Fig. 7) in the 3G grew somewhat smnger by oo00 m on 21 November. It was completely absent in the buoy spectra, whereas in the 1G it was weak and is barely visible in Figs 8e and h. It appeared in the 3 around oo00 m on 20 November and continued to exist until around 1800 G m on 21 November. There was no source region for this system as generated by the 3G for the CMC winds. However, this north-northwestward mvelliing wave system generated by CMC winds was absent in both the 1G and 3G for MMM winds as seen in Figs 9e-f and 9h-i. This fourth wave nain may, thmfore. be the result of wind errors present in the CMC wind field assuming that the MMM wind field was more realistic. The measured and model 2-D specm at oo00 UTC on 24 November sbown in Figs 8k-1 for CMC winds and in Figs 9k-1 for MMM winds illusmu the case

.- . .

..

...

CSOWM Performance During the 'Grand Bath ERS-I SAR Experiment 151

0
.-

of an almost fully Swell dominated sea (we Figs 6b and c) in which the major wave nain was the southward travelling swell labelled as system G in Fig. 7. In the MEIW: analysis in Fig. 1Od valid at 1200 UTC on 23 November the wind sea and swell in the B a n BayLabrador Sea area arrived at the buoy as a swell some 12 h later. This swell in the buoy spectrum exhibited multimodal spectral peaks with periods ranging from 7 to 12 s. The weak east-northeasward moving wave train, appearing in the 1G and 3G for both CMC and MMM winds but not in the buoy'specaa. was not dissipated rapidly enough by the model since this was an earlier sweil system moving towards the northeast that decayed completely before reaching the buoy and was, therefore, not detccted An examination of model and buoy specha for the duration of the field experiment but. not presented here showed that the spectra for MMM winds were generally in closer agrxment with the observed spectra than the corresponding s p e m generated.using CMC winds.
4 Summary and condusions

In this study two msions (a first generation and a third generation) of the operational wave model WWM were used to simulate sea-state conditions during the Grand Banks ERS-I S A R wave spectra validation field experiment Two sets of wind fields were u k d t drive the.model in a hindcast mode: one set of wind o fields was obtained from the RFE model of CMC in Montrial and identified as CMC winds: the other set of wind fields was obtained through a hand analysis of 3-hOurly surface pressure cham and an application of a marine planetary boundarylayer model and identified as MMM winds. The model wind and wave products were evaluated against buoy measurements available inside and outside the field experiment area in the Canadian Atlantic. The evaluation suggests that both msions of CSOWM performed reasonably well in the hindcast mode and generated error sratistics that were quite comparable wt ih those generated by some of the operational modeis implemented at national weather services of many counmes in Europe and No& America The wave model produns generated using the MMM winds had the closcst overall agreement with buoy measurements and this in turn produced improved error statistics compared wid^ the corresponding error statistics generated using CMC winds. The wave heights generated using MMM winds during the active periods of the field experiment showed a closer agreement with buoy-measured wave heights. The 1G version of QOWM has slightly outperformed the 3G version for the duration of the field experiment; funher the 3G version has consistently undcrpdcted the sea states during the active periods o the field experiment compared wt f ih the 1G version. The present 3G version is being tuned through the adjustment of vadous source terms. The dissipation tcrm (Eq.(A-17) in the Appendix) used in the prcsent 3G version appears to provide a realistic sjmulation of extreme sea states associated with intense storms in the Canadian Atlantic (Kbandekar. 1994). For moderatc sea states that prevailed over the Grand Banks area during the field

52 I M L Wand*,

R Lalbehay and V. Cardone

10

w a n Malyru valid 81 (a) 12M) rm on 14 November. (b) OD00 rm on 20 Nonmhr. (c) 1200 rm on 20 November. and (d) 1200 rm on 23 Novgmber. C o n m E w a n i m ( ~ . mows indicate wind sed% a d wavy y m w s indicate rwsb. n rotid n
M -

0

. ’

e

CSOWM Performance During the Grand Banks ERSl S A R Experiment / 53

e
Fig. IO lCwul&dJ.
-.

54 /

M L Khandekar, R Lalbeharry and V. Cardone .

experiment the dissipation term of (A-17) may have produced increased dissipation, thus leading to consistent underprediction of wave heights by the 3G version em compared with the 1G version. A wind-dependent dissipation t r together with a revised drag law of (A-16) is being considered for further testing of the 3G version. Both the 1G and the 3G versions of QOWM have simulated wave trains in fair agreement with the buoy-measured wave trains for the duration of the field experimeni. The g e n e d o n of four instead of three wave systems by the 3G version using ie CMC winds at two of the ERS-1 overpass t m s may be amihuted to emrs in the CMC wind field since the MMM winds generated three wave systems in agreement wt the number of wave systems measured by the buoy and as diagnosed by m ih chans.The more accurately analysed MMM wind field produced 2-D wave spectra that in general were in beaer agreement with the buoy spccna. The results presented here have provided a degree of confidence in the use of wave model products generated by the two sets of wind fields. The two sets of wave model products provide a possible range of variation that can bc expxted assuming of coune that the wave generation and dissipation mechanism i adequately r p s e resented by the wave model physics. The two sets of CMC- and MMM-generated wave specua can provide a useful fint guess for S A R data assimilation in wave models. Acknowledgements We wish to express our sincere thanks to Dr Fred Dobson of the Bedford Insti~tc of Oceanography. Halifn, for his valuable suggestions on various aspects of this study. Thanks are also due the Meteorologist and Computer Analyst of oceanwcather Inc.. U.S.A.. for their assistance in the preparation of the surface pressure and wind field analyses. IO Mr Arun Bishnoi of the AES Forecast Research Division for his assistance in data archiving and retrieval. and to Minh Nguyen and David Brown, Coop Sbdents h m the University of Waterloo. Ontario,for their valuable computer suppolt. We are also grateful to mCenm. Halifax, for makii~p available surface and wave height cham for the duration of the experiment Finally, our appreciation is expressed to Ms Ruth Tung for expertly w i n g the manuscript. This study was pardally funded by the Panel on Energy Research and Development (PERD) of the Government of Canada. Appendix: Definitions of wave parameters and specilication of source terms a The Wave Spectrum The two-dimensional wind sea specrmm associated with all grid points of the model is expressed as

EU, = E W G e)

-.

(A-1)

Here EV) is a hquency dismbution for a fully developed s e a and G is a'normalked angular spreading function that depends on frequency. relative angle between

CSOWM Performance During the G k d Banks ERS-1 SAR Experiment I

jj

0

wind and wave directions and wind sped. The function G is applied to the wind input term in (A-11) to obtain the two-dimensional PM s p K w m and also to the partitioning of the total energy into wind sea and swell components. The significant wave height (If,) at any grid point of the model is given by

HI = 4.0&

(A-2)

where 'E is the total wave energy defined in (3). The peak period, T, is defined ,, as the inverse of the peak frequency. f. which is the frequency at which the wave , energy is maximum for the onedimensional energy spccrmm.

b Energy Pomrioning INO' Wind Sea and Swell Components Each spectral component is made up of a wind sea and a swell component To sepmte the swell energy from the wind sea energy, a onedimensional PM s p m m for a fully developed wind sea is applied. Equation (4) when integrated over all frequencies yields
.. .., .

!.:>fT:

EPM = 0.2849 x 1O4C&

64-3)

The significant wave height corresponding to this value of wave energy is given by
Hs(p,w) = 4 . 0 6 = 2.12 x 10-2
UlpJ
2

(A41

-. . ...

A wind speed of 10 m s-I gives a fully developed wind sea with a significant wave height of 2.12 m and a peak specaal energy at a frequency of 0.1372 Hr Thus. with a specified wind speed, the onedimensional specuum can be esrimatcd assuming fully developed s a Given the wind direction, cacb frequency component of the e PM specaum i then spread over all wave directions using an angular spreading s function that is propomonal to the fourth power of the cosine of the relative angle b e o m n wind and wave directions. Having defined the Wind sea specrmm as given by the PM spectrum, the spectrum of the swell energy is then given by

E ~ Ue), = EU, 8)-

8)

(A-5)

where thc subscript " S W denotes the energy of the swell component Summing over all hquencies and directions. the total energy is obmincd as

E = E ~ +E,; ,

thus H I = H:,+H;

('4-6)

In (A-6) the subscript "Ws" tefers to the energy of the wind sea component: the wave height for the total sea is given by the square root of the sum of thc squares of the wave heights for the wind sea and for the swell. respectively. The PM specuum is also w d to define an initial wind sea specrmm wbow specaal energy values are spread in dflerent directions as discussed earlier. Some wave models use a JONSWAP spectrum (see Hasselmann et al.. 1973) to define an i i i l Wind sea sp+crmm. nta

56 I M . . L
c Source T e r n
1 IGSOURCETERMS

Khandekar, R Lalbeharry and V. Cardone

0
(A-7)

Based on the landmark studies of Phillips (1957) and Miles (1957). the wind input term S .is expressed as i

s ,

= AU, e.

u6)+

BU, U.)EU, e) e,

Here the term A q m e n t s Phillips’s resonant interactions between waves and atmospheric turbulent pressure fluctuations in the overlying atmosphere. and the term B represents Miles’s exponential growth term. The term A. which is commonly identified as the linear growth term. is exprcssed as a function of wind speed (U6) ai the 6-m level: the term B is a function of U. the friction velocity. Based on an . expression for h e atmospheric nubulent pressure specmm proposed by Priestley (1965). the linear growth term has been formulated in terms of model parameters 0 k. (@ e), and . computed from the logarithmic profile for the wind speed given U and uI5.3. as .

-

Acf, 8,

u6) =

(‘4-8)

where. w = 2xf is the angular wave frequency and k = 2x/L is the angular wavenumber with L being the wavelength. The total contribution of the A term af any grid point and at any time step is obtained by integrating (A-8) over all wave dmctions, 8. and over all frequencies,f.The expression in (A-8) is a modified version of the linear growth term developed by Pierson et al. (1966) and Inoue (1967) and has been used in a wave model developed by the Meteorological Research Institute. Japan (see Uji. 1985). The exponential growth t r B is essentially the em same as that developed by Pierson et al. (1966) and is written as

B

- = {5.oC~p[-7ooO(ARG - 0.031)2]
0

+ 2612ARG2exp[-0.0004(ARG)-’1}/3600
(A

- 9, 0

In (A-9). ARC = (. cos(@-e)/C. where C is the phaw speed. and the expression I inside the braces represents the growth rate per hour:The friction velocity U is . obtained from an empirical relationship between &, the roughness paraniccer. and

u. as

CSOWM F'erfo~rmanceDuring the Grand Banks ERS-1 SAR Experiment I j 7
01 & = -+azu:+ao

U.

(A-IO)

Hm

ao, a1 and a2 are empirically determined constants (see for example. Arya. 1977); The solution for U is obtained through an iterative procedure using (A-IO) . and the logarithmic wind law for neutral stability, given the wind speed U19.J. Thc.pwtb equation for the 1G version of QOWM is expressed as

(%)- =

(A + B E ) 1

[ - (&y]

(A-11)

Equation (A-11) allows the transition of the wind sea to fully developed sea as defined by the two-dimensional PM spectrum obtained using a spreading function of the form suggested by Pierson (1982). The dissipation t r Sb in the 1G version em represents the rate of loss of energy due to wave breaking in deep water. The dissipation of wave components in the upwind direction and down to 90" on both sides of the upwind direction is effected through an exponential t r given by em

E ~ v ; ,e = EOU, ) , e,)expc-KfdW

(A-12)

..

In (A-12). EO is the speed energy component travelling against the wind before dissipation, ED is the energy component after dissipation over one time step. E, is the energy summed over all frequencies and those wave dmtions lying within f W Oof the wind direction, N is a dimensionless power with a value of 7 in the upwind direction dccrcasing t 3 in the outermost directional band. and K is a o coefficient that controls dissipation and depends on the growth and propagation time step. Morc details on the dissipation term can be found in Pierson (1982).
2 xi somcF.TERWS In the 3G version of CSOWM, the wind input term is based on the study of Snyder et al. (1981). which obtains an expression in terms of U, (wind speed at the 5-m level) and phase speed C. The wind input term is written as
S,=A'+PE

(A-13)

In (A-13) A' is half the value of A given in (A-8) and this allows the model to f o a flat sea without the need for specifying an initial wave energy field rm based on a PM or a JONSWAP spectrum. The growth rate coefficient B is expressed following Snyder et al. (1981) as
start

(A-14)

Here (+ - 8) is the angle between the wind vector and wave propagation direction.

58 /

M L Khandekar, R Lalbfhay and V. Cardone

o = 20f. and p. and py are the densities of air and water, respectively. T k n ai
U as an appropriate fraction of UIO(wind speed at the 10-m Ieucl) and using a 5 drag coefficient representative of the measurements of Snyder et al.. an expression for B in terms of U has been proposed by Komen et al. (1984) as .
B=max

0,025-

::
C
6.2 m s-'

COS(+

- 8)-

1

11

(A-15)

In (A-15). the coefficient 28 is subject to a slight variation depending on the r p e resentativeness of the drag coefficient. The solution of U utilizes the neutral drag . law proposed by Wu (1980) and modified in this sNdy for low and high wind speeds. The modified drag law is given by
C = 1.2 x d

U s 6.2 m s-'

= (0.80+0.65U) x
= [(0.80+0.065U)

< U S 14 m S-I

- O.O021(U - 14)'l x lr3, m s-' < U S 29.5 m s-' 14 = {(0.80+0.065U) - 0.0021[(U - 14)' - (U - 29.5)']} x I-, O'
U > 29.5 m s-'(A-16)
In (A-16) U refers to the wind speed 81 the IO-m level. Given the logarithmic wind law. the wind speed U,~J. the relationship U = U a .an iterative procedure and . I is used to obtain the appropriate (. for (A-15). The drag law proposed by Wu (1980) is found to produce higher wave growth at higher wind speeds. Equation (A-16) allows the drag coefficient to increase asymptotically to a limiting value for exmme wind speeds. Such an asymptotic behaviour of the drag coefficient has been hypothesized by Frank (1984) in a study relating circularion of a marurr hurricane. With (A-16). CSOWM appears to produce a more realistic wave p w t h since the drag coefficient tends to slowly converge to a limiting value as the wind speed becomes more exmme. The non-linear wave-wave interaction term S., is explicitly calculated in the 3G version using the discrete interaction appoximation first proposed by Hasselmann and Hasselmann (1985) and applied here to the full two-dimensional spectrum of 23 frequencies and 24 directions. The discrete interaction approximation provides an efficient numerical technique for explicitly calculating the non-linear interaction term S. Nevenbeless. the inclusion of the non-linear term significantly inmesses , the computation time required to obtain model products using the 3G version o
CSOWM.

0

A

The dissipation term S h represents. as in the IG version, the energy loss due to wave breaking in deep water. The whitecapping model of H a s s e l m a ~ (1974) has been modified by Komen et al. (1984) to obtain an expression for S* in terms

i
CSOWM Performance During the G m d Banks ERS-I S A R Experiment I 5 9

i
1

of a. the angular wave frequency and a an integral wave steepness parameter. The expression by Komen et al. has been modified h m by increasing the dissipa- tion constant and by changing the dependence 10 instead of 0 . ctr' ' The modified. expression is given by
(A-17)

In (A-17). the panmeters 6 . Group (1988).

and

~ P M arc

the same as defined in The

WAMDI

References
IPS. 1957. Suggutd d o n s m cenain boundary layer pnrammizotioo schemes uwd in armosphsric ckcuhion models. Mon wcnrhcr R ~ IC 215-in. . C*ILDONE VJ, 1969. Specifisarion of the wind distribution in the marine boundary layer for r ~ y forecasting. Gmpbys. Sa. Lab. TR-69-1. e school of Engineering and Snencer New Yo& Uni-cy, New Y o h 118 pp. . 1992 on Lhe sullmue of the marine surfac+ wind field i urranopical stomu. Rcn prim. Tiid I ~ Worfrrhop on Wave HindssriL i n g a n d n d n g . I9-22hlay IWZMontrtaL w c Armm. Envimn Sew. Dovnrvinu. br ON pp. 54-66. ; WJ. RE~WN and LG. w m . 1976. Hindcasting the dirc~tionalspccna of hurricanegenerated w i d . J. PcrmL T c h L 1 : 3858 394. , JA GMWD and HI c m 1990. On mends in historical marine wind data J. Clim 3: 113-127.
*IIY*.

Mntal Smdicr RcwMh h n d s k p . No. 076. 0taW.oN. 111 pp. W H 1984. A Composilc analysis of the a r e of a matwe hurricane. Moa Wearher Rev. 112: 2401-2420. GEPLWG. T.W. 1991. A com-ve matomy of Lhc LFYEX SySrCUlS. In:DinctioMl Occm

-

W v Spcom RC. B d (d) The J o h Hop ae E.. khs Univmiy h pp. 182-193. .
.G I: * VJ. URCCNE and LM UWSON. 1985. LuUc0-n mt -io" of the S A L wave nmdd. In: Ocean W v Modelling. The ae n w Gmup. Plmvm Chap. ZZ pp. 221-233. v HUSEWN. K 1974. On the -pal h i p a n o n of o ~ n due m whir^ w i n g . Boundaryk-r McrcomL 6: 107-127. et aL 1973. Mcaurrcmtm of wind-wave gmwh and decay during the IoiN Nonh Sea Wave h j (JONSWM). ~ Dlrch Hydmgr. Z Re& AI2.95 pp. M U . and K M E L M A N N . 1985. CornS. pulationr and parameterimion of the nonlinear e &cr in I pvicy-wave ~p"mun. Pan 1. New method for efficient compluadons of the uart nonlinear m f uintegral. 1. Phys. Oc-~r. Is: 1369-1377.

-

-

.

DUE.

Y.

1985. Surface mrbukm Bux fannula-

don in smble conditions for a m p h e i s circulation d L . Mon Wmrhcr RN. 113 89-98. . 1988. The poridon of Lhc IoVcIt leveb in the boundary layer of the armolpheic cimlation models. A T M O S ~ 26: 329-340. .
DOWN.

MWET. I967.Onrheg1mw~hofthcrpecmunof a wind gcnuaud sea according m a modifid

Milcr-Phillips mshsnrrrn and is applicadon m ' F.W. and P.W. VICXON. 1%. ' I h C Grand wave f o m t i n g . TR-67-5. Gtophys. k i . Lab. ERE1 SAR wave specp. didadon exk p . . New York Univ.. %mi of Eng. Sci. perimmt:hgramOvminvanddaQd.8~. U U N D E u l L HL 1989.0pmiOMl AMiySU ond A M S W X. T OF E32 (rhir Lnu). Prediction of Ocean W n Wnw.Coastal and id EID. n and VJ.CARCONE n 1987. Opentional KSI of wan-forssPning models during the C ~ M - Estuarine Studies No. 33. SprinpVcrlag. 214 pp. divl Arlnndc Sih - (CASP). Envimn-

E &

60 I hU. ghandekar, R Lslbebarry and V. Canlone

.

-

1994. smrm w a r n in the CaMdian AI. lauric: A numcrid dmuladm In: AdvMccr in N m d Md Tecbno&#icnl H a w d s Resewrh, VoL 4. Kluurcr. Cmdrcch~ Nsthulandr. lix and a LILBEHMRY. 1990. A f o m t u ' s -guide l the Canadian speed o o c wave model. Int Rep. ARD-9o-M7. Amos. Envimn. Sm..hwntvicw. ON.41 pp.

G . 0. H S J; Aand L IUSSELMLMI. 1984. On the exismu of a fully developed wind-sa spcsrmm. J. P e s . OceMogr. 14 1271-1285. MILS. J.W. 1957. On the g a d o n of surf= waves by shear B w . 1 Fluid Mcch 3 1 8 s .

204.
HTlcHaL KL: C ClUREHE SJ. LAMBERT: 1. Hw

ad c QIOUM*PII 1993. Thc Canadian global dam -miladon system Description and cvald o n . M a Wmher Rev. U :1467-1492 1 -.OM 19S7.0nthe gaerdrion of w a r n by nvbulrm w i d 1 Fluid Me& 2: 4 1 7 4 5 . . masord. WJ. 1 8 : 9 2Spectral OCW Wave Modcl (SOWM): A Nonhrm Hcmispphas cornputu mcdcl for spscifying and forecasting ocean wave s p x n F d Rep. No. DINSRDC82/11. US.Naval Oaanography C o d DsrachMtu. Ahvillc, N.C.. 201 pp. nnd I MOSKOWIIZ 1964. 'A pmpmcd s p s n n i form for fully developed wind seas

b+rcd 011 the S&kriV mcOry Of S A Ki pmdskii. 1 Gcophyr Res. 6 :5181-SIW. . 9 ; u.ncx and L 81966. computu bpvd pmcedurrr for przpming global wave f o m o and wind field analyses capable of YIing wave dara obtained by a spandsh. Ra. Sixth Naval Hydmdynami~ Symp.. Off. Naval Res.. Washington. D C .pp. 499-532. .. m m Y . 1T 1965. Correlation mdiu of .. pusus Ausruodonr on the gmund beneath a turbulent boundary layer. National Busnu of Srandds. NBS Rep. 8942.92 pp. SANDEW. F. 1%. Slid= analysis O w the O C U l U ScYEhing for sa mrh. W m k r and Fore. &a. 5: 5 w 1 2 WmElL ILL: F.W. DOBSON.JAU Q T and R.8. LONG. 1981. A m y meauvemunr of armmpheric pluUUC fl&N aha- rvrfosc giXnty warn. 1. Fluid MIclr 102: 1-59.

-

-

UT. 1985.ThcMRIwavemodcI.In:Occnn Wave Modrllin#. The W*MP Gmup. Plenum Qop. IS. pp. 157-165 THE WAMDI GROUP. 1988. WAM model A third genemion ocean wave prcdidon mode 1. Phys. Oce-gr. 1 177S-1810. % w. J. 1980. Wind s u e s caffiicicntsover the sa surface during llcar neu1181 conditions A revisit. 1. Phyx OcCMogr. 10: 727-740.

-

0

-

I
Reprinted hum J ~ ~ R N M ~ F A T M O S P H W CN D ~ C T E C X N O L O G Vol. 13. No. I . F e b w 1996 A Y.
A-M-m

Evaluation of Contemporary Ocean Wave Models i Rare Extreme Events: The n “Halloween Storm” of October 1991 and the “Storm of the Century” of March 1993
V. J. C-NE
Onmwather I%.. COSm,Conneaicui

R E. JENSEN D.T -10 .
V. R. S w m
A.

T.Cox

B

. .

.

.. .

F
.<I. I.
3.: 1

.......................

..,I

.........

..

. . .

~

. -

.-

-*

Ii

198

J O U R N A L OF A T M O S P H E R I C A N D O C E A N I C T E C H N O L O G Y

Vowve 13

...
...

. . ..

..

. .. ., ...., ,.
!..

8. "

0 1996 Amaican MeoomlOgiCal sosiny
.

. . -............

. . . .

.. .

.......

0

FEBRUARY 195%

C A R D O N E ET A L .

199

i.

..

e
I

model (Resio and Perrie 1989). and the 3G WAM4 (WAMDI 1988: Gunther et al. 1991) arc adapted to the western North Atlantic basin on grid systems with common resolution and am driven by a common wind field derived for each storm. The wind field was developed using a subjective-objective man-machine mix approach, which w d all conventional data. including ship and buoy observations received too late for use in reat time. The al(anao've wave hindcasts were evaluated against time series of measured significant wave height (HS). dominant wave period (TP). and onedimensional (frequency) wave spccua obtained at nine U.S.and Canadian buoys moored in deep water beween offshore Georgia and Newfoundland. Wave models have been compared in previous studies. For example. the behavior of about a dozen 1G and 2G spefnal wave modcis were evaluated in several standard test cases and intercompami over a decade ago in the SWAMP ( 1985) cxmise. Large differences in behavior WQC found even for the simplest consrant wind fetch- and duration-limited growth tests. suggesting large diffcrcnccs in behavior as well in rcal applications. In a more recent intercomparison study. nine North Atlantic wave models (one 1G. seven 2G. and 1 Introduction . one 3G rypc) were applied to hindcast an 1l-day period Numerical o e n wave m d l have advanced sig- of the Labrador Sea Exacme Waves Expuimcnt ca oes nificantly within the past decade, parriculady with the (LEWM) using both model-specific wind fields and a inucduction of the so-called third-gcnmtion (3G) common wind field. with hindcasts evaluated against class of m d l . First- (IG) and second-genedon in SIN and airborne remotely ScIlsed directional wave oes (2G) models have also been improved and remain in specaa The sea s m regime of LEWM was swell widespread use for climate assessment, engineering dominated and the skill of all hindcasts appeared to studies, and operational forecasting. These three wave model classes am differentiated mainly by the simulation of the physics of dccp water wave growth and decay. Due to r m i i g uncnrainties in the underlying eann physics, however, all models rely to some deon empirical uming, based mainly on observations of wave growth in stationary fetch-limited wind fields of moderate strength. Such models are then oftcn applied c 1.4 ..; _ _ . . _ . ... _. to detcrmine the extreme wave climate in hmh basins -. . ... ._ ..- ._\..... .. .. for design plnposu and to provide guidance in rcal W 1.2 time to shipping for optimum ship muting and heavy 1 ' I weather damage avoidance. An extensive suite of wave 0 15 measurements made in two very extreme western North Srncoth Hs" Atlantic storms from buoys moored in deep water provides a rarc oppdnunity to vaIidate these models in wave regimes far removed fmm those used for model IPOZ)UOQ.O.W~I~~~R'~I c tuning, but comparable to rhwe encountered in the y 1.6 model applications citcd above. LL In this SNdy, four widely applied spectral wave models, namely Oceanweather's 1G (Cardone et al. 1976) and 3G (Khandckar et al. 1994) m d l , a ZG wave oes

Y

1

:: ..

I

-

.

i

..

200

J O U R N A L OF A T M O S P H E R I C A N D O C E A N I C T E C H N O L O G Y

VowMe I3

e.

44142
44141

HOS.SOC

44140 44139

HOS
HOSSOC

1500 4500 1400 1 I00

. 4 m . 0 ' ~ 5609.1'W
42'43.8'N 44'19.2'N 44'13.8'N 41'11.6'N 40.15.O'N 36'35.0'N 4Tll6'N 38.47.6'N 4104.9'N 38-n.O'N 4(P30.O'N 4331.6'N 4237.9'N 3F3U'N 24'383'N Zk518'N 2870.O'N 30'43.CN 3 m 32'17.7'N 34'553N 5036.4'W 577.ll'W 5338.0'W

4T29.6'N

6Q.IZO'W

5.4 m

5.4 m 5.4 m
5.4 m 5.4 m

44138 44137

HOS.SOC HOS.SC€

I500
4500

C C C C C C D D

C6NOI C6NlZ C6N09 C6N06 C6NIO C6NII 3D21 3D3 lL3DZO LNBll

La Have B M k
h m d m Fan Tail ofthe Bank

Bpnqucrrnu

SW G a d Bsnks
k l SCOllan

40. 425

40
48

44014

SCC
HOS
H S O

44013
44009
44012 44011

30
24 88 28

44008
MMn
eiM)s

60
47
202

soc

44004sOC
41016 41010 41009 41008 41036

3231 1586

833
41 18

S W
HOS.SG€

1042

41004 4lOU2
41001
I ,

z9w.m

6107.8'W 73'lOa'W 74'50.0'W 70'46.6'W 74'34.6W 6612.6W 74'420'W 6925.7'W 7005.4W W33.l'W 7W423W 76.313W 78lZO'W 80'10.8'W BI'M.8'W TN9'W

5.0 m 3.5 m
13.8

5.4 m

m

13.8 m
5.0 m 13.8 m 13.8 m 13.8 m 5.0 m 5.0 m 10.0 m IW5 m 5.0 m 5.0 m 5.0 m 5.0 m 5.0 m 5.0 m

D D D D
D

LNBOI
6N16 LNB06 LNB07102

Long Island Virgmta Beach Boston Harbor Eve Fathoms

SIW

eager B d

D e l a m Bay

D D D
G D

LNBo9
6N02 6NI916N29
12D9

NMNC~C~ Ponknd G d f of Maim "HM'
Elcuthcn ECapcCMsvd

79.I-W
75.14.4'W 7757.1W

D D D D D

3658
4206

WD

HOS

IODU6N28 3D17/10Dll 3D16 6N06 3m9 6Nu 6NOI

Cape Canavcral E Jackmnvillc

E Dapom
mt so
S. Ctpc Hancra~

E*

. .

Hl:6N-6 ul

m N O W . 3D--3-m diwur. LNB--Largs Paylo& C-Canadk D-DA(3T. G-GSBP.

navigational buoy. IZD--12-m discus: IOD-IO-m

discus.

..

.. ..

suffer from unavoidably large errors in the wind field in swell source zones, which happened to be located outsidc the N o d Atlantic shipping h c s (Beal1991). The evaluation of the 1G and 2G models that have been applied opcrationaUy for real-time analysis and forecashing OVCI the past decade (e.g., Clancy and Sadler 1992; Guillaume 1990) and in historical hindcasting studies (e& Cardone and Ewans 1992) indicate greater and more uniform skill than implied by the SWAMP experience. Typically. in Nodem Hemisphere basins. and when driven by conventionally PICduccd wind fields, time series of HS arc specified with bias of lcss than 0 5 m rms difference ranging between about 0.75 m i summer to 1.25 m in winter. and scatter n index (mi0 of rms difference to mean of verification data sample) of 25%-50%. S t peak puiod is wd usually specified with bias of less than 1.0 s and standard deviation of about 1 5 s S t a t k t i a of wave dircc. tion skill arc still rarcly reponed. There is a $rowing mnd toward implementation of the WAM model (WAMDI Group 1988) for global and regional analysis and forccasting [e.&, the global implementation of cycle 4 of WAM is described by Gunther et a. (1991 1; the Nonh S a implementation l of WAM known as NEDWAM is described by Burgen (1990); and the Mcditcrranean Sea version of WAM run at the U S Navy Fleet Numerical Meteorology and .. Oceanography Center (FNMOC) is described by C h c y and Sadlcr (1992)l.

WAM4 w s also used to hindcast an Il-day stormy a period sampled during SWADE (Cardone et d. 1994). driven by six altcrnadvc wind fields. One of the wind fields w ~ S car~fullydCV~10ped mainly haDd kineby matic analysis with the aid of the SWADE meteomlogical data the other five were produced by objective analysis with b e e of those five a c d y the standard products of operational centers. The winds derived by manual kinematic analysis provided the lowest wave hindcast errors reporrcd thus far. Tbe scatter index varied between about 10% within the SWADE m y to about 20% on the open sea boundary of thc buoy array. Errors for the objectively analyzed winds were ~ W Oto four times greater thaa for the kinematidy dcriv wind fields. This tinding is conditioned on the rath modest range of sea states available for model verificaa'on in SWADE where b e peak sea state mcar& was about 8 m.Since WAM4 requires a supercomputer for most practical applications, it would be of intenst to quantify the performance of highly developed 1G and 2G m d l relative to that of 3 models for meoes G teorologicd forcing covering a wide dynamic mp. Such a range is provided by the two well-documentcd storms studied hemin. O the many intense exhatropical storms generated f off the cast coast of N r h America within the last rhne ot cold seas~ns (since 1991/92), fwo BTC also characterized by sea states of unprrcedcntcd magnitude: the "Halloween smnn" of 26 October-2 November 1991

m

F ~ R U A 1996 RY

C A R D O N E ET AL.

:

201

80’

-75’ .

-70’

-65’

-60’

-55’

-50’

FIG. Tns* and m d ~ R U L V C of Ihc Hdlowten swnn and wuthsm I m p i d syarcm. 3. u

(HOS) and the ‘*stormof the century” (SOC) of 1215 M r h 1993.At Environment Canada buoy 44137, ac moored in deep water south of Nova Scotia the measurui peak significant wave height (HS) exceeded 1 6 m in both storms, with maximum crest-trough amptitudes exceeding 30 m (nearly 100 ft). These wave heights exceed current estimam of 1y r m peWr e r wave height exmmes in deep water south of Nova i d Scotia ( E d et al. 1992)by about 5 % Ar NOAA buoy 0. 41002,m o o d east bf South Carolina, the peak mea, sured HS in the SOC was 15.7 m an all time record high for NOAA buoys and again exceeding cumnt estimates of design wave heighu in that area by a wide margin [c.g., the WIS estimate (Coma et al. 1981) for this site is about 1 m for the 1y condition]. We 2 Wr shall refer IO sea nates of such magnitude (say HS. greater than 12 mor so) as extreme storm seas (ESS). The immense single muximwn waves t a may occur ht within such sea states have b a n dubbed recently exm m e storm waves (ESW) (Nickemon 1993) as an alt e d v e t the m m popular descriptors “rogue o waves’. and “freak waves..’

This paper is organized as follows. In section 2 we give a concise description of the meteorological and wave data refemd to i the study of both storms. Secn tion 3 describes the wind field analysis methodology and the evolution of the major kinematic properties of the generated wind fields. Section 4 gives a brief description of each wave model used and the strategy of the modeling experiments. Section 5 presents the assessment of the accuracy of the alternative wave hindcast models drawn from evaluation of the wave hindcasts analysis against buoy measurements. F d y . in section 6 we discuss the m i ~ s u l t s assess their an and significance.

2

~~h sources

A large amount O meteorological and oceanof graphic data was available for analysis and vcrification in these two stoms, inc!uding observations fmm the Canadian and U.S. buoy networks operating off the east coast of North America coastal stations, both manned aad automatic. and observations

202

JOURNAL OF ATMOSPHERIC A N D OCEANIC TECHNOLOGY

VOWME

13

SI

m

from ships of opportunity in the northwest Atlantic surtmcnts provided the vezi6carion for the wave model Ocean. The following paragraphs briefly describe output Environment Canada opcrarcs six 6-m NOn the characteristics and interpretation of these data MAD buoys, offshore and i dap water; the National Oceanic and Amospheric AdminisPation ( N O M ) SOUTCCS. The most important observationsfor this s ~ d were maintains a series of buoys, both along fbt? coast in y the measurements of winds and waves made from the relatively shallow water, and offshon in d w water. The locations of the buoys are shown i Fig. 1; detailed n Canadian and U.S. metcMological and -ogr@k buoys. The buoy wind mcasurunents were the Primary information on the parameters obsmed by the buoys al input t the kinematic analysis, while the wave mea- is given in T b e 1. o

FFBRUARY 1996

C A R D O N E ET A L .

203

.

i .

I
!

I
e,
The m s imp0rtan.t cbaracraistics of the buoy meaot surements in reiation to this study arc
(i) winds fromthe NOAA buoys arc 8.5-minscolor average specds, and dirrctions are unit vector averages: (ii) winds from the Canadian buoys am 10-min YCCfor average speds and directions: (iii) winds f o the NOAA buoys may be at either rm a 5. 10. or 13.8-m level, and wind observations from the Canadian NOMAD buoys arc a 5.4 m; t (iv) Canadian buoys also rcponcdthe higttst 8-s running scalar mean peak wind speedin the IO-min sample;
'

!

( v ) Canadian and U S noniiecrional wave mea.. suremenu uscd a "strapped down" acalcrometcr aligned with the buoy's m c with the wccption of ai s buoy 44139. which employs a *baled Datawell heave sensor: the dirrctional buoys (44014,44025) w a gimbaled Hippy 40 sensor; (vi) Canadian buoys sample waves at 1 Hz for 35 ethe NOAA buoys sample at 256 Hz (DA'X) or . 1 5 HZ (GSBP)for 20 min, and (vii) significant wave height in tenths of m e w and peak period in tenths of seconds computed from the sample is recorded along with the 1 (or2D) spectra D

..

204

JOURNAL OF ATMOSPHERIC A N D OCEANIC TECHNOLOGY

Voruwe 13

...

.

.. . ~.
I

, .

i.

1

I . : .! .

FIG.5. Twelve-hour (a)-(d) surf= wind field d y w r in thc immediate vidnily of thc Hnlloupn Itom Showing cvolunon of major drcukion fa- including jetsbcwcsn Ddoo U T 29 Ocmba Md 1800 UTC M C k r o k .
. .' .

.

.
I

.

.

?be detailed specifications for the NOAA buoy payloads in operation during these events is given in NCDC (1993a) and Axys Environmental Systems Ltd. (1992). For waves. the told measurement system accuracy is usually quoted as 50.2 m or 5% for HS and z 1 s for TP. However, a larger uncertainty results from the effect of sampling variability. which for typical buoy sample lengrhs imp& an uncertainty of 10815% in HS and 5 5 % in TP (Donelan and Pierson 1983). Winds f o shipsof-opponunity wcre a mixture of rm anemometer measuremenu and Beaufort estimates. These were adjusted to provide consistent input to the kinematic analysis as described in section 3. Wave estimates f o ships w e n not used in the study. Winds rm f o manned and automated coastal stations were also rm used i the kinematic analysis, with subjective c o c c n tions to account for laud-based effects (e.g.. increased surface fiction).

to hindcast the maximum sea statcs experienced off the cast coasu of the United States and C n d . After conaaa sideration of the extent of the stonn circulations and the time of o c c m n r r of the peak waves at the buoys, the following wind field atmbutes were adopted:

spatial domain 23'-67W, 30"-82W grid system 0.5' X 0.5" latitude-longitude t m step 3-h analyses (interpolated to 1 h) ie reference 20-m effective neunal wind speed storm history HOS oo00 UTC 25 October- 1800 UTC 1 No her 1991 SOC oo00 UTC 1 M r h 1993-oo00 UTC 17 1 ac March 1993.

a -

In addition to the measurcd dara the following surface analyses w m refemd to
N O M Northern Hemisphere surface analysisf n l analysis ia NOAA N&onal Hluricane Center (NHC) surface analysis METOC Halifax surface analysis.

. .. ~. : . . . ..

. .. ..: . .I .. . . ..".~. .. i . . .

..

,.

3. Wind fields
a Analysis method

- . .~

..

.

The particular wind a~~alysis approach followed. deThe objective of the wind field analysis is basically adn to &be the evolution of the surface wind field over scribed by C r o e et al. (1980). attempts to blend the a domain of spatial and t m o a resolution sufficient advantages of classical manual analysis with modern eprl

..

a

FKBRUARY 1996

C A R D O N E ET A L .

205

objective techniques. The advantages of manual analysis for marine surface weather analyses have been recently emphized by Sanders (1990) and Ucccllini et al. ( 1992). and t m is indted a crcnd at Nwp centers h to maintain and facilitate the use of hand analysis through the use of analyst-fritndly graphical intcxfaces implemented on powerful workstations (e.g., Kocin et al. 1991; Swail 1992). The analysis method consists of the foilowing basic steps.
1) DATA PREFXOCESSING

The first step is data preprocessing to identify and filter from the synoptic data sueam obviously moneous data For buoys, measurements arc made by calibrated insnumenu so errors arise mainly in miscoded or mislocated data and transmission errors. which art easily scmned due to the relatively small volume of data involved in this sNdy by visual scans of time series of observations from each platform. The quality control of ship pressure reports follows the approach of Sanders (1991) to screen mislocated reports, miscoded dam. and biased surface pressure observations. The preprocessing of the data strcam also includes the adjustment of all measurementslobservationsof the marine surface wind to a common averaging interval, reference heighf and to quivalent neuual boundary layer stratification. The intent of the analysis is to resolve the "synoptic scale" wind field at 3-h intmals on a grid of spacing 0.5' in latitude and longitude COVering the analysis domain. To minimize the effect of "mesoscale" variations on the buoy measurements, hourly samples of the 85-min average wind are smoothed by averaging three successive measurements with weights 0.25. 05, and 0 2 . For the buoys that .5 obtained continuous 10-min averages. an hourly average wind is obtained directly by averaging six successive 10-min averages. The averaging is done on meridional and zonal wind components of the wind to derive average wind direction and on the scalar wind speed to derive average wind speed. Canadian buoy vector average wind speeds are first transformed to effective scalar average winds using a wave height dependent factor. Gilhown ( 1987) has suggested that the underestimate of the scalar wind by vector averaging for a sample of data from U.S. buoys is about 7% for wind speeds greater than 8 m s-' ; this mult was derived from analysis of measurements in rather modcrate sea states (HSless than about 5 m). However. the underesrimatc may be greater in higher sea states. and in fact forecasters in Canada tend to regard the Canadian buoy reponed gust value as a better measure of the average wind at standard height in storm conditions. (The buoy dand nansmits the highest running 8-s mean gust.) This impression was also formed by the kinematic analysts in this sNdy, after comparing winds to ship r e p and winds derived f o prcssure gradients using a calibrated boundrm

ary layer m d l A comparison of the dependence of oe. the gust factor (nponed gust spetd/rcponcd average speed) on reponed HS for Canadian and U.S. buoy observations in the HOS (shown in Fig. 2) does suggest a dependence of vector-averaged speed on sea state. which does not appear for scalar-averaged winds. Assuming that the gust factor itself should not depend on sea state (and hence that the scalar-based factor from N O M buoys is correct) provides a linear (in HS) cord o n factor (by n o d z i n g the ratio by its base value) for the vector wind speed average. which varies from496 at HS o f 5 m up to 12% at HS of 15 m. 'Ihe adjusrmmt of measured scalar (or adjusted v s t o r ) average wind spccd to a commoo reference level is carried out following the procedure originally suggested by Canione (1969)(see also Cardone et aL 1990). The p n ~ ccdurr uxs srabity-dqmdmtsrafan wind profile laws and an assumed depndcnce of the roughnss paramncr on wind sp&d to calculate an "effective neunal" wind spccd at a rcfcrcnce height The effmive neunal wind specd u, is the "vhntal" wind spd that at height 4 impam. in neutral t h d surface boundary layer &fication and for the Bssumcd drag law. the same SRS I as impaned by the mcamul. Urn, measuRd at height t in . a thcrmally snatified boundary layer

where L is the Monin-Obukhov length and # is the "profile" stability function [for a more detailed d s c r i p tion of the iterative procedure used to calculate U from . the three hown quantities: measured wind sptcd mcasuremat heighf and air-wa tcmpaaturedifference, x c Cardone (1969) or Ross et a. (1980)l. The calculation i also yields estimates of the f i t o velocity U and rcin . roughness length a.Oceanweather. Inc, ( O W use the effective height of 20 m, so that buoy winds can be comparcd in an analysis to winds reported by ships. Ship reponed winds arc c0nvem.d to 2Gm "Cffdve n e d ' winds for input to thc wave models following the proctdurrs also dacribedby Cardone et E (1990). L whcrc adjustments arc made depending on whether thc wind was estimated or mcasrmd and the height of thc ancmomem where applicable (and horn, from the World Mertorological organization Ls of Supplemenit tary and Selected Ships). If the anemomcm height wm not known, then an esrimate was made based on average heights. Where the wind spd was esdmaed acc(rrding to the Beaufort scale, Cardone et al. (1990) demmined rhc effective neutral s p e d from the rehionship V , = 2.16V:'9, whcrc V, is the reported wind Speea No attempts were madc to othenvisecomctthe winds from thc volunwy obsaving fleet (VOF), such as possible flow distonion effecn. Therc arc two additional subtle and related assump tions implied in the adjustment procedure. The first is that the mughucss parameter does not depend on wave

I

I

206
i

JOURNAL O F ATMOSPHERIC A N D OCEANIC TECHNOLOGY

--am*

y.md W M h m v U 3 3 s

- 1 D I I y N cs

Voruue 13

yE*sum

a

NOV

I

I

!

..

age, but a number of rcecnt studies (e.g., Maat et al. 1991) suggest that it das. To account for this effect would have q u k d making the wave hindcast a pan of thc wind analysis pmess itself. The s s o n d assumption is that normalizing the effect of t h d suatiiidon through tbe fiction velocity U . as is implied in the adopted pmedurr. also accounts for the effectof srabity on mC wave generation proccss itself. However, as suggested by Kahma and Calkocn ( 1992) scaling of wave g ~ ~ wwith fetch in t e r m th

of nondimensional variables scaled with U as o p . p o d to U a p p m to be insufliuenc to fuuy explain . , the effects of thermal stratilkation of the surface boundary layer on wave development with fetch.
2) B~~~~
svwAa

s

An Si l esrimatc of the 3-hody Wind field ta throughout the analysis domain is made using a mariac plancCW boundary layer model (MPBL) applicd t o

."
Feanu~ar1996

C A R D O N E ET AL.

207

FIG. (Continued) 6.

gridpoint-specific gradients of sea level prrssurr, air e'm and air-sea temperature difference. Since its development (Cardone 1969). this particular model has been used a many national centers for routine analt ysis and f-ting and in many studies (see. c.g., Overland and Gcmmill1978; Cardone 1991). For this application, the required MPBL variables were derived from carefully hand-analyzed and digitized maps drawn t carefully maintain continuity of centers of o action and fronts. A Large n u m b of evaluations of surface wind fields produced i this manner (Cardone n

1991) indicate that a scatter of about 3 m S - I in wind speed and 30" in wind direction is about the intrinsic limit of skill in specification of synoptic-wale surface wind fields over the ocean using simple analytical steady-state MPBL models. To reduce t i error in a hs surface wind analysis scheme requires the assimilarion of m d surface wind data

3) KnuhcA.ncmwLusrs Kinematic analysis (KA)is carried out only for the most critical parts of the wind field. 'Ibis criteria gen-

Et - ..

-.

it.

.. .

208

JOURNAL OF ATMOSPHERIC A N D OCEANIC TECHNOLOGY

Vowme 13

5' 0

45'

Qo'

35'

30'
I1

25'

0

. . . . . . erally excludes the model spinup period (generally the first 36-72 h of the modeled period), areas well Outsi& the circulations of the major cyclonic centers of action and also over the open North Atlantic Ocean well cast of the buoy array. whrn the sparsity of ship dam tend to minimize any incremental skiU to be gained from KA. In KA, streamlines and isotachs of the effective neutral 20-m wind arc manually drawn at 3-h intervals. subsequently hand-digitizcd at 0.5" latitudelongitude intervals, and where available they simply supersede the MPBL gridpoint winds. The KA winds blend smoothly with the MPBL on the borders of the KA region because they arc drawn to do so by the KA analysr

.. .. . .

The evolution of the main meteorological systems comprising the HOS event have been described in deti by Cameron and P a r b ( 1992) using mainly o p al erational analyses produced at AES, and by NOAA ( 1992). using mainly NOAA National Meteorological Center (NMC) products. The emphasis h m is on the surface wind field evolution associated with the formation movement and interaction of the two cyclones comprising this event. one subtropical and the second extratropical. Figure 3 shows the mcks of these systems over the period of intercsr Figure 4 is a composite of the surface wind fields over the subdomain of the analysis grid shown, at 24-h intervals over the peri oo00 UTC 25 ~ b e r - o o o OUTC 1 November 1 : The conventional wind barbs rcprtscnting the 1-h a erage 20-m wind an ploncd at 1" spacing. b. Evolutioh o the windfield f The southern subtropical system formed late on the 25th ahout 300 nm south of Bcnnuda, intensified to 1) HALLOWEENSTORM tropical storm sucngfh early on the 26th and hurricane The HOS has been studied h m t o f o n mainly in smngth (Grace)late on the 26th. Grace moved slowly terms of the devastating impact of its storm surge and northwestward on the 27th. The surface wind field heavy swell on the coastal communities from Maine about Gracc during this period was charactcrizcd by (whrn, e.g., U S Prcsident Bush's Kennebunkpon winds of gale to storm force over a large area, up to .. home sustained sevcrt damage from wave action) to 300 nm from the center. Grace slowed and avned as far south as the us.v i Islands. The storm dam- northward on the 28th while weakening. then tumed aged or desmyed over a thousand beach front homes. toward the cast and moved for a time eastward toward caused 12 dcaths (including five in a fishing boat loss Bermuda early on the 29th. While the storm was losing offshore) and iaSurrd lOsscS the U d d States 4 its tropical charactcristics and its inner core of near hurmated a between $15 and $2 b l i n (NOAA 1992). ricane force winds (not well resolved on the 05' grid) t ilo

QI)

FEBRUARY 1996

C A R D O N E ET AL.

209

t it still a this time possessed a large circulation with winds in excess of 20 m s-' covering the large area i. shown in F g 4. Late on the 29th. the remnants of Grace m e d northeastward. sparing Bamuda si@cant imoact and on 30 October became absorbed into the norrhcrn cyclonic system. The nonhem system formed out of a weak wave thac had been moving eastward out of New England along a strong basically cast-west-oriented cold front laid OUI along the St Lawrence Valley early on the 27th. As the wave reached the eastern tip of Nova Scotia at abouc 1200 UTC on the 28th. it evidently tumed abruptly southward and began to intensify rapidly. Twenty-four hours later, the center was near 41W. 55%'. its cennal prcssurc having fallen about 24 mb. The cyclone then tumed toward the west-southwest and continued to intensify over the next 24-h during the 30th. before slowing and filling on the 31st The center executed a counterclockrvise loop on the 31 October and 1 November while 6lling:Beyond 1 November. the cyclone cumcd . northward and for a time intensified and acquired to r p ical characteristics before weakening and entering Nova Scotia just west of Halifax on 2 November. The surface wind field about the northem system intensified rapidly late on the 28th. first in the nonhern and westem quadrants of the circulanon. By early on the 29th, surface winds of greater than 25 m s-' developed from the Grand Banks of Newfoundland southwestward. and a continuous northerly airsof surface winds of 20 m s-' or greater extended over the western pans of the two cyclones. East of the two cyclo.ncs, the circulations remained d s i c a this time. itnt t The derails of the interaction of Grafc and the expatropical cyclone (EC) on the 30th may never be completely resolvable since there was no aimaft m o n naissance of Grace aftcr this time. The N O M National Hurricane Center (NHC) analyses carried a distinct circulation of Grace until 1800 UTC 29 October after which NHC considered its circulation absorbed into the northern system. O r kinematic analysis supports the u s w i v a l of a distinct circulation c a t e r of the remnanu of Grace until at least o600 UTC 30 October. Regardless of the exact evolution of the center,however, there appears little doubt that the tropical system w s rea sponsible for the northward migradon of a band of gale force southerlies cast of the center of the southern system and toward the fmnral boundary associated with the northern system. By oo00 UTC 31 October. the wind field hsr simplified into the pancrn of an extensive circulation about a single center (hence the "Halloween storm") located just south of Georgcs Bank The strongest surface winds in the HOS, up to 33 ms-I. occumd between oo00 and 1200 UTC, and were found nonh of the northern system, in the area of the most intense prrssurc gradients. 4 t 1200 UTC 30 October there was a total pressure drop of 60 hPa between the 972-hPa low near 40"N. 62%' and the 1042hPa anticyclone centered near 50% 65%', with over

two-thirds O this gradient concenaatcd between the f low center and the southwest tip of Nova Scotia The availability of the Canadian buoy winds and ship rcpom allowed the K A to detect fine structure usually not resolved in the evolution of the surface winds in the airstream nonh of the cyclone prior to and during the occurrence of the maximum sea states in the Canadian buoy array. This fine structure is shown in the composite of KA covering the period o600 UTC 29 October and 1800 UTC 30 October given in Fig. 5. (The 3-h KA are sampled at 12-h intervals in this figure.) The broad a m of greater than 50 kt wind speeds north of the storm center during this period is actually a combination of two distinct wind maxima, which could be tncked southwestward during this period. The first maximum seemed to originate near 45W, 5 1W at o600 UTC 29 October and to propagate west-southwestward at about 20 m s - I , as estimated from its movement over a 12-h period. The core of this feature passed almost directly over 44139 inducing the indicated speed maximum of 29 m s-' (at 20-m height) at 1200 UTC as s e n in the time history of measured winds at that buoy in Fig. 6a This feature, which we shall call a surface wind "jet sIrcak" or HOS-JSl, lost definition aftcr 1800 UTC 29 October and was overraken by a second JS. HOS-JS2. which developed ups near 45W, 52'w at 1500 UTC 29 October. HOS-JS2 could be u c west-southwestward a little aM south of the track of HOS-JS1. the con speed maximuin of 33 m s-l passingjust nonh of 44141 between oo00 and 0300 UTC 30 October and passing direftly over 44137 at 0300 UTC 30 October, as reflected in the measured time histones in Figs. 6b and 6c.The time history at 44137 also shows a secondary spxd maximum at o900 UTC 29 October. which is the effect of the COR of HOS-JSl passing nonh of this site earlier. HOS-JS2 maintained its identity over a longer period than HOS-JSl. roughly the 48-h period 1500 UTC 29 October-1200 UTC 31 October. This JS propagated down the flow at an average speed of about 17 m s-l as it approached and passed 44137. This speed is about 50% faster but basically in the same &tion as the movement of the parent low center during this period. By 1200 UTC 3 1 October. HOS-JSZ seemed to become associated with the parent low and propagate slowly westward wt it thereafter as the maximum core wind ih speeds gradually diminished. Figure 6d shows the arrival of the core of this JS over 4401 1, now associated with northerly rather than east-northeasterly flow late on 30 October. These interesting finescale feacurcs of the wind field seem to be intimately related to the pat. tclll of exmme wave generation observed in the Canadian buoy array and will be discussed in the next

&on.
2) STORM OFTHE CE?iCURY The SOC is one of the greatest weather catastrophes ever to affect the United States. The death tl of 270 ol

!

210

JOURNAL O F ATMOSPHERIC A N D OCEANIC TECHNOLOGY

VOLUME I3

J

cxcccdcd the combined dcarh toll of Hurricanes And n w and Hugo by thrtc timcs. Insund 1Os.s~~ esrimaced at $6 billion. while the impact of SOC is USUally expressed in terms of its associated biizzard conditions and rccord snowfalls from Alabama t Maine, o the extnme surface winds and low temperanvs and the property loss on land (e.&, NCDC 1993b). the impact at sea was also grcaL At least 48 lives were lost

at s a including all hands on the bulk Carrier Gold Bond Conveyor south of Halifax on 15 March.

TheSOCwasforrseenawcckinadvanccbyNOAA
medium-range numerical weather pnsdiction (W) models and very well forecast by short-term I ” models. Figure 7 shows the hack of the surf= low. The surface development began over the western GuLf of Mexico carly on 12 March 1993, and intensification

.
0
!

FIG. (Coruinurd) 8.

was explosive h n the beginning. 'Ihc c e n d
dua-cadh loo0 to 969 Wain mC Whperiod during

which h e storm aossed the nonhcm G l of Mexico and uf entered Georgia Even in the Gulf of M x c the storm eio &vdopd its "lO&ytar" signature. ' E storm gencrared ? wind gusu t 98 mph on South Tunbalier platform south o of Lnuisia~.The maximum HS at N O M buoy 42003 (25.9w. 85.9w) was mcaslllcd at 92 IIL Figure 8 shows the evolution of the wiid field at 12-h intervals between oo00 UTC 13 March and 1200 UTC 15 March. The full cyclonic circulation is not shown on these plou hecause winds over land are not plotted, and the center remained on the coastal plain as it moved rapidly northeastward. Al of the extreme l wave generation occurred in the basically southerly (southeast to southwest) wind field in the Es am Hemisphm of the circulation. These strong winds dthe Florida Peninsula and emerged off the southeast coast (Figs. 8a and 8b) carly on the 13th. By oo00 UTC 14 March. surface winds (1-h average at 20 m) anained maximum spcrds of 30 m s - ' cast of the Carolinas. These winds reached the Scotian Shelf about 24 h later. The lowest cenhal pnssurr recorded on land was 958.4 mb at White P a n , New York, though a lis ship docked in Baltimore Harbor recorded 956 mb. The ccnhal prcssun increased and the maximum winds gradually weakened as the storm moved through the' Canadian Maritimes on the 15th. At least over the msin part of the inlcnse circulation that passed over the buoy array off the East Coass the K A allowed the resolution of several 6necale s t w N m including two Well-defined J . The more merS
.#

getic J of the two, termed here SOC-JS 1. emerged off S the northeast coast of Florida late on the 13th. and accelerated northeastward during the 14th and 15rh, as shown in the composite of KA (sampled at 6-h intervals) in Fig. 9. The movements of thee feanrrcs domi a e the time histories of the measured winds at the nt buoys in their paths. as shown in Figs. loa-d. which compare modeled and measured surface winds (again the measured winds are s o t e from three COIISCCmohd utive hourly mcasurrments, adjusted to 20-m neutral stability) at buoys 41006. 41002, 44004. and 44137. The core of SOC-JSl with maximum winds of 30 m s-' passed just n o d of buoy 41006 inducing a maximum wind sped of 25 m s-' there at 1800 UTC 13 March, thenpassedapparentlyrightover41002at2100 UTC 14 March where maximum winds spceds w m 30 m s-', just east of44004 at o600 UTC 1 4 M m h (maximum wind spnds of 25 m s - ' ) . and just west of 44137 at around 2100 UTC 14 March (maximum wind specds of 24 m s-' 1. The propagation velocity of the core of JSI averaged about 13 m s-' off the east coast of the United Starcs accelerating to 17 m s-' over Canadian waters. A second JS. SOC-JS2, is resolved but not as accurately as SOC-JS 1 because it evolved in part over sparse data BICBS just cast of the buoy array. The surface flow was basically southeasterly in this JS, but the JS itself propagated north-northeastward The core of SOC-JSZ ( i h maximum winds about 28 m s-') wt passcd east of 44004 at 2100 UTC 13 March and east of 44137 at 0600 UTC 14 March where the secondary local wind speed maxima are induced as scen as indicated in Figs. 1Oc and 1Od.

.

0

.-

--

. -

.

.

-

VOLUME 13

212

J O U R N A L OF A T M O S P H E R I C A N D O C E A N I C T E C H N O L O G Y

-

.... . .. ... .- -. ,. . . -

.

.

FEBRUARY 19%

C A R D O N E ET AL.

213

4. Wavehindcssts
a ModcLc

All of the wave models are of the discrete spectral
type. This general type of wave model has gradually replaced earlier “significant wave” type models at most (but not yet all) major analysis-forecast centers and has been used in most wave climate assessment

This second-generation model, also known as WAVAD. evolved from an earlier 2G model first described by Resio (1981 ). ’The source tcrms currently used are demibed in Resio and Perrie (1989). In WAVAD. Si. is a tuned quadratic function of dimensionless peak fxewhere E ( f . 8 I. f ) is the energy density at position x quency ( Uf,/g). Nonlinear source term S is modeled . , and time f and is a function of frequency f and prop- in terms of its effective energy nansfer onto the forward agation dmction 8; C, is the group velocity vector; S. face of the spectrum. The form for this nansfer was ; is the source term for energy input to the spectrum from determined from calculations of the total Eux of energy the atmosphere; S is the source term representing dis- to the forward face, which is expressed in terms of the , sipahve processes; and S represents nonlinear energy migration of the peak spctral period toward lower fxe, transfen within the spectrum. The numerical solution quencics; S is based upon the assumption that wave , of this equation is carried out in separate propagation breaking removes all energy that is input to or transand source term integration steps. The basic attributes f a d into frequencies above some threshold fxe(spectral banding, grid systems, propagation scheme, .quency. The propagation of spectral component energy etc.) of the four models are summaxized in Table 2. is by Iinite differences using a fint-order upsacam differencing scheme. Model evaluation can be found, for example. in Hubem et al. (1993). This model is cur1) OWIlG rently used by the US.Army Corps of Engiuecrs in its The source-term representation of this model, also Wave InformationStudy ( W E ) and is used by several known as ODGP. was developed originally by Pier- private companies worldwide. son et al. (1966, hereafter FTB) and was later modified and retuned to provide accurate specification of 3) wAM4 directional wave spectra in tropical cyclones (CarThis is the fourtb cycle of the third-generation model done et al. 1976). The ODGP variant of PTB was later found to work equally well in a wider range of (WAMDI) originally described by WAMDI Group meteorological regimes (Reece and Cardone 1982). (1988). In WAM each source term is specified explicIn ODGP. separate source terms for S,”, Sa,, and S., itly and the source-term balance is integrated to yield are not resolved but are replaced by an algorithm that the net development of the spccuum over a time step leads to at first linear and then exponential growth of of integration without arbitrarily forcing of specmi spectral energy. The growth algorithm is designed to shape or specification of an external limit to growth. allow the spectrum to gradually approach the Pier- However, in the original development of the model. son-Moskowitz (PM) fully developed spectrum at considerable experimentation with and N I h g of the long fetch or duration under the action of a steady input and dissipation source terms was carried out (KOwind. The spectral growth algorithm was later mod- men et al. 1984) m achieve growth rates and asymp ified (ODGP2, see Khandekar et al. 1994) when an totic behavior under constant winds in w m e n t with aa “equilibrium-range relaxation” was added follow- field d t . The present version uses the input source ing the model of Resio (1981) to allow the high- terms described by Jansscn (1991). The most imporfrequency tail of the spectrum to adjust to the stage ’ tant difference between basic WAhtDI and WAM4 is of wave development ( a consequence in nature of the that the growth rate is computed not only in t e r n of balance between three source terms in the high-fre- the wind alone (as in all of the other models applied quency tail of the spectrum). The propagation hen) but is also is allowed to v q x a function of the scheme (Greenwood et al. 1985) is a downstream stage of wave development The input Si. is basically . . interpolatory scheme that is rigorously energy con- quadratic in the friction velocity U and since U is.a

studies carried out over the past IWO decades. In a discretized spectral model. the dirccuonal spahum is resolved at each model grid point in terms of a number of hquency-dircction bands and the evolution of the sea state is found by the numerical solution of the spectral energy balance equation. which in deep water with small ambient currents can be written

serving with p a t circle effects included. The ODGP2 growth and propagation schemes am the basis of the global wave model (GSOWM) operated by the U.S. Navy Fleet Numerical Meteorology and Oceanography Center (FNMOC) from 1985 until mid-1994 (Clancy and Sadler 1992). when WAM was implemented for operational dissemination, aild of models used in proprietary and public domain wave Climate assessment studies for numerous basins worldwide (e.g., Cardone and Ewans 1992; Swail et al. 1992; Eid et al. 1992). 2) Resio2G

214

S O U R N A L OF A T M O S P H E R I C A N D O C E A N I C T E C H N O L O G Y

VoutMe 13

t

function of roughness parameter, which is itself a function of sea state, the wave and wind fields arc wcakly coupled. Also, whercas in WAMDI the whitecapping dissipation source tcrm (Hasselmann 1974) is quadratic in frequency. in WAM4 it is pmportional to the founh power of frequency. H m S., is a parametcrization of the exact nonlinear interactions as pmpossd by H ~ e l m a a n(1985). This sc-callcd diSCnte intenr~tion approximation (DIA) i described in W W Z G s (1988). Propagation is by hrst-ordcr finite differencing.

4) O W G

This model consists of a source t c ~ m representation and integration scheme based upon WAMDI combiucd with the propagation scheme used in OWIIG. The s o m terms follow the theoretical forms uscd in WAMDI but with different numerics and code and with the following modifications. Fink a linear excitation source term is added to S,. taken a~ a d o d e d varjant of the (urn used in ODGP. This allows the s a to e grow from a flat calm initial condition in O W G . un-

Fesnu~ar1996

215

i
1

0

FIG.IO. (CoruuuUd)

Wrc all cycles of WAM. which quire an artificial quency is cubic rather than quadratic. Fiially. for S , . warm s m f o a +bed rm initial spccmun. The ex- the DIA is adopted except that two modes of interaction ponential input tcrm is the empirical form of Snyder et arc included (in WAMDI the second mode is ignored). al. (1981) in which S . takcn as a linear function of The modifications just described wcrc adopted and re8 is U However, unlike WAMDI in which U is com- fined based upon tuning ~ l l against thc fetch-limited . . s puted f o Ulofollowing the drag law of Wu ( 1982). growth benchmark for 20 m s-’ wind spccds under rm
in OWI3G. a different drag law is used in the model hlning stage. TEai drsg law follcws Wu closely up to about 20 m s - I and then becomes asymptotic to a consfant at hurricane wind speeds. Dissipation , is also S taken f o WAMDI except that the dependence on fnrm comtaut win& used to tune WAMDI initially, as well as in rcwated trial hindcasts of a well-documented moderak cxaatropical cyclone (SWADE IOP 1) and two intense Gulf of Mexico hurricanes (Camille 196% Frcderic 1979).

.. 'I

I

-.
216

!

J O U R N A L OF ATMOSPHERIC A N D OCEANIC TECHNOLOGY
TmLe 2 Wave model mbulp.
sparmducrauvlon
Gnd S
F r n

VOLUME

13

Tm ucp i
(11

0

WAM4
OWUG

26
2)

Aflf 0.1
Aflf
0.1

-

24 14

I S I S

C 23'-6"i

0.5.

wB w - z
C: 'IC-67.N

r

G u n k u d. (15911

Fini-=rdn
f i n k diR. Encorder

6m L

6Lm sa,

Kkmmszkarn

5w

C: COMC nest; F: frne n a : No.: number of bands.

0

b. Model grids and outpur
A basic grid conspaint was that at least within the general domain of the buoy array (say 30"-45W.50"8 0 W . see Fig. 1) all mcdcls use a grid spacing 0.5" in latitude and longitude. ensured that resolution effects would not contribute to differenriol behavior between the different wave models in the energetic parts of these s t o m and that the wave model grid point selected to represent a buoy would be the same for all models. As shown in Table 2 most m d l adapted a oes two-nest grid layout c satisfy this consmint. o Since only deep-water physics was considered in the propagation and soum terms of the model used, model results are compared to measurements from buoys moored in deep wafer, or basically those shown in Fig. I. These include buoys 44008 and 4401 1. which are in marginally shallow water. Comparisons were made at these buoys in HOS only because of the high sea states observed there, but separate comparison statistics were computed for smctly deep water buoys as well. For each run, hourly timC histories of the directional specm w m saved at the closest agreed model grid point to each buoy. These were ruluced to the following integrated propems of the spectrum:

16"1
VMD = m-'

sinOE(f. O)dfdO

[coseE(f. e w e

.

(6)

In Eq. ( 5 ) , f , is derived from a parabolic fit to the equation dE(f )/df = 0. VMD is the vector mean dtcction.
5. Evaluation of alternative hindcasts
An objective evaluation of the accuracy of the common wind field is not possible in terms of measured winds because every wind observation and measurement had to be used in the wind field analysis process. Figures 6 and 10 contrast the difference ' skill at the buoy locations at least, between fields computed from the pressure field and stability. using a boundary layer.mode1 (as shown by the comparisons for times outside the KA periods) and wind fields derived by KA. which relies on observations. However. as shown in the comparative analysis of alternative wave hindcasts with WAM4 driven by six different wind fields in the study of SWADE IOP I (Cardone et al. 1994). the KA wind field used is probably the most accurate wind field derivable from the available database. We caution, however, that since the K A weights the buoy in situ wind measurements so heavily, there remains the possibility that the wind speed fields may still be biased at the higher sea states even after our comctions for the height. stability. and averaging method if the ability

w e

iw)= p w-, e ' w
0

(3)
103

r ea. c

(4)

0

i .-A

218

JOURNAL OF ATMOSPHERIC A N D OCEANlC TECHNOLOGY

VOLUME 13

NO.

Avg
msy

Avg

Ma en

E OrnlG

pu
108

hind

diff

-m d i n Sedde*
0.60
139 OS9 032 1.01 0.95 OM 1-71 0.81

NO.

Arg
my

Awl

Mun

my

SI

R

CC

pu

hind

din

diR

Sddm

SI

R

CC

ami
4lmZ

5.74

98
108

5.11
5.19 625
674

44ma
UOI I

I M
I08
108

44137 44138 44139 44140
44141 All

Rpi07.G

-

I08 I08

4.99 531
413

I08
954 746

5.93 531 5.45
S.14

6.05 537 621 6-54 S.98 5.s 5.69 4.95 5.05 5.68 5.48

4 3 1

+ n 1.04 o
-1.07.
-031 0.16 1.42 - 3 0.81 0 -032 0.69

0.09 0.20

0.18

0.85

026 0.63
0.92
099 099 0.96

0.12 MI1 0.26 0.18 0.69 0.16 MI

ala

108 98 103 105
108

1325 I413 11.65 i2.m

1338 13.40

I202
I139 10.99

-0.13 od3 -037 1.11

1.63 1.22

1.62 0.97

0.12 0.07

OUI 0.82
050

( I I LL

094
0.61

2.57
1.90 1.93 1 8

2.9

02 .2

M 2
054 1.16 I.04 I.w
0.69 I.02
093

-0.72
069 -0.17 -OM

o m
1.44

an

0.28
a75 034 0.61

a8 9

0.u 0.15

097

o m
0.19 0.19

108 108 108

12.48 10.91

iim
1132 1213 11 21

10.53 1021

1.49 039

I54 0.12 12 aro 1.78 . 0.16
119 0.14

0.80 0.84 0.91 a 9 3
0.58 W OS9
011 0.87 038 0.90

MI

1.70

o s
0.94

1. in 11.06
10.13
11.46

an
1.19 0.61
0.75

108
9% 746

1.W
1.04

1 m I1 d ldl

1.75 I3

0.94

1136

L74
LW 1.76

m

1.14

0.15 0.12 a14

053

au

0.68 063 0 3 1 an

41001
41002

10s
98
103

J.S
4.96

4.1s
0.15
-0yI

0.70

a n ab(. 0.73
020
0.11

108
98
103

1325
1423 11.65

44ma
UOll

5.11 5.19

1.03
1.M 0.78 1.41

039

I M
108 I08
I08 108 108

bu
674 4.99 537
4.23

443 417
44138

5.61 6.60 6.12 4.86

ala o s

0.61 0.90

13.29 1322
10.46 11.43 12%

-om
1.01 1.19

zm
1.45

0.1s 0.10

OSI 0.76
021 025 0.21

0.87
033

-035

0.70

06 .2
a3 1
-1-23 -0.06

I34
1.02 1.43 0.60 I20 Ill 12 .9

1.03
1.89

44139
44140
44141

660
430 4.11

o m o n
0.14

o m

0.72

043

0.97 0.96

I M
108
108
108 108

izm

i n

1 s
1.93 261 1.84 3.64
214

1 s 1.44
261 1.83

a3 1
0.11

M I
1.71 IO 129

9 031 a 4 a93 0.0 O S 0.95
0.18 o s
lU3 OM
0.91

All
Dap

9 %
146

5.93 S ~ I
5.45

s~

1.22

-a01
0.10
416

o m
0.22
024

108
954 746

535

0.90

1248 10.91 11.02 11.77 1132 12.13 lZll

a14

021

o . u
0.48
0.66

0.

10.72 a9 1 1281 -1.115 10s 1.18 9.n I35 11.65 DT Ll ll.81 OJO
13.01 13.W
024 1.14

202
ZT)

237
26 0

0.17 3.14 018 1.79 0.15 I . 0.12 M 223 0.18 2.35 0.19
25 0

:# :

0.m
a74 0.88

03)

om
.pu

om .

038 0.10

OWUG 41mi 41002

108

5.74

5.90
5.14

5 a
103 105

44ma
UO1I 44137 44138 44139
44140 44141 All Dap

108
108 108

5.11 5.19 625 6.74
4.99

-om
-0.86 -0.11 0.92 0.26 0.03
116 0.21
0.41

6.05
636

5.82
4.73

049 06 7 . 129 0.98 IY

OJt
a76

a n 03 033 o ais OM o m
0.19 0.16 a5 1

108
98 103 105 108 io8 108 108 108 954 746 108 98 103

13.25
14-23

09 .6
0.98

0.13
OS3

0%

0.97

ass
odo
0.m 0.61
0.94

0.86

0.98

537
423

534
3.72 4.68

odl 0.m
137

a16 0.76
0.150.75asS 0.14 0.81 097 A16 0.94 0.98 ais 0.63 0% 0.11 0.29 a%
0.11

108 108 954
146

032 0.80
Lo4
1.01

11.65 iim izm11.70 1248 11.10 10.91 10.19 11.02 954 11.71 i o n
1132 It13 12.11 9.63

-om

1.m
249 201

1.13 248
1.75 1.19 205 1.71
1.66
I54

1m .
138

0.15 033 0.78 0.080.a.i 0.93 011 0.43 053
0.14

i.m

0.72 1.48

217 226

0.10 0.19 0.15
a14 0.14 0.15 0.14

om om

OdO
0.94 0.80 0.90

038 0.86
0.74

5.93
531 5.45

120 ZM 1.69 229

an

s3 .4
5.01
5.93

1.02
092

11.16
10.99

09 .7
1.12

210
205

1.87
1.77.

0 9 0.93 .4 0n 0d3 . 0 2 aK? .3
033 0.83
0.43

wAhl4
41001 41002
l(m8

108 98
103

5.7i

5.11
5.19 625 6.74
4.99

5.m
6.08 627

-0.19 0.07 -0.89
4 0 1

0.0.78
1%

UOI I 44137 44138

lM
108 I08

5n .
4.66 536

44139
44140 44141 AI1

J3l

0.97 033 -0.19

I30 1.41 0.98 I18

0.61 0.78 126 I29 I.02 0.93 1.26

0 3

0.90

0.15
0.14

0.9
0.26 0.67 0.87
0.74

as0
0%

13.25 14.3 11.65

13.33

-0.08

23 .2
1.47 2.80

u z
I33

1338

0.85

im

021 0.15 0.19
0.24

am
0.83 0.94

0.97 0.98 0.97 0.97

10s
'

M
954
746
S3I 5.45

5.28

rep

5.m

0.23 0.42

1.21

I a6 1.19

0.13 0.18

097
0.97 0.93 0.94

1.15

1.m

0.22 0.66 Q.33 a 9 .. 2 .

108 108 108 108 108 954 746

1270 12.48
10.91 11.02 11.71 1132 12.13 12.11

12.48 -0.83 1221 019 1 s 0.95 1 10.46 0.4s 9.91 1.12 10.81 0.96 9.96 136 1154 039 I I J I 0.m

267 090 210 1.67 1.73 LEI 1.90 1.73

0.17 0.08 0.21

0.73 0.92 039

1.60 131
215 201 1.98

0.12 a70 o m 0.87 0.19 05 .0
0.15

0 .

1.88
1.99 1.91

0.15 ai1
0.16
0.14

0.81 0.61 0.90 0.69 a28

091 0.73 0.95 0.82 0.86

B
o

SIonly.

index; R-rad0 of pinu nbovc and below 45. Line; CC-mlation

cocffiiCicns Dsep-staIirtics

for d deep W a m buoys l

bias (hindcast-mcasurrd)
1 b ="

2(h, - rn, 1;
i

(9)

..

FEBRUARY 1996

CARDONE ET AL.
b Y r

219
i dr H . u O w a u SDXm for four wave &Is n
OWDG
wAM4

T~8Le4. SImm palr maamd md hiDdcpa wave heinht md Prid

0
Buoy

Meantrcd

OWllG

Rnio2G

HS
7.7 7.8
9.5 11.8

TP
20.0 20.0 12.6 17.8 17.5 13.7 14.2
I23

DWHH

HS

TP
18.7 13.7 15.7 16.1 16.7 14.2
14.0

DD/HH

HS
7.8 5.9 9.8 12.9
153
111 123

TP
12.5
143

DD/HH

HS
7.8 6.6 11.3
128

TP
18.2 163 15.6 16.1 16.8
14.4 14.4 12.5 15.3

DDIHH

HS
8.1 6.7 11.8 13.7 16.2 12.7 12.8 73

TP

DDlWH
31/07. 3oR2 311w 3W15 3M)6 29/20 29/22 30/01 30/01

41WI 41002 44008
44011

44137 44138 44139
44140

16.9
11.5

IO3
8.1

44141

14.6 .l6.0

6.8 31/05' 31/10 6.2 3 0 ~ 105 3W16 12.0 3W5 '133 29/16 11.3 29/16 ,10.6 2 9 N .. 8.8 '10.3 3M)I

31/01
30/18

132 14.1

31ms MllS 30/06 29/20 29RO 30102 301M

8.6
10.1

29/07 29/19 IU 31/01 14.3 W19 16.7 3010) 14.3 3woO 14.3 29/21 14.3 UMI 14.3 UM5

14.7 11.7 11.7
'

7.0 11.5

31/M 3MI 31m1 3W16 3M)6 29RO 29/20 3OlU2 30102

18.2 16.9 16.5 16.9 17.0
14.4

15.3 129 1 1 3 ~ 15.6

HS-sipificant wave height ( ) TP-peak pi (I); DDMi-dayhur of muimvm HS. m: ad Meamred wave height and penk paiod arc smoothed using a 1.1.1 d n g average.

standard deviation of differcnec
. L. ~ " r ml
v
,L

L

- " .. - 1 m

L,2 01-

(

)

scatter index . ,.

SI=-; cd

(12)

ratio (where 6 = 1 ifhi Z mi. 6 = 0 if h, < m i )
(13) correlation coefficient

cc = [X ( m i - ni)2/noJl[X (h, h)2/nIoJ '
i
i

7 [(h,- x ) ( m , - K)]/n

-

( 14)

Time histarin Of Hs and l T buoys sites evaluated.
a Halloween sfom

PIancd at

0

A general impression of the behavior and performance of the alternative wave models is gained through

Fig. 11. which gives scatterplots for each model bemeen hindcast and measured time histories of HS for a l l nine buoys, and Table 3. which gives the comparative sratistics of HS and TP at all buoys and aggxegate statistics over all nine buoys combined and a ; & 6nly the Seven buoys, which arc moored in deep water ( i s . , excluding 44008 and 44011) The first 72 h of the hindcast period, which was reserved for model spinup, arc omitted from these comparisons. We have indicated in boldface in Table 3 for each buovlnatistic thc model with the mast favorabie behaior. Fir. it should be nofed that all of these wave modcls provide exucmely skillful hindcam in the aggrcgatc. For HS. the mean difference varies from between -0.01 m for Resio2G to 0.23 m for WAM4. H r SI varies from only 0.17 for O W G to 0.24 for ee Rcsi02G. The correlation cocfficicnts'vary over the m w m e 0.91-0.94. No mcdcl clearly oucperg forms any other in specification of time histories of HS in this storm. Differences between the models in specification of TP arc slightly more apparent. Both 3G models show about a 1-s negative bias in Tp, common a properry of previous validations of WAM (e.g., Cardone et al. 1994). Resio2G is the least biased in l T albeit af the expense of its having the greatest sca[trr with SI of 0.19 relative to the OWIlG modcl, which exhibits the smallest SI in TP of 0.13.

.

Mods1
OWlG Rerio2G
OWUG

Avg meas

Avg

Mcan
diA

mu

Avg
Stddrv

Avg

MWI

mu

hind
10.0

diff
201 1.71

S I
0.16
0.15

R,

CC
0.81 0.84

mep(

hiod
15.1

diff
0.87 1.85 0.46

diff
258 3.49 1.83
1.87

Stddev

S I
0.15 0.18 0.11

R

CC
032 0.09

WAM4

10.9 10.9 10.9 10.9

IO3 10.6
112

0.94 0.41
035 -028

139
1.73

1.78 1.66 155

1.71

0.14 0.16

036 036 0.44 0.44

0.M 0.84

16.0 16.0 16.0 16.0

243
2.96 1.77

0.67
033

141 15.6 16.0

056
O S

0.05

1.87

0.12

0.83 0.79

SI-

indrr:R-Rfio

of poma above ud W o w 4 P k CC-con'ch 'on socffidmr

220

JOURNAL OF ATMOSPHERIC AND OCEANIC TECHNOLOGY

VoLuue 13

.
f

W I G

llELLNRED

1

FG. I Comparison of b i n h r and m e a r d dmc hisoria wave height and wid 2
i Ihc H a l l o m s n
m al six wlcnsd buoys for four m&b.

The scatterplots show in addition that below HS of about 6 m there is relarively lit~le difference between the model hindcasts of HS except for slight level shifts, which pmbably arise from small differences in model cuning. I the range of HS between 6 and 12 m the n

s c a m incrcascs for d models but the hindcasts arc l basically unbiased. Above 12 Q w h m most of the points are contributed by buoy 44137, tm i some h s stratification in pcrformancc by model wirh WAM4 closest to the 45" line though still slightly negatively

'1
. .
I

FWRUARY 1996

C A R D O N E ET A L

22 1

1
...

IS

30

31

1

-

1

nOV

FIG. I (Continued) 2

biased at the highest sea sfales. OWIlG is the most negatively biased at ESSs. while OWUG and ResioZG behave vcry much like WAh44. Table 3 also shows that then arc some variations in model sM1 rank from site to sire that may be illustrated through comparisons of time histones of measuredhindcast time histories of HS and T' at individual F

t buoys as follows (sec Fig. 12). At 44140. which is a the extreme upwind end of the principal nonheasterly fetch of the s t o m there is about a 2-m spread in the time histories of HS (OWIlG highest, OWUG lowa t ) , with this envelopc generally straddling the measurcd time history during the 48-h period centered on F the peak The measured and hindcast T' histories an

I

222

IOURNAL OF ATMOSPHERIC A N D OCEANIC TECHNOLOGY

VOLUME

13

all in close agreement during this period as well. At 44138. the models lag the p w t h stage and lead thc decay stage in the local rime history of HS by 3-6 h. with very good apeerncut between hindcast and modcled storm peak HS and associated Tp for all models. At 44141. the southernmost of the buoys, all models underspecify the part of the storm during which HS

exceeds about 8 m. In this regard, this time history stands out from all of the othm. This buoy is in the vicinity of the north wall of the Gulf Strtam during

HOS and perhaps the deficiency in the hindcasa h m arises i interaction of the growlh with cumncs in thc n
sueam or a detached eddy. At 44137. one of two buoys

that experienced ESS's in HOS, Rcsio2G lags in

. .

. -... . . ._. . . .

1
I

FEBRUARY 19%

C A R D O N E ET AL.

223

'0

20

1

/

5 E
0
1

wbhlcyaer

0

5

10

15

20

HS mcdel (m)
FIG.13. Average of measured and model hindsast uorm pa HS i HOS and S O C o v a p u p 3 of sites -"ring n mearurrd HS psolu above and below 12 a

In model applications to assess extreme wave climate it is the specification of the peak sea state at a site in a storm that is ciia. Table 4 compaw b d c a s t rtcl and observed storm peak HS and associated TP at nine buoys. and in Table 5 these differences arc summarized statistically. The average measured storm peak HS over this sample set is nearly 11 m. The scatter index for peak HS varies only over a very narrow range of 0.14 (OWUG) to 0.16 (WAM4). Only WAM4 exhibits slight positive bias at 0.28 m, and OWIlG bas the largest negative bias of 0.94 m. However. as noted above in the scan~rplots, models fare &t in specification the of peaks in the range of HS up to 12 m or so. Considering the extreme storm peaks only at the two Canadian buoys that observed ESS (44137,44141). the models peak HS were b i d low by an average of 3.9 m for OWIIG. 2.7 m for Resio2G. 2.6 m for OWUG, and 2.0 m for WAM4. This behavior is also shown in Fig. 13 as a scatterplot of mean peak HS by m d l for each oe HS range.
b. Storm o the century f

Tbc cornpararive scaacrplots for SOC, bawd upon time historia a eight buoys,att shown in Fig. 14.The paaans t of tbc scatter for each m d l resemble those for HOS oe remarkably. even though the two storms are v a y diffcmlt al growth(asseenalsoat44141)withrrspccttotheother in movement and suucnm. T b e 6 givcs the cornpawive models but ultimately yields the second highest spcci- smistics bascd upon time historia, excluding a 4 8 h Bed peak sea state and matches the decay stage more spinup period. As &own forbe HOS comparisons all closely than the other models. WAM4 and OWUG fol- hindcansarcbasidyvayskillfu~compsrcdtonalrimc of storms provided by tbse same modeis. Tbe low the observed history closely on the growth phase ad.5 of the storm here and dccay too rapidly on the decay scaucr i n k for HS varies &om 0.18 for OWBG to 0 2 phase. WAM4 grows to witbin 1 m of the observed for Rcsio2G. Mean diffamcc in HS is smallcst for Remaximum ESS peak by virtue of an extra spurt in sjoZG(+0.04m) andlargstforOWIlG(-039m).The ~ g o t during the 3 b before the peak, which the other s c a in TP is lean for OWIlG and largcst for RaioZG, rwh models fail IO simulate. Tbis "exna" growth may be a m a n i f d o n of the tendency of the Rcsi02G model to a at some buoys in the ratc of dea result of the wind-wave coupling included in l g tbc m-nunts WAM4's atmospheric input source rum S. Despite crease of TP following the time of mer.of the peak , the magnitude of the peak HS at 44137 in HOS,the sea sea statc. Typical time histories of hindcast and measured HS state is stiU underdeveloped for a wind speed of 32 F m s-' , implying an enbancement to the coupling even and T' in the SOC at five buoys arc shown in Fig. 15, in extreme storm s a conditions. (Note that when we in order from southwest to northeast. At 41006 and e are describing thesc time histories we w the tenn 41002, the relatively simple s u u c n m to the measured "growth" in the sense of local change in HS and Tp; and hindcast time histories reflect the approach and an such changes may.arise in propagation andlor wave passage of the m i JS (SOC-JS 1) described in the last l generation and dissipation.) Downsrrcam of the a~ of section. A l models adck the observed rapid increase rm maximum waves, at 44011 and 41001. propagated in HS and TP during the approacb of this JS f o the oes swell becomes an increasingly more important com- southwess and BS seen in HOS all m d l fail to grow ponent of the total sea state. and all models seem IO all the way to the measured peak. As seen in HOS at agree more closely with cach other and with the mea- 44137 also, ResioZG lags the peak by about 3 h with surements in both HS and TP.At 4401 l, the tendency. lag in HS also on tbc dccay side with linle change in esrd for all models to slightly ovmpccify HS may be due TP.The record peak ESS (for a US. buoy) m a u e t to neglect of shallow-water processs. Dcspite these a 41002 is undaspcded by all models by between 2 differences in detailed m d l behavior there appcarj to and 4 m, in the same order as seen at 44137 in HOS; oe rm be no systematic effect amibutable to the m d l class. that is, f o highest to lowuc WAM4, RcsioZG. oe Indeed. when the models depart f o the measure- OWUG. OWIlG. Even though HOS and SOC were rm very different typcs of exhampical storms, 44137 in menu they all tend to depan in the same sense.

I

224
18

JOURNAL OF ATMOSPHERIC A N D OCEANIC TECHNOLOGY

VOLUME

13

MWel: Resio 2G

15

l5

i
E

12

12.

i
.
.
.,Dot

9

g9
6

. .

.-I
Ull?

UIIE
44119

3
0
Measured Hs (m)

n

18
MWel: WAM Cycle 4

. .

ItW2

'1ooo .am4

. . .
0
3
6 9
12

* Ut17

. . I n

.om

494,

.

Measured Hs (rn)

HOS and 41002 in SOC both lay in the path of the core of propagating energetic JS's that could be tracked UP stream over the previous 24 h a lcast with COR s p d s t ofabout30ms-'. At 44004.which lay ouOide the core of SOC-JS2 there is a double s f f u c ~ ~the time histones reflecthg fim in the influence of SOC-JSl hat approached from d~ south, foUowed by the influence of SOC-JS2 and ih associated wave field, which approached from the southwest All modeled time histories barely reflect the fim maximum

associated with SOC-JS2 strongly suggesting that the common wind field is deficient in specifying the pxioUS history and intcnsiry of this featurc. 'Ihe main peak i s generally well hiudcast w t the models exhibiting about ih a 2-m spread about the m a u e pcak Fanher downesrd Bt 44137, the dual-peak S CK ULE II in the timC histories is also evident. Again BE ~44004. fim peak i the s barely indicated i the modcls. ?he main HS peak asson ciami w t SOC-IS1 is undcrpredined by about 2-3 m ih by all models. However, TP is wdl spefifieed during the

-

:..

/

F e s a u ~ n rI 9 9 6
T A B U 6 Comprism .

C A R D O N E ET AL.
andbcs of 3-h .ndhindcra wave bcight and period .L buoys in mc swm of mC ccnony for four wave mukk.

225

a
Buov
nu

Signifimt wave height (m)

Ral:md (S) o
SI

No. Avg Avg
mcar hind

Ma m en
diff

diff Slddev 139 0.73
1.26 0.69 1.45 0.93 0.73 1.12 I .08

R

CC

No. Avg vu meas
33 33 33 33 33 33 33 33 264

Avg

hind

M a diff

m diff Stddev
1.01 0.64 2.10 3.61
13

SI

R

CC

,

4lWZ 41006 44004 44005 44137 44138 44139 44141

AII
RnioZG
41002

33 33 33 33 33 33 . . 33 33 264

4.91 4.04 -0.86 3.82 3 3 6 -0.45 5.03 4.48 -055 4.01 4.26 025 6.85 5.81 -1.04 5.42 552 0.10 ~. 5.85 5.36 -0.49 5.87 5.77 -0.09 522 4.83 -0.39
~

I .09 057

1.13
0.65

0.22 07 0.98 . 2 0.15 0.18 0.97 0.22 0.42 0.97
0.16
0.64 0.97

1.01 0.93
0.54

1.12 1.00 1.24 I .02
0.79

0.15 0.17 0h . 0.19 0.19

0.15 0.97
.OAS 096

9.23 8.92 8.71 838 9.64 9.78 8.19 1054 11.76 11.51
1235 13.05 1221 12.71 1216

12.35

0.15 0.99 0.39 0.93 033 0.95

10.71 10.73 9.23 11.89 8.71 11.01 9.64 9.00 8.19 9:ii I 1.76 1i.n 12.35 10.52 13.05 11.47 1271 I 2 2 5 10.71 10.98

1.78 -0.84' 1.62 -0.55 1.03 0.02 186

-0.31 -033. 0.14 2.35 -0.25 0.00

0.96 055 2.10 2.73
1.36 1.78 139 0.87 186

0.10 0.39 0.97 0.06 0.21 096

0 2 0.67 0.83 .2 0.33 0.88 0.73 0.12 0.39 092 0.14 0.45 083 0.11 0.24 0.93 0.07 0.30 0.96 0l 0.44 0.86 .7

41006 44004 44005 44137 44138 44139 44141
All

33 4.91 33 3.82 33 5.03 33 4.01 33 e 6.85 33 5.42 33 5.85 33 5.87 264 5 2 2

5.37
4.15

5.14 4.74 0.73 650 -035 4.w -0.81 5.42 0.44 6.16 0.29 0.04 5.26

0.46 133 033 1.07 0.11 0.79

I35
1.41 1.93 0.84 153 133

1.13 136 1.76 0.72 I50 133

0.25 0.76 0.95 33 0.27 0.61 0.89 33 0.16 0.67 0.98 33 0.28 0.73 0.94 33 0.20 033 0.94 33 0 3 2 0 3 0.83 33 0.12 0.18 0.97 33 0.26 0 4 0.84 33 0.25 O S 0 0.91 264

2.65 2.31 0.17 093
0.01

3.99 3.73
1.68 1.67 155

-1.83 3.70 -158 238 -0.46 1.27 0.28 2.71

2.98 2.93 1.67 139 155 3.21 1.78 1.18 270

0.32 0.82 0.56 034 0 6 0.42 .4 0.17 0.61 0 8 8 0.17 0.73 0.87 0.13 0.45 0.87 0.26 0.33 0.50 0.14 0.15 0.85 0.09 0.36 0.91 0.25 051 0.68

41002

41006 44004 44005 44137 44138 44139 44141

Au
WAMd

33 33 33 33 33 33 33 33 264 33 33 33 33 33 33 33 33 264
~~

-0.63 0.95 -0.41 an -0.14 0.97 0.68 130 -0.88 136 -050 0.82 5 3 1 -034 0.74 5.87 5-50 -037 1.06 522 4.87 -035 Lo2
4.27

4.91 3.82 5.03 4.01 6.85 5.42 5.85

3.41 4.89 4.69 5.97 4.91

0.71 059 0.96 1.12 1.03
0.66 050 1.00 os5

0.15 0.12 0.98 0.15 0.27 0.97 0.19 0.61 0.97

028 0.73 0.94 0.15 0.18 0.96 o o n 0.95 x 0.09 0.15 0.98 0.17 036 0.93

a18 034 os5
0 3 0.12 0.99 0.17 0.24 0.97 Or19 055 0.97 035 0.67 0.90 0.14 0.15 0.97 0.14 036 0.96 0.11 0.15 0.98 0.19 039 O S 0.19 033 0.k
~~~~ ~ ~

9.23 8.86 8.71 8.25 9.64 10.01 33 8.19 11.04 33 11.76 1157 33 IUS 1130 33 13.05 1206 33 1271 12.31 2~ l a 7 1 1.0~0 33 33 33 33 33 33 33 33 264 9.23 8.71 9.64 8.19 11.76 12.35 13.05 12.71 10.71 9.30 853
10.55

33 33 33

-0.37 -.6 04 0.37 285 -0.19
-0.05

0.89

0.81

0.09

0.88 1.90 4.09 1.49 200
2.01

0.75 1.86 2.94 1.48

200
1.75
126

-0.99

-0.41 132 0.09 206
0.06 1.07 -0.18 0.70

26 0
1.07

0.09 0.19 036 0.13 0.16 0.13 0.10 0.19

0.33 0.30 0.64 0.82 05. .7 055 030 0.42 0.48

0.97

0.95 0.86 0.63 0.90 0.81 0.89 0.93 0.83

.-. "...
41wZ

41006 44004
44005

;a
!

44137 44138 44139 .. . 44141 All

4.91 4.42 -0.49 081 3.82 3.41 -0.40 0.75 5.03 4.88' -0.15 0.97 056 1-72 4.01 4.57 6.85 5.90 -0.96 135 04 5.42 4.95 - . 6 0.88 5.85 5.24 -0.61 0.89 5.87 5.47 -0.40 1.16 5.22 4.86 -036 1.07
~ ~~ ~~~

465 06 .4 0.96 I .42
0.95 0.75 0.64
I -09 1.01
~~~~

1. in
12.41 1293 12.53 1283 Il!Z9

1.80 220 -052 1.75 0.12 129 0.59 2.26

0.91 3.08 0.65 058

2.35
4.51

0.67 217 3.38 1.68 212 1.67 1.29 2ilB

0.12 0.45 0.95 0.08 039 096 023 0.64 0.82 0.41 0.76 0.55 0.14 055 0 9 .0 0.17 0.61 083 0.13 0.42 0.94 0.10 055 096 023 055 0.83

SI-

index; R--ndo of poinu .bow and below 4 P lins;CC--corrslarion coeflicicnr

.
groowrh and approach t the peak by aIl models. Downo sueam of 44137. at 44138, w i h lay w l to the right of hc el the core of SOC-JS1. the models arc in much bemagme-

mentwithrhcmcarurcments. Table 7 c o m p m hindcast and observed peaks of HS and associated T p at eight buoys, and in Table 8 . these differences arc summarized statistically. The average measured storm peak HS over t i sample set is hs 1 1.70 m. slightly greater even than for the HOS dataset The scatter index for HS varies only over a vely narrow range of 0.10 (OWI3G and WAh44 tie) to 0.15

(OWIIG). All models exhibit-a negative bias in HS. which in both O W models is about 1 m. Resio2G exhibits the smallest mean difference in peak HS and the lowest mean difference and scatter index in T. HowF ' ever, as also seen in the HOS comparisons of peak;, he model biases in peak HS arc negligjble for HS less than 12 m and negative for HS greater than 12 m. Considering only the thrce buoys at which peak HS altained ESS levels (41002 44004.44137). the mean negative bias in modeled peak S W H was 2.7 m for OWIIG, 1.4 m for ResioZG. 2.0 m for OWUG. and 1.1 m for

226

JOURNAL OF ATMOSPHERIC A N D OCEANIC TECHNOLOGY

VOLUME 13

YIII

1.7

13

14

15

16

17

FIG.15. Cornpariron of hindcast end m u r u d time histories wave height md period i h e stom of thc m u aI five V l a c d b y , for fW n ay modclt.

W-4.

The pattern of bias for each repime. shown in 2G wave model. and the latest cycle (WAM4) of the Fig. 13. i r m r a l like that found in HOS. s eakby WAM 3G mcdcl are adapted t the wcstcm N o d Ato lantic basin on grid systcms with common resolution, and are driven by common wind fields for both storms. 6 Snmmarg and condusiom . The wind fields were devclopcd using carcful manual In this study, four widely applkd spectral wave h d kinematic reanalysis using d conventional data, inoels. namely, the Oceanweather 1G and 3G m d l , a cluding ship and buoy observations received (00 late oes

...

..

--

/

0

F-RIJARY19%

C A R D O N E E T AL

227

-1

1

-i

1

0
FIG.IS. (Conrinucd)

for use i real time, A crucial part of rhis analysis is n the adjustment of buoy wind spccds to the 20-m effective neutral surface wind for anemometer height, stab l t . and avMaging interval effecu. A sea-statedeiiy pendent bizs between vector- and scalar-avemged derived measuIcd wind spccds was inferred f o the rm m a u e e t and was used t iesrmns o n vector-aver-

.

aged buoy wind spccds by up to 12% in the most exneme sea states. Wave hindcasts were evaluated against time series of measured HS, dominant wave period and wave spectra obtained at nine U.S. and Canadian buoys moored in deep water between offshore Georgia and Newfoundland. We conclude f o our analyses that rm

FIG. (Continued) 15.

all wave models investigated here performed very well for wave conditions up to about 12 m HS. These hindcasts exhibit considerably greater skill than rcaltime wave analyses derived by some of these same models operating at U S , Canadian and European .. centers, confirming that the typically large errors in operational surface marine wind field analyses arc the dominant source o emrs in operational wave f analyses and forecasts. In conuast with the successful performance of the wave mod& in sea states up to 12 m. all models tended to underpredict the most extreme sca states. It k unclear whether this disagreement is due to r m i i g uncereann tainties in the wind field or to biases in the wave models

at high sea states. Dcspite the large differences in the parameterization of the physics between the models. differences in skill in specifications of time histories of HS,TP.and spectral shape (spectral comparisons were made but arc not shown) averaged over all buoys wen relatively slight. It was noted rhat when the models disa p e with the meaSmmenu in time series cornparisons. they tended to disagree in the same sense, suggesting a source either in residual wind field errors or r m i i g deficiencies in wave model physics common eann to aU classes of models. A common phenomenological ln between the wik trcmely high waves in these storms appears to be wave generation along a dynamic fetch associated with in-

TABU 7. Smrm peak mcasurtd and hindsarl wave kiphl and pricd 81 buoys in rhc rtonn Of the cmnuy for four wave nude M a d
Buoy
41002 41006

a
1410)

OWllG HS
10.4 72 11.7
93

Resi020

OWUG
HS
11.9 8.0 12.3 10.1 9.2 13.4 8.5 10.9 9.7

wAM4

HS
14.6 9.7 13.1 8.9 6.7 15.9 9.4 11.7
103

TP
16.7 12.5 153 126 13.7 I73 173 16.6
18.4

DDM
14#Q4

TP
14.0

DDMH
14/04 1 -

HS
12.5
9.0 13.8 113

TP
143 123
143

DDM
14/05

TP
14.6 11.8 15.9 163 145 17.4 17.8 17.4 17.5

DDMi
14/04 lM)I 14/14 1 0 14/06 15/02 15/15 15/09

HS
127 8.4 132 10.7
93

TP
15.0 11.8 16.7 16.7 153 18.0 183 18.1 18.2

DDM
14101 14/14 1 0 14/m IM1 15/14 1M8 15/07

IY23
14/10 14/13 13/19 15/m 15/14

1I.S
153 15.6 14.3 17.6 17.6 18.0 17.7

la3
14/11 1 0

44004
-5 44014 44137
44138 44139

I1 U3
14/19

83

(3.5
10.8 11.7
10.9

1 W 15rW

14.3

ISM
IM8 15/10

15108
IM)9

44141

132 10.1 10.6 11.4

16.7 16.7 143 16.7

1Mn
1m 5 1m S 15/07

14.4
9.3 11.5 10.6

IML9

HS--rigniiBcanl wave beighc (m):T P@ period () OD/HH--drymWr of muimUm HS. s; Mcsnocd wave height d peal: p i arc rmoolhd Wing a 1.1.1 d n g I-. ad

.. .

. ..,
' .

0

1

FEBRUARY 1996

CARDONE ET A L

229

TABU8. Suurtlcr of dtffnnro bclwecn smnn peak mcLtlllcd and lumbsl -Ye k g h t and p o d a buoys f m he srom of thc anmy for four wave models.
Sigmficanr wave hcighr (m)
Avg

Peak pnod (5)

Avg
lund

Ma en

Model
~~

meas
1. 17 11.7 11.7 1. 17

&
-.1 10
-0.21

m &ff
206 16 .3 139 11 .7

Srddev
'

SI
0,lS 01 .4
0.10

R

CC

Avg Avg rn lund

Man

nnr.
dtff
137 1.25 1.60

&ff

Srddedev
1.25

SI

R

CC
'0.73 08 .1 0.69 07 .0

OWIG
Ruio2G OWUG WAM4
SI--rcytcr

1. 07
115

1. 06 11.4

-1.10 -0.35

1.80 16 .2 1.15
1.12

0.10

03 0 6 .8 . 7 0.50 0.76 0 1 0.90 .3 038 0.90

15.9 15.9 0.0s 15.9 15.0 - . 1 09 15.9 1. 61 0.20 15.9 .16.6 0 7 .4

1.79

157 . 0.10 0.63 0 0 0.13 .8 0.10 0.50 159 0.10 0.63 16 .3

.

index: R--ratio of p i n u above and below 4f line: CC-correlation coeffisicnr

tense surface wind maxima or jet shlaks. which maint i high spatial coherency o v a at least 24 h and p o an rp agate at about speeds of 15-20 m s-' . The ESSs arc observed in these two storms only at those buoys directly in the path of the jet core of these features. Unfortunately, such detailed surface wind field feaNnS arc resolvable only through kinematic analysis and only when these features evolve within the relatively dense network of buoys just off the a t coast 'Ibis suggests that operational objective analyses and forecasts will have even greater negative biases in extreme s t o m . Consequently, such information should not be used for applications that require accurate specification of exWme sea states. These results suggest strongly that for applications where supercomputers are not available. and especially for most operational applications when only integrafed properties of the spectrum (e.g.. HS,TP)are required or where errors in forcing wind fields are t p y ical of real-time objective analyses and forecasts. highly developed and validated 1G and 2G wave models may continue fo be used. However, accurate specification of ESS is especially critical for application of wave models to determine the extrcme wave climate for ship, offshore, and coastal suucture design. Therefore, further study i required to isolate the cons mbution of remaining wind field errors and model physics and numerics to the underprediction of ESS in these storms. Finally. this study raises the possibility that very extreme sea states may be much more common than earlier thought in the waters off the East Coast of N o h America and suggests that the extrcme wave climate needs to be updarca if not reevaluated altogether.

gram; permission to publish is granted by the Chief of Engineers.
REFERENCES
Axys Envimnmmlnl Systems. 1992: Ops-rationrlmninPMnceIrcfc~cascm dfor NOMA0 ZEN0 payloads. A X Y S Envimnu mend Systems Sidney. B C @appendices. pp. plus I

Beal. R C, 1991: LEWM: MMvatioa objcctivcr and rrrulu. Dinnionol O C ~ M Wovc S p m m R C. Beal. Ed.,Johns H~pkins
univarity Rc+s 2-12 Burgm. G.. 1990: A guide to the Nedwave

model. Scientific

Rcpolu W R - W , ffininklijk Nederlands MeMmlogiwh hrtituut. 81 pp. [Available fmm I M. P.O. Box 201,3730AE M I

de Bik Wilhclminalun IO. Naheherlandr.1 CMlemu D.and G. pahu.1992: A mercomlogid ovcwicw of thc .
WlOWCCU

SrOLm Of 1991. h p f h U . T id IN. WOrkhoP m h r Wave HindcMing ad Foomprring. Monawd PQ. Gnad.,Ai-

morpberic Envimnmmr saia.SI-S44. cardone. V. I.. 1%9 Specification of thc wind dimibution in rke marinc~lsyaforwavcforsarting.131pp.[~AD 702 490.1 , 1991: The LE\KEx wind fields and barelinc hiDdcpr~ Dincr i o ~OCIM Wave Specnu R C. Bcal. E . IhsJohns Hopkins l d. Univmity m.136-146. , 1992: On thc s o w n u t of U d n c wind field m c x m . c a l C I storm. h p r i n m . T id Inr Workhop M Wme Hindcpning h r rmd FOtTC&#. M a d PQ. CaDada AlmorphcriC EnVimS

-. -.
-.

men1 Sewice. 54-66.

K. C. Ew- 1992: Validdon of the hindcaw appmach the spxificzion of wave conditions at the M.ui I d o n off h e (yc11 c w t of New zcaland. Rcprinu. Third I n t Workhap an Wave Hindcawing ond Forecasing, M a d . PQ. casad.. Amorphcric E n v h n m l Senice. 232-247.
and
10

W. I. Pi-n. and E G. Wad, 1976 Hindcming thc &dad spcnra of hlmiavle gaumai waves 1. Per ZechnoL.

Acknowledgments. Suppon for this project was provided by the Canadian Federal Panel on Energy Rcsearch and Development. The manual analyses were performed at Oceanweather by a team of MeteomloSphrM rmd & C M sySI.3lL a C C C Q l d gists including B. T. Callahan. S. T. Keeney. and ckncy.R M.. and W. D.sadln. 1992: 'Ihc R a N u m u i d 00.nography Onvr Suile of Oceanographic Models d RoQum. M. J. Parsons. The work conducted by R. E. Jensen a Fon&g* 7,3U7-327. w s performed under the Coastal Flooding and Storm C-R W eW.D .D. 1.R n i a R M. B r m k B. A Ebarolc. R. E. a . Rotection Program of the US.Army Engineers GcnJ m n . D S. Ragsdak and B. A. T m . 1981: A h d C amsl . hindcart ~ w rignificanr wave inforrrmion. WIS Repon m eral Investigation Research and Development Pro-

28.385-394. A 1. Broccoli. C. V. Grrcnmwd, and 1. A Grrmwood 1980: . . ErmrchMspnroCI of u n a n u p i d rronn wind Bel& m d f k fmm hiswid dWd. 1. PCh TrrhnoL. 3 2 873-880. -,I. G. Glcmwcd, d M . A Cane. 1990: On& inhinnical marine wind dam 1. C l h c . 3 113- 1 7 , 2. H. C. Glabcr. R. E Isnwn. S.Harselmanhand M. 1. CMUO. 1994: In JcMh of he m e surface wind field during SWADE IOP-I: Ocean wave nmdelling #vc. 7he Glob01 A m

-.

i

c

230
2 US. Army E. n

JOURNAL OF ATMOSPHERIC AND OCEANIC TECHNOLOGY

VOLUMe 13

1 :

W-ya Expaimmt Srmiop Vi&burg,MS.I2l2pp. ' Donclan. M. and W. 1 piason. 1983 Ibs -ling . . variability of atinma of spuxrn of luind-garrpred gaviy wayel. 1 G . o pes. Res.. s(I(C7), 4381-4392 E d B.M. C. M ~ l l E Dunlop. and T. pime. 1992: Development i . ~ h of wind and wave climate ulnr for rhc a t c o s t of Canada and the Grrat Lakes. Rcprilur. Third IN. Workhap on Wave HindCaning ond FOrrcMing. Monncd FQ. Canada Atmorphcric Eovimnmmt Service. 362-372 Giihouws D B. 190. A field evaluation of NDBC mrmcd buoy .. awdr 1. AaML occnnic TechnoL. 4 M- 104. Glcenwoob 1. A.. V. I. cudonc. and L M. Lawn 1985: Interconpison rc~t mia of du SAIL wave model. Ocean Wave MadeUin8. rhc SWAMP Gmup. P l a m Rcu.221-233. GuiUaumc. A.. 1 9 : Sednicd 1 0 for the comparison of surface 90 0 gaviry wave rpecaa with lppiicndoo t model validation. 1. o A m w Ocermic TechwL. ?,SSI-561. Guntha. H S Hauchnann. and P. A E. M.Januee 1991: Wpmodel , . cycle 4. T a h Rcp. 4. DcuuchsJ I(lim0 Recheat Z n m edited sm by ModeUbe~mgr(puppc. Hamburg, Germany. 91 pp. Hsuclmans K.. 1974 On du qaud dirripafion of ocean w a v u due m whirccrpPing. BOund-Layer Merear.. 6 106- 121. Hasdumu S. K Huvlmans 1 H. AUmdn. and T. P. Bamcg . . 198% Compuwiona and paramcrcri2adons of du nonlinear acrgy &a i a gaviry-wave ~posmrm n panmemiran Pan : tion of the w n - l i i anrufn for applicuion in wave models. 1. Phys. Oceanagr.. E 1378-1391. , Janrrss P. A E M. 1991: Pupri-linear msOry of wind wave gcn.. d o n applied m wave forarrdng. 1 Phy& Oceanagr.. 2. . l
~

Komss G. J.. S. Hauclmann, and K
8

Ihdmam.

1984 On

Oceana8r..

14,1211-1285.

145-154.

Kahmn. K K. and C. I. Calkom, 1992. Reconciling divrcpMcicr in the obsaved groowlh of wind-gaemd waves. 1 Phys. .
Oc-gr.. 72, 1389-1404. Khandek, M. L R Lnlbch.y. and V. J. cnrdonr 1994 Ibsp a . formnnsc of du c.rud*n specual OSCM wave modcl (CSOWM) during Q Grand B& ERS-I S A R wave rpcpa

vliid.donapcrimmt.AaML-Ocem, 32,31-60. K o c i s P. L D. A Olson. A F. Wick. and R D. Hamn. 1991: ~ m c rM c I ~ o ~ . n,459-472 . sor. SufaceweathcrdyaisathcNaIionalMacornlogidCcnrer: . Cvrrcnl pmblpnr and fume plaru W m Fomcanh8.6 289- W u 1. 1962 Wd-rrrsu cafficienu ova the x a surface fmm breeze m hunicmc. 1 Gophys. Re*. 87,9704-9706. . 298.

Maat. N. C.. C. Kmau and W. A 0 % 1991: The rnughws of the 0 wind wavu. & d - & y e r Mereor.. 54,89- 103. Nickenon. 1. W.. 1993 Freak w w u ! Mor. Wea Lox. 37.14-19. NCDC. 1 9 9 3 N O M Marine EnvimnmenlalDr on CD-ROM. 14~ an disc set available fmm N a i o d Climatic Dr Center. Ashean ville. NC. 1993b Ibs Big OK! National C l i d c Dua Cater. AsheWUe. N C 25 pp. NOAA. 1992: 'Ihe Hallorvepl Nor'carter of 1991. N d Disaster Survcy Repon Available fmm US. apamncnt of Commerce. 61 pp. plus Appndicsr. OvcrlanQ J. E. a d W. K G & U . . 1978: Marine w i d in the Ncw York Bight. Man Wea RN.. 105.1W3-1008. Reex. A M.. and V. J. M n c . 1982:Tol of wavc hindcast modcl r u d U against masuremenu during four metcornlogical sysBins. offJhomTechnology Con$. Hnumn TX 269-282. Ruio. D. T, 1981: The erdmarion of a wind wave semm in a pc u d h t t e qaud modcl. 1. Phys. O-r. cg. 11,SIO-SZ. and W. m . e 1989:Impticarions of M r equilibrium m g for wind-gennodwave% 1 Phys. Oceamgr.. 19, 193-204. . R 0 s h D . B - V . 1. Cadone J . E . OvaladRMcPh-n. W.1. piason. and s. w. Y u 1980: o x a n i c surface winds,A.tvrmcu in c m p h y r u , Vol. 27. Academic Rczr 101-139. Ssnden F.. 1990: Surf= wendm annlysis o v a the ceca.@ Scnrcbing for ICO.mh W m F o m c h g . 5.5%-612. t. Snyder. R F. W. Dobron 1 A Elliolf and R B. Long. 1981:A n a y . of lmKuphcric prusmt All-2lldoru above sur. face g d i w a v a 1. F W Mech. 1oL,1-59. r vy SwaiL V. R. V. I. Cardone.and B. M. E. 1992: An exmme wind d and wave hindcart off the west coast of cvlada Rsprinu. Third I n t Workhap on W a n Hindcaning Md Forecprring. M o n d . FQ. Canada Armorphcric Environment savice. 362-372 Ucallini.L W , S . F . C d d i , N . W.Junka,P. J. Kocin.andD.A Olson 1992: Rcpon on the Surface halysis Workshop bcld I I du NadDnal Mnmmlogid Cater. 25-28 Mvcb 1991. Bull

-.

-.

a

Appendix F. Additional Storm Hindcasts. Following completion of the baseline hindcast study consisting of 30 storms and 3-continuous years, four additional storms were hindcast as follows: January, April, April, March, 1979 1979
1982 1988
..

0 4 / 0 0 0 0 UTC

20/1200 UTC
2 5 / 0 0 0 0 UTC 0 6 / 0 0 0 0 UTC

-

ll/1200 UTC
2 3 / 0 0 0 0 UTC 2 7 / 0 0 0 0 UTC 09/1800 UTC

. ...- .
:i. ht
.,.\
.I-

~.-

,.. 0 ”’
. ,

The January, 1979 event was selected because a time series of wave measurements was available in very shallow water offshore Yanbu. The basin-wide wind and wave hindcasts were made using exactly the same procedures as used for each of the 3 0 historical storms. The wave data were not referred to in the hindcast, and only the results from the DHI nearshore model were compared to ‘the measkements. The attached figure shows the time histories of hindcast-wind speed and direction and selected wave parameters at a grid point in deeper water offshore Yanbu. The storms of April, 1979, April, 1982 and March, 1988 were hindcast in order to potentially provide more data for the .-directional analysis of wave extremes in the central Red Sea, especially for southerly wave approach directions. The attached tables give storm maxima of wind speed and wave height at grid points which are used to drive DHI’s nearshore wave model in the centra1,regions. It is seen that only the storm of March, 1988 is a significant southeaster, and in fact this event ranked as the second highest southeaster in this region relative to the base’line population of 30 storms.

.~

0

RED SEA STORM 7901 - G.P. 759

h

c

.

c/3

I

I

I

I

I

1
-I

360

o ~ ~ " ~ " ' ~ " " " ' l " ' l " ' l " ' ~ "
1/5/79 1/6/79 1/7/79 1/8/79

-__-

1/9/79 1/ I 0179 1/I 1/79

/

ws 15.0930 13.8722 12.9830 11.9928 11.7349 11.2829 12.6680 11.5806 11.4854 9.3787 14.0564 13.4649 12.6176 12.0981 11.6338 10.1554 11.1824 9.8419 10.4714 10.7504 10.2515 8.4389 7.7959 13.0021 12.6824 11.8136 11.7471 10.8147 10.9289 10.7200 10.4435
i

wd 137.2448 140.1706 131.8115 135.9455 132.0406 133.1116 149.8912 142.9117 149.0016 155.1783 179.1013 184.5687 184.7307 189.4738 159.0269 164.3786 161.6170 201.8700 147.8710 332.5787 335.2945 191.4708 352.4751 322.2114 323.0365 319.8576 322.7745 319.7823 335.1564 320.3435 322.9832

hs 2.6337 2.7244 2.2729 2.3770 1.17841 1.0914 1.8778 2.1172 1.5212 1.7909 2.2998 2.3341 2.2163 2.2893 1.4474 1.6311 0.9045 1.0323 2.1043 1.8676 1.7673 0.0415 1.2584 2.2931 2.3243 1,8937 1.8992 0.9913 0.9280 0.8837 0.8956

tP
7.1754 7.3204 7.0520 7.3050 4.5437 4.5680 5.9745 6.6348 5.5618 6.6482 6.6848 6.7472 6.8430 7.0581 5.3068 5.9284 4.4723 4.4562 5.4198 6.5793 7.1114 4.5730 7.2567 6.8085 6.8624 6.3077 6.3203 6.1857 6.1633 4.3354 4.4322

vmd 330.4122 329.9443 330.5777 330.0375 325.9048 324.0522 328.5757 326.4211 323.6289 321.1217 345.9742 347.1053 355.3525 355.9715 339.4593 337.0224 359.8095 21.6013 335.1548 145.7534 129.2022 14.7364 49.0007 144.5504 145.2706 137.7353 137.4081 111.8515 116.1966 147.8517 148.0542

dhm 211200. 197904. 197904. 211400. 197904. 211200. 197904. 211400. 197904. 211200. 197904. 211300. 197904. 210600. 197904. 210900. 197904. 211200. 197904. 211600. 197904. 221200. 197904. 221300. 197904. 221200. 197904. 221300. 197904. 211200. 197904. 211500. 197904. 211200. 197904. 221300. 197904. 201900. 197904. 230000. 197904. 230000. 197904. 201300. 197904. 222300. 1'97904. 221800. 197904. 221900. 197904. 221200. 197904. 221300. 197904. 221300. 197904. 221800. 197904. 221200. 197904. 221300.
Ym

9P 262. 262. 295. 295. 335. 335. 419.
419.

572. 572. 630. 630. 681. 681. 694. 694. 699. 699. 710. 710. 759. 776. 776. 877. 877. 944. 944. 946. 946. 1006. 1006.

juldate 28966.500 28966.584 28966.500 28966.584 28966.500 28966.541 28966.250 28966.375 28966.500 28966.666 28967.500 28967.541 28967.500 28967.541 28966.500 28966.625 28966.500 28967.541 28965.791 28968.000 28968.000 28965.541 28967.959 28967.750 28967.791 28967.500 28967.541 28967.541 28967.750 28967.500 28967.541

!

I i j
!
i
I

!

j
i

!

i

1

ii
I

!

ws 11.2213 10.9213 11.8241 11.4083 10.4426 10.3028 13.0086 13.0915 11.9304 14.5163 13.1471 15.9344 14.4273 13.6892 12.3881 10.2477 6.9583 12.4490 11.9019 6.2066 4.3050 7.0081 2.8488 7.2245 3.3978 10.5479 9.4502 10.4104 9.9254 9.5311 8.3684 10.2323 10.6670

wd 142.1279 144.6372 144.9677 146.9555 144.0879 146.4738 148.7914 141.6137 144.0112 142.5208 145.8413 145.4907 147.9244 159.5122 162.0446 158.7404 264.7839 172.6543 173.0702 168.9141 332.5125 181.3350 335.6594 180.1284 349.2123 16.1454 318.9538 8.6668 320.2020 8.1478 319.4814 334.3059 333.9224

hs 1.7283 1.7509 1.7689 1.8232 0.9393 0.9522 2.3698 2.4660 2.6010 2.5941 2.7742 2.3192 2.4074 2.2288 2.4146 1.5288 1.9019 1.1545 1.1735 0.8940 1.6005 0.7813 1.4560 0.7772 1.2541 1.0124 1.4354 1.0281 1.3219 0.4949 0.7173 1.1430 1.1032

tP 6.1051 6.1670 6.0697 6.1883 4.3754 4.4245 6.8505 7.0807 7.4407 7.0377 7.4388 6.7491 7.2595 6.5774 7.1440 5.9140 7.4168 5.3510 5.3741 5.1391 7.9784 4.8261 7.9870 4.9001 7.8995 4.4232 5.8478 4.4690 5.4326 3.5900 3.7861 4.3997 4.5451

vmd 332.8807 332.6173 332.9952 332.7794 331.2665 331.0556 327.5539 324.2003 321.3189 332.8394 331.2477 344.8318 344.5577 345.5519 344.9999 334.0561 333.6844 6.1004 7.1735 330.5297 326.6567 340.6288 339.5965 353.0845 352.6660 174.9796 147.5398 165.5227 143.2430 140.4633 119.8607 161.4015 161.4008

ym dhm 198204. 260600. 198204'. 260800. 198204. 260600. 198204. 260800. 198204. 260600. 198204. 260700. 198204. 261200. 198204. 261200. 198204. 261400. 198204. 261200. 198204. 261400. 198204. 261200. 198204. 261400. 198204. 261200. 198204. 261400. 198204. 261200. 198204. 261700. 198204. 261200. 198204. 261300. 198204. 261200. 198204. 262200. 198204. 261200. 198204. 270000. 198204. 261200. 198204. 262300. 198204. 260600. 198204. 262000. 198204. 260600. 198204. 261900. 198204. 260600. 198204. 262100. 198204..260500. 198204. 260600.

gP 262. 262. 295. 295. 335. 335. 419. 572. 572. 630. 630. 632. 632. 681. 681. 694. 694. 699. 699. 710. 710. 759. 759. 776. 776. 877. 877. 944. 944. 946. 946. 1006. 1006.

juldate 30067.250 30067.334 30067.250 30067.334 30067.250 30067.291 30067.500 30067.500 30067.584 30067.500 30067.584 30067.500 30067.584 30067.500 30067.584 30067.500 30067.709 30067.500 30067.541 30067.500 30067.916 30067.500 30068.000 30067.500 30067.959 30067.250 30067.834 30067.250 30067.791 30067.250 30067.875 30067.209 30067.250

ws
15.7502 15.0708 13.9105 13.5606 12.7669 14.4627 13.5935 19.2854 18.4873 17.9772 17.2570 16.1791 14.6852 14.7355 13.4900 14.5992 13.4080 13.4389 12.6303 15.0190 14.1176 16.3480 15.1392 6.9270 13.5582 12.9851 11.4017 11.6956 10.7770 10.9509 10.7764 10.4965 9.5702

WD
149.8852 152.9041 149.7924 156.5118 145.1750 140.6445 14 1.4644 164.9839 169.3916 165.0756 171.0393 157.0486 179.2866 173.4257 189.1833 175.1903 187.9304 159.0831 172.7659 336.5316 336.8428 324.8174 324.2637 329.7906 311.6964 321.0153 328.4004 322.5155 329.5509 . 323.0556 325.2429 306.7661 34 0.57 11

ES 3.3950 3.5771 2.7962 3.0016 1.3112 2.4.081 2.5571 4.1419 4.1628 4.4296 4.5009 3.0450 3.7549 3.6342 3.6992 3.6183 3.7214 1.1479 1.2426 2.7727 2.9794 2.8010 2.9335 2.3971 2.0008 1.7739 2.3247 1.5149 1.8574 0.3101 0.9305 0.7725 0.8253

TP 8.2703 8.6983 7.9477 8.6055 4.9124 6.6559 7.0266 8.6127 8.7185 9.0787 9.2476 7.9859 9.0400 9.0070 9.4137 8.9306 9.2942 4.8525 4.9580 7.2724 7.5932 7.1861 7.4363 9.6268 6.6665 5.5498 6.6654 5.4638 5.9399 4.9935 5.3525 4.1910 4.2907

VMD

YM

DE
90000. 90200. 90000. 90400. 91200. 80600. 80800. 81200. 81300. 81200. 81300. 80600. 81400. 81200. 81400. 81200. 81400. 80600. 80300. 91200. 91400. 91200. 91400. 81800. 91200. 81200. 30800. 81200. 30800. 81200, 81300. 80600. 81600.

338.5100 ' 338 .E286 338.2021 339.9999 335.3146 319.2154 320.6613 335.2579 335.6700 347.4186 348.4880 357.4688 2.9233 353.3083 356.8338 342.2593 345.1570 355.6675 6.3465 152.7406 153.9837 137.5540 138.0401 359.6677 114.5381 124.2472 142.6732 127.9386 133.8692 1 1 1.2242 112.6927 113.9426 164.6698

198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803. 198803.

GP 262. 262. 295. 295. 335. 419. 419. 572. 572. 630. 630. 632. 632. 681. 681. 694. 694. 699. 699. 710. 710. 759. 759. 776. 776. 877. 877. 944. 944. 946. 946. 1006. 1006.

JULDATE 32211.000 32211.084. 32211.000 32211.166 32211.500 32210.250 32210.334 32210.500 32210.541 32210.500 32210.541 32210.250 32210.584 32210.500 32210.584 32210.500 32210.584 32210.250 32210.375 32211.500 32211.584 32211.500 32211.584 32210.750 32211.500 32210.500 32211.334 32210.500 32211.334 32210.500 32210.541 32210.250 32210.666

APPENDIX C
DHI’s EVA Program

Danish Hydraulic Institute

EVA

A Programme for Extreme Value Analyses

EVA is a computer programme for performing extreme value analyses on a sample of data and for graphical presentation of the results (see on the verso for example of output from EVA). The programme performs extreme value analyses on data samples extracted as annual series data or as partial series data (ie peaks-over-threshold data). Standard statistical tests for independence and homogenity are carried out on the data. The data can be fitted by up t o six different candidate probability distributions: exponential distribution lognormal distribution Gumbel distribution Frechet distribution Weibull distribution Weibull distribution (upper bound) The statistical parameters in the selected distributionk) are determined either by the maximum likelihood approach or by the linear least squares method. Chi-squares tests and Kolmogorov-Smirnov tests are carried out for analysing the goodness of the fit. Determinations of events with user-specified T-year recurrence period can be performed. A Monte Carlo simulator is included for calculating the standard deviation of the T-year estimates and for deriving confidence bands covering the whole data sample and the extreme values.

Application of EVA

Computer Requirements

EVA can either be used as a standard module in the UNlX version of MIKE 11 or as a stand-alone programme available for UNlX computers. A DOS version for personal computers is also available. EVA is available a t DHI on general commerical terms.

Use of Programme

INPUT DRTA FILE ANRLYSIS INPUT FILE

: Recorded uave heights, partial series : C00leva.prt

B A S I C DATA ANALYSES

=
I,

M c

0,

..* ""-

Sample size Mean value o f sanple Threshold value St. de". of sample Events Per Year Skewness of sanple Median of sanple

-

34 6 9 . 0

I

5.35
1.02

=

=

-

2. 125 0.425

6.87

'

STRTISTICAL ANRLYSIS Rnalysis type Distribution Fftting nethod aIpha beta

-

Partial GUIlBEL ML :.E5 6.228

GOODNESS OF FIT TESTS Chi s w a r e d : Highest level of significance = 10.071 Z Kolnogorov-Snirnov test : Accepted at 20% leuel o f significance

'Id'
2..
S . '

'..

H .

EST I MATES Recurrence period (years) Non-exceedance probabilitY Estinated Value Estimated Standard deviation

10 0.9529
9.18

50
0.9906 1 . 3 0 7 0.34

1 W

0.54

0 9 5 . 9 3 11.38 0.97

Example of graphical output from EVA using a Gumbel distribution on a sample of data.

For further information, please contact: Danish Hydraulic Institute Ports and Offshore Division, Agern All6 5. OK-2970 Hsrsholrn Phone: +45 45 76 95 55 Fax: i 4 5 45 76 25 67

-

0

APPENDIX D
Short Description MIKE 21 HD

0

a
MIKE 21 HD Release 1 1 .

HYDRODYNAMIC MODULE

A SHORT DESCRIPTION

a

DANISH RYDRAULIC INSTITUTE Agern All6 5 DK-2970 Hamholm Denmark

0
Telephone Telex Telefax +45 42 86 80 33 37mdhicphdk +45 42 86 79 51
-

MIKE 21 HD -

HYDRODYNAMICMODULE

WORT DESCRIPTION
Introduction

MIKE 21 is a comprehensive modelling system for 2dimensional free surface flows where stratification can be neglected.

MIKE 21 HD is the basic module of the entire MIKE 21 system. It provides the hydrodynamic basis for the computations performed in most of the other modules, for example the Advection-Dispersion, Water Quality and Sediment Transport modules.
MME 21 HD simulates the water level variations and flows in reponse to a variety of forcing functions in lakes, estuaries, bays and coastal areas. The water levels and flows are resolved on a rectangular grid covering the area of interest when provided with the bathymetry, bed resistance coefficients, wind field, hydrographic boundary conditions, etc.
The system solves the equations of continuity and conservation of momentum using implicit finite difference methods.

Applications

MIKE 21 HD is applicable to a wide range of hydraulic phenomena:

Tidal exchange and currents

Storm surge

a
H E L O O L A N D . WEST

----

GERMAN*

COMPUTED RECORDED

m

ai

Secondary circulations,eddies and vortices

AYPLICICATION
WAVE P E R I O D
8

FaCIODS

I2OOSEC

...
2

MIKE 21 HD - Short Description

0

Dam-bmk

.

..
0
0
-

Tsunamis

SEA SURFACE IMMEDIATELY AFTER EARTHQUAKE

TSUNAMI AFIER 14 min.

MIKE21HD Short Description

3

Equations
The hydrodynamic module of MIKE 21 solves the vertically integrated equations of continuity and conservation of momentum in two horizontal dimensions. The following effects are included in the equations:

0
-

-

-

convective and cross momentum wind shear stress at the surface barometric pressure gradients Coriolis forces momentum dispersion ('eddy') sources and sinks (both mass and impulse) evaporation.

continuity

F+g+p=~-e t Y
X

x-Momentum

4

MIKE 21 HD - Short Description

... .

e

g
f

C

PW

n

X,Y

water surface level above datum (m) flux density in the xdiection (m’/s/m) flux density in the y-dimection (m’/s/m) water depth (m) source magnitude per unit horizontal area (m’/s/m? source impulse in x- and y-dktions (m’/s/mz. m/s) evaporation rate (m/s) gravity Ws’> Chezy resistance No. (m%) wind friction factor wind speed and components in x- and y-diiections (mls) barometric pressure (lc Imls2) density of water ( k g h Corioliscoefficient(latitudedependent) (s-7 eddy or momentum dispersion coefficient (m’/s) space coordinates (m)

9,

t

time (s)

Calibration Factors

A MIKE 21 HD generated model has only three calibration factors, namely bed resistance C, wind friction factor f. and momentum dispersion mfficient E (normally called an eddy coefficient). Using these factors alone, calibration of a model is normally quite easy. In practice, the calibration of a model depends far more on the accuracy of the data, e.g. bathymetry, boundary data and wind speeds.

The bed resistance, C, and momentum dispersion coefficient, E, can both be specified as a function of space. Further, E can vary in t m in accordance with a special formulation of ie
Smagorinsky’s theory for turbulence.
MIKE 21 HD - Short Description
5

Solution Technique

The equations are solved by implicit finite differnce techniques with the variables defined on a space-staggered rectangular grid as shown below.

A 'fractioned-step' technique combined with an Alternating Direction Implicit (ADI) algorithm is used in the solution to avoid the necessity for iteration. Second order accuracy is ensured through the centering in time and space of a l l derivatives and coefficients. The AD1 algorithm implies that at each time step a solution is fist made in the x d m t i o n using the continuity and x-momentum equations followed by a similar solution in the y-direction. As a simple example, the finite difference scheme for the time derivative of flux density is shown below.

truncation error

The truncation error is of second order, O(Afi. The application of the implicit finite difference scheme results in a hidiagonal system of equations for each grid line in the model. The solubon is obtained by inverting the tridiagonal matrix using the Double Sweep algorithm, a very fast and accurate form of Gauss elimination.

6

MME 21 HD - Short Description

0

The implicit scheme is used in MME 21 in such a way that stability problems do not OCCUT provided, of course, that the input data is physically reasonable, so that the time step used in the computations is l m t d only by accuracy requirements, iie

Data Requirements

The necessary data can be divided into several groups as briefly described below. Basic Model Parameters: Model grid size and extent Time step and length of simulation Type of output required and its frequency Latitude and its onentation

....
I "

0

Bathymetry. Calibration Factors: Bed resistance Momentum dispersion coefficients Wind friction factor Initial Conditions: Water surface level Flux densities in x- and y-directions
Boundary Conditions: Water levels or flow magnitude Flow direction

Other Driving Forces: Wind speed and direction Sourdsink discharge magnitude and speed.

0

Pre- and Post-processing Software
The MIKE 21 HD module includes a range of pre- and postprocessing software which eases the input of data and analysis of simulation results. The software can be applied to the data and results of all the MIKE 21 modules. Some examples of the software capabilities are:

-

-

Input of time series 2nd 2-D data Isoline plots of any variable 2-D vector plots of current patterns Plots of the variation in space of a variable along any line through the model 3-D plots of bathymetries, surface levels, concentmtions Statistical analysis of time or space variation of any variable.
7

MIKE 21 HD

- Short Description

AU graphical presentations can be in colour, produced with the
UNIRAS graphics package.

List of References
H.I. Bundgaard, I.R. Warren, A. Barnett, "Modelling of Tsunami Generation and Run-up". EGS Congress, Copenhagen, April 1990. To appear in Science of Tsunami

Hazards.
Abbott, M.B., H. Shaoling, I.R. Warren, "Auto-dispersion processes in numerical models". Computer Modelling in Ocean Engineering, Venice, 1988,Ballcema, pp. 723-727. Madsen, P.A., M. Rugbjerg, I.R. Warren, "Subgrid modelling in depth integrated flows". Proc. 2 s Coastal Eng. Conf., 1t Malaga, 1988,pp. 505-511. Jensen, O., I.R. Warren, F. Nielsen, C. Laustrup, "Extreme water levels for design in the Baltic Sea Conference", September 1987. 8 pp.

.. Amdisen, L.K., J.B. Nielsen, I R Warren, "A prototype verification of a largeeddy simulation". Proc. 22nd Congress IAHR, 1987,Lausanne, B, pp. 53-58.
Warren, I.R., H.I. Bundgaard, "A Comparison between physical and numerical models of tsunamis". Proc. 22nd Congress IAHR, 1987,Lausanne, B, pp. 282-287. Abbott, M.B., J. Larsen, "Modelling circulations in depthlw. integrated f o s Part 2: A reconciliation". Journal of Hydraulic Research, Vol. 23, 1985. No. 5. Abbott, M.B., J. Larsen, J. Tao, "Modelling circulations in depth-integrated flows. Part 1 The accumulation of the : evidence". Journal of Hydraulic Research, Vol. 23. 1985,No. 4,pp. 309-326. Abbott, M.B., A. McCowan, I.R. Warren, "Numerical Modelling of Freesurface Flows that are TwwDimensional in Plan". T a s o t Models for Inland and Coastal Waters, rnpr Academic Press, 1981,pp. 222-283. Abbott, M.B., Computational Hydraulics, Elements of the Theory of Free Surface Flows, Pitman, London, U.K.,1979.

8

MIKE 2 HD Short Description 1

-

0

Abbott, M.B., C.H.Rasmussen, "Onthe numerical Modelling of Rapid Concentrations and Expansions in Models that are Two-Dimensioanl in P a " Proceedings 17th Congress IAHR, ln. 1977, Baden-Baden, 2, pp. 229-238. Abbott, M.B., A. Damsgaard, G.S. Rodenhuis, "System 21, Jupiter, A Design System for Two-Dimensional N a l ery H r z n a Flows", Journal of Hydraulic Resources, Vol. 11. oiotl 1973, pp. 1-28.

iI.

.*-

..

a

1991-01-21 . hl21HD.RW/CHC9011k i d

MIKE 21 HD - Short Description

9

10

MIKE21HD ShortDescnption

-

e

APPENDIX E
Short Description MIKE 21 NSW

Danish Hydraulic institute

MIKE 21
A Modelling System for Estuaries, Coastal Waters and Seas

Wave Modelling A Short Description

-

MIKE 21 Wave Mwlelling

a

INTRODUCTION
A total of five wave modules are included in MIKE 21, each with their particular area of application.
The individual'modules are: MIKE 21 OSW (Offshore Spectral WindWave Module) MIKE 21 NSW (Nearshore Spectral WindWave Module) MIKE 21 PMS (Parabolic Mild-Slope Wave Module) MIKE 21 EMS (Elliptic Mild-Slope Wave Module) MIKE 21 BW (Boussinesq Wave Module)

The models can be divided into two groups: models based on wave energy/wave action concept (OSW and NSW) models based on momentum concept (PMS, EMS, BW)

GENERAL OVERVIEW
A short description of applications, basic equations and solution techniques for each of the MIKE 21 wave Modules is presented below.
An overview of the computational ability of the five wave modules is indicated in the overleaf table.

Coastal Regions

Wave modelling in MIKE 21.

a

A Short Descripthn

1

MIKE 2 I Wave ModeiGng

0

2

A Short Desctrption

0

0

MIKE 2 7 Wave Modd6ng

MIKE 2 1 OS W OFFSHORE SPECTRAL WIND-WAVE MODULE
MIKE 21 OSW is a spectral wind-wave model describing the propagation, growth and decay of short-period and short-crested waves in offshore areas. The model comprises the effects of refraction and shoaling due to varying depth and wave generation due to wind. Furthermore the model includes the effect of interaction between waves with different frequencies. MIKE 21 OSW is a time-dependent discrete spectral wave model based on the energy balance equation. In the description of the spectral energy, an upper bound is applied representing a saturated sea state. The limiting spectral shape is a scaled Pierson-Moskowitz spectrum, where the scaling factor (Kitaigorodskii factor) depends on the water depth and the total energy. Hence, the effects of water depth, limiting the growth of waves and the shape of the spectrum during growth, are considered.

e
-

Application Areas A major application area is the design of offshore structures where accurate assessment of wave loads is of utmost importance to the safe and economic design of these structures. Measured data is often not available during periods long enough to allow for the establishment of sufficiently accurate estimates of extreme sea states. In this case, the measured data can then be supplemented with hindcast data through the simulation of wave conditions during historical storms using MIKE 21 OSW.
Another application area comprises studies of the wave conditions in coastal areas in connection with a nearshore wave model like the Nearshore Spectral Wind-Wave Module, MIKE 21 NSW. These wave models require wave conditions along the offshore model boundary. Based on information about the wind conditions, the boundary conditions can be determined by applying MIKE 21 OSW for the offshore area.

MIKE 21 OSW application. Hindcasted wave parameters, significant wave height, zero-crossing and mean

0

wave direction.

A S h r t Description

3
D a n I Hydraulirk lnnitm I Danish Hydraulic lmtitute

MIKE 21 Wawe Madehg

0

MIKE 21 OSW application in the Central Mediterranean. 7he upper chart depicts the hindcast windJeld as wind speed and direction. 7he lower chart shows the computed wavefield illustrated by the mean wave direction and signrpcant wave height.

0
4

A Short Descriptmn

MIKE 2 7 Wave Model6ng

Basic Equations MIKE 21 OSW describes the wave field by the directional-frequency wave energy spectrum.
The basic equation states that a component of the directional-frequency spectrum is moving by its own group velocity, being subjected to an increase or decrease in energy, depending on the bathymetry, wind speed/direction and spectral shape. This equation reads:

Solution Technique MIKE 21 OSW is basically a discrete spectral model, ie the energy is calculated in a number of discrete bins or mesh points of a rectangular Eulerian grid for a number of discrete frequencies and directions. However, a parametric model is used to describe the high frequency energy. This energy is 'fed' into the discrete model as the sea grows. The discrete model thus covers the main frequency range using the parametric model as a trigger function.

In MIKE 21 OSW,a semi-Lagrangian explicit higher-order difference scheme has been
The left-hand side of the basic equation takes into account the effects of refraction and shoaling, while the net source term on the right-hand side includes the effects of energy input from the wind as well as wave-wave interaction.
...... .. .
.,

formulated. The energy spectrum E (f,@ is calculated in mesh points of a fixed Eulerian grid using a Lagrangian approach.

.

.

. .

Input Basic input data to the MIKE 21 OSW Module is: bathymetric data wind fields boundary conditions The main task in preparing input for the Module is to select and prepare the (timevarying) wind fields. The wind field can be given as wind speed and direction or wind velocity components, and can be given either as constants for the entire computational domain or as a 2D map. If not directly available from eg a meteorological model, the wind fields can be determined by using the wind-generating programs in MIKE 21 PP, the pre- and postprocessing Module. Along all essential boundaries, the time and spatially varying directional-frequency wave energy spectrum should be specified.

:.

. .direction.ofivave propagation (deg) .: .group velocity (m/s)..

:.

.

...

MIKE 21 Wave M L n &6g

output
Two types of output can be obtained from MIKE 21 OSW: integral wave parameters - significant wave height, & - peak wave period, T, - mean wave period, To, - zero-crossing wave period, To2 - peak wave direction, 8, - mean wave direction, 0 , - directional standard deviation, u directional-frequency wave energy spectrum at selected grid points or areas

Hasselmann, K, Barnett, T P, Bows, E, Carlson, H, Cartwright, E, Engke, K, Ewing, J A, Gienapp, H, Hasselmann, D E, Kruseman, P, Meerburg, A, Muller, P, Olbers, D J, Richter, K, Sell, W, Walden, H (1973) Measurements of Wind-Wave Growth

and Swell Decay During the Joint Nonh Sea Wave Project (JONSWAP). Deut. Hydrogr. A., Suppl A, 8, No 12.
Hasselmann, K, Ross, D B, Muller, P & Sell, W (1976) A Paramem'c Wave Prediction Model. J. Physical Oceanography, Vol 6, pp 200-228. Kitaigorodskii, S A, Krasitskii, V P & Zaslavskii, M M (1975) On Philips' Theory of

References Bows, E, Giinther, H, Rosenthal, W & Vincent, C L (1985) Similariry of Wind Wave Spectrum in Finite Depth Water. Journal of Geophysical Research, 90 (Cl), pp 975-986.
Brink-Kjier, 0, Knudsen, J, Rodenhuis, G S & Rugberg, M (1984) Errreme Wave

Equilibrium Range in the Specna o Windf Generated Graviry Waves. Joumal Physical Oceanography, Vol 5, pp 410-420.
Resio, D T (1981) 7he Esrimarion of Wind-

Wave Generation in a Discrere Spectral Model. Journal Physical Oceanography, Vol 11, pp 510-525.

Conditions in the Central Nonh Sea. Proc. 1984 Offshore Technology Conference, Paper NO OTC 4809, pp 283-293.
Brink-Kjrer, 0, Kej, A, Cardone, V & hshparatnam, E (1986) Environmental

Conditions in the South China Sea mshore Malaysia: Hindcast Study Approach. Proc. 1986 Offshore Technology, Conference Paper No 5210.

a

6

A Short Description

a

Danrk Hydrduliik lnnituf /Danish Hydraulic l n a i l ~ t e

MIKE 2 7 Wave M d L n osGg

%
Application Areas MIKE 21 NSW can be applied to transform an offshore wave field to the nearshore area, typically in open coastal areas.
'

MIKE 21 NSW NEARSHORE SPECTRAL WINDWAVE MODULE
MIKE 21 NSW is a spectral wind-wave model, which describes the propagation, growth and decay of short-period waves in nearshore areas. The model includes the effects of refraction and shoaling due to varying depth, wave generation due to wind and energy dissipation due to bottom friction and wave breaking. The effects of current on these phenomena are included.

0

MIKE 21 NSW is a stationary. directionallv decoupled, parametric model: To include the effects of current, the basic equations in the model are derived from the conservation equation for the spectral wave action density. A parameterisation of the conservation equation in the frequency domain is performed introducing the zeroth and the first moment of the wave-action spectrum as dependent variables.

The assessment of the wave conditions in coastal areas (ie wave heights, wave periods and wave directions) is essential for the estimation of the wave forces at a shoreline. Another important problem in coastal engineering is the simulation of the sediment transport, which, to a great extent, is determined by the wave-induced littoral current. The wave-induced current is generated by the gradients in radiation stresses which occur in the surf zone. MIKE 21 NSW can be used to calculate the radiation stresses. By inclusion of the simulated radiation stresses in the basic hydrodynamic module, MIKE 21 HD, the wave-driven currents can be calculated and applied in sediment transport modules of MIKE 21.

MIKE 2 1 Wave ModelGng

Wave refraction, shoaling and breaking calculated by MIKE 21 NSW. lhe vectors indicate the significanr wave height by their length and the mean wave direction by their orientation.

d

2 meter

Oepths below DNN

-

8

0

1 ' 2

3

4

5

6

7

8

9 1 0 1 1

Fetch-limited wind-wave growth computed using MIKE 21 NSW.

d

1 meter

Depths (m)

5

10

15

20

25

30

8

A Short Dezeripthn

a

MIKE 2 7 Wave ModeLGm

a

Basic Equations The basic equations in MIKE 21 NSW are derived f o the conservation equation for the rm spectral wave action density based on the approach proposed by Holthuijsen et al. (1989). A parameterisation of this equation in the frequency domain is performed introducing the zeroth and first moment of the actjon spectrum as dependent variables. This leads to the following two coupled partial differential equations:

The description of the bottom dissipation is based on the quadratic friction law to represent bottom shear stress (Svendsen & Jonsson, 1980) and the description of the wave breaking is based on the expressions given by Battjes & Janssen (1978). The effects of the bottom dissipation and wave breaking on the mean frequency are also included.

From m&3) and m,(& two wave parameters can be calculated; the directional wave action spectrum &(e) = %(e) and the mean frequency per direction q ( 0 ) = ml(t9)/rn&%

The spectral moments

%(e)

are defined by:

where w is the absolute frequency and A is the spectral wave action density. The propagation speed cg and group velocities cgr and c, are obtained using linear wave theory. The left-hand side of the basic equations accounts for the effect of refraction and shoaling. The source terms So and SIaccount for the effect of energy input from the wind, bottom dissipation and wave breaking. The effects of current on these phenomena are included. In MIKE 21 NSW, the source terms for the local wind generation are derived directly from the Shore Protection Manual (1984) formulation for fetch-limited wave growth in deep water.

.

.
' : I '

:x and y

9 :.

.

. . . .. .
'

so, . ~.

s,

., .:.

.

: C r e i n coordinates .(m) atsa .': direction,of.wave. . .. .' propagation' (deg:).. .. :. source terms..(mz and
m*/s)
.
'

.
::

.., . .

. . .. ..

. .

..

. . . . .. . . . . . . . .

MIKE 2 1 Wave Modeffing

Solution Technique

Input

The spatial discretisation of the basic partial differential equations is performed using an Eulerian finite difference technique. The zeroth and first moment of the action spectrum are calculated on a rectangular grid for a number of discrete directions. In the xdirection, linear forward differencing are applied, while in both the y- and &directions it is possible to choose between linear upwinded differencing, central differencing and quadratic upwinded differencing. The best results are usually obtained using linear upwinded differencing both in the y- and 0directions.
The source terms due to the local wind generation are introduced explicitly, while the source terms due to bottom dissipation and wave breaking are introduced implicitly. Hence, a non-linear. iteration is performed at each grid point.

The following basic input data are required in MIKE 21 NSW: bathymetric data stationary wind field (optional) stationary current field (optional) bed friction coefficient map (optional) wave breaking parameters (optional) boundary conditions
The offshore boundary conditions can be given in the form of significant wave height,
mean wave period, mean wave direction and maximum deviation from this direction and finally the direction distribution of wave energy. The boundary can also be applied in the form of transfer data from a previous simulation.

Output
Two types of output can be obtained from MIKE 21 NSW: integral wave parameters in the computational domain:

The non-linear algebraic equation system
resulting from the spatial discretisation is solved using a once-through marching procedure in the x-direction (the predominant direction of wave propagation) restricting the angle between the direction of wave propagation and the x-axis to be less than 90 deg. In practice, this angle must be less than about 60 deg due to numerical stability considerations. MIKE 21 NSW can be applied for waves with wave periods between 0.21s and 21s.

-

significant wave height, , H - zero-crossing wave period, To2 - mean wave direction, 0, - directional standard deviation, (r 2D-maps of radiation stresses S,, , and S

0

s,.

MIKE 21 Wave Modelling

References Battjes, J A & Janssen, J P F M (1978) Energy Loss and Set-Up Due to Breaking of Random Waves. Roc. 16th Coastal Eng. Conf. Hamburg, pp 569-587.

CERC (1984) Shore Prorem'on Manual. U.S. Army Coastal Eng. Res. Center, Corps Engineers, Vol 1. Holthuijsen, L H, Booij, N & Herbers, T H C (1989) A Prediction Model for Stationary. Shon-crested Waves in Shallow Water wirh Ambient Currents. Coastal Engineering 13, pp 23-54. Svendsen, I A, Jonsson, I G (1980)
Hydrodynamics of Comral Regions. Den

Private Ingeninrfond, Lyngby .

MIKE 21 Wave Madehg

MIKE 2 1 PMS PARABOLIC MILD-SLOPE WAVE MODULE
MIKE 21 PMS is based on a parabolic approximation to the elliptic mild-slope equation governing the refraction, shoaling, diffraction and reflection of linear water waves propagating on gently sloping bathymetry. The parabolic approximation is obtained by assuming a principal wave direction (x-direction), neglecting diffraction along this direction and neglecting backscatter. In addition, improvements to the resulting equation, cf Kirby (1986),allow the use of the parabolic approximation for waves propagating at large angles to the assumed principal direction.
An additional feature of MIKE 21 PMS is the

-

Application Areas MIKE 21 PMS can be used to determine wave fields in open coastal areas, in coastal areas with structures where reflection and
diffraction along the x-direction are negligible, in navigation channels, etc. Furthermore. MIKE 21 PMS can produce the wave radiation stresses required for the simulation of wave-induced currents, which is very important in the computation of coastal sediment transport.

meter

ability to simulate directional and frequency spreading of the propagating waves by use of linear superposition.

MIKE 21 PMS can be applied to any water depth on a gently sloping bathymetry, and it is capable of reproducing phenomena. such as shoaling, refraction, dissipation due to bed friction and wave breaking, forward scattering and partial diffraction. MIKE PMS is extremely cost-efficient and requires a low computation effort. since the numerical solution is based on a single marching procedure from the offshore boundary to the coastline.
0

200

400

600

1 meter

MIKE 21 PMS used to calculate wave heights behind a detached breakwater.

12

A Short DescnWon

0

Danik Hydrauliik Inmlu /Danish Hydraut lnllltule

a

MIKE 21 Wave MadeEhg
rn

m

1800.

1800

1800

1800

1400

14M

1200

1200

Elm
800

0

"
400

zoo
0

m4008008001m

o m 4 M 6 0 0 8 0 0 1 m o

MIKE 21 PMS application to predict wave conditions in a groynefield, instantaneous surface elevations QeJ) and wave height and direction (right).
rn

rn

2500

2ooo

E

E 1500

\
0

*
Ls
1.0

.far

500

05M-03-1.0

-

-,a

-20 Bb-

0
0 500

bd

lmo

1500

2ooo

500

lam

1500

2ooo

Wave conditions at a harbour entrance calculated by MIKE 21 PMS. 7he bathymetry i depicted (left) s together with the instantaneous surface elevations (right).

a
A Short Descnpthn

13
Danik Hydraulisk lnnitut I Danish Hydraulic InnitUte

MIKE 21 Wave Madelhg

0

Basic Equations The parabolic mild-slope equation applied in MIKE 21 PMS is:

Generally, the (1,l) Pad6 approximation is accurate for wave propagation direction within +I- 45 deg around the principal wave direction. For larger aperture widths, the minimax approximation can be used, although noticeable errors are present at small angles to the principal direction when the aperture width is greater than 70".

where
I

For the parabolic approximation, three different techniques are implemented via the coefficients of the rational approximation PI, Pz, and &:

. . . . Symbol.:List :A .(x,y) : slowly varying. comple . . amplitude (m) C : phase velocity (m/s) ". . cg : group velocity ( m k ) k : wave number ( d ) . : average wave. number ,o.v.ec. k, . the-y-direction (m') . . .~ ..
'

simple approximation (also known as (1 ,O) Pad6 approximation) (PI = 1 , pz = 4 ,and P, = 0) (1,l) Pad6 approximation (Dl = 1, p2 = -3/4, and p, = -1/4) minimax approximation for different apertures (10, 20, ..., 90 deg) Each aperture width has a set of coefficients, cf Kirby (1986). The formulation of bed friction is based on the quadratic friction law. The description of the dissipation due to wave breaking is based on the expressions given by Battjes & Janssen (1978).

.

.

and:,wave breakirig':(s.'

Solution Technique The parabolic mild-slope equation in MIKE 21 PMS is solved using the Crank-Nicolson finite difference techniques with variables defined on a rectangular grid.

e
14 A Short Description

MIKE 2 1 Wave ModelEm

Input In MIKE 21 PMS, the following basic input data is required: bathymetry data bed friction data (optional) wave breaking parameters (optional) boundary conditions
For monochromatic unidirectional waves, the incoming wave conditions are specified by the wave height, wave period and wave direction. For irregular andlor directional waves, the incoming wave conditions are given by the directional-frequency wave energy spectrum, prepared using the MIKE 21 preprocessing program m2lspc.

References
Battjes, J A & Janssen, J P F M (1987) Energy Loss and Set-Up Due to Breaking o f Random Waves. h o c . 16th Coastal Eng. Conf. Hamburg, pp 569-587. Kirby, J T (1986) Rational Approximatio& in the Parabolic Equation Method for Water Waves. Coastal Engineering 10, pp 355-378.

0

output MIKE 21 PMS produces four main types of output: integral wave parameters - the significant wave height - the peak wave period - the mean wave direction (MWD)

0

2D map of instantaneous surface elevation 2D map of vector components H.cos (MWD)and Nsin (MWD) 2D map of radiation stresses

MIKE 21 Wave Modellim

MIKE 2 1 EMS ELLIPTIC MILD-SLOPE WAVE MODULE
The Elliptic Mild-Slope Wave Module, MIKE 21 EMS, simulates the propagation of linear time harmonic water waves on a gently sloping bathymetry with arbitrary water depth. MIKE 21 EMS is based on the numerical solution of the Elliptic Mild-Slope equation formulated by Berkhoff in 1972 and is capable of reproducing the combined effects of shoaling, refraction, diffraction and backscattering. Energy dissipation, due to wave breaking and bed friction, is included as well as partial reflection and transmission through for instance pier structures and breakwaters. Sponge layers are applied where full absorption of wave energy is required. In addition, the model includes a general formulation of radiation stresses, based on Copeland (1985) which is valid in crossing wave trains and in areas of strong diffraction.

Application Areas MIKE 21 EMS can be used to study wave dynamics in smaller coastal areas and in
harbours. The Module is particularly useful for the detection of harbour resonance and seiching due to for instance long-period swell. The Module can also be used to study short periodic wave disturbance in harbours, but since it operates with monochromatic and linear waves, MIKE 21 BW is generally recommended for this purpose. MIKE 21 EMS can alternatively be applied for the computation of radiation stresses in the surf zone of smaller coastal areas. By inclusion of the simulated radiation stresses in the MIKE 21 HD Module, the wave-driven currents can be calculated and applied in sediment transport modules of MIKE 21.

MIKE 21 EMS application to harbour resonance. The plot shows the computed relative wuve heights.

16

A Short Description

.,

.. , .. ,

MIKE 21 Wave M&ffing

e
m

.. ...

e
0

50

100

150

200

250

e

Harbour resonance studied using MIKE 21 EMS. l3e plots show the relative wave heights and depth-averaged panicle velocities for wave periods (a) 15 s and (5) 30 s.

1 l/S

0

50

100

150

200

:

I

MIKE 21 Wave Modeffim

Basic Equations
The Elliptic Mild-Slope equation valid for time-harmonic problems is formulated as:

By introducing the pseudo fluxes P and Q', ' this equation can be rewritten as a system of first-order equations, which are similar to the mass and momentum equations governing nearly horizontal flow in shallow water:

where L,
* fi
C

The harmonic time variation is extracted from the hyperbolic system of equations by using
t

P '

-

= S (x. y, I) e'"'

P (x,

y. I)

C'W

Q' =

0 (x. Y.

I) e'"'

Notice that the remaining time variation in S , P, and Q is a slow variation, which is due to the solution procedure (iteration towards a steady state).
This leads to the following set of equations which have been generalised to include internal generation of waves, sponge layer absorption, partial reflection from breakwaters, bed friction and wave breaking:

. i.

'

.

ss :: '.

::'imsginary..mW . . . I :source magnitudejper..unit. :

.

.

. .

'.

.

frictioni(s:.!), .. .: . ;., . . . .:. energy.dissipation: due'. e.. . . .. r&ing::(s;') . ." ..: .. . .: .Cartesi&.coordinates (m).' ... . . .. . . . .' . . .. ... . . :. time (s) : : .horizontal gradient operator
' '

nergy.:dissipation due..to:bed . .
,

. . ..

, ,

'

... . .

MIKE 21 Wave M d n o&g

For further details, reference is made to
Madsen and Larsen (1987).

output Three types of output can be obtained from MIKE 21 EMS: 2D map containing the wave height (relative or absolute) in selected areas
2D map of depth-averaged particle velocity components 2D maps of radiation stresses Vector plots of the depth-averaged particle velocities are very suitable for visual detection of nodes and antinodes in studies of harbour resonance.

Solution Technique The equations are solved by implicit finite difference techniques with the complex variables S, P and Q defined on a spacestaggered rectangular grid. An Alternating Direction Implicit (ADI) algorithm is applied together with an efficient alternating time step strategy to fmd the steady state solution. A constant time step may also be applied. The difference equations are solved by use of the double sweep algorithm, a 'very fast and accurate formof the Gauss elimination.

Input The basic input data to MIKE 21 EMS consists of: bathymetric data bed friction coefficients (optional) wave breaking parameters (optional) partial reflection from structures wave absorption (sponge layers) boundary conditions
The essential boundary conditions concern waves entering the computational area. The time harmonic waves are generated internally inside the model boundaries applying a source term in the mass equation based on Larsen & Dancy (1983). Only the incoming wave height and wave period is required.

References Berkhoff, 1 C W (1972) Computation of Combined Refram'on-Di@ktion. h o c . 13th Coastal Eng. Conf., Vancouver 1972, ASCE. New York, Vol 1, Chap 24, pp 471-490.
Copeland, G J M (1985a) A Pracrical Alternative to the Mild-Slope Wave Equation. Coastal Engineering, Vol 9, pp 125-149. Copeland, G J M (1985b) Pracrical Radiation Stress Calculations Connected with Equations of Wave Propagation. Coastal Engineering, VOI 9, pp 195-219. Larsen, 1 & Dancy, H (1983) Open Boundaries in Shon-Wave Simulations - a New Approach. Coastal Engineering 7, pp 285-297.
~~

Madsen, P A (1983) Wave Refectionfrom a Vertical Permeable Wave Absorber. Coastal Engineering 7, pp 381-396.

MIKE 2 1 Wave Madefiw

Madsen, P A & Larsen, J (1987) An Eflcient Finite-Direreerence Approach to the Mild-Slope Equarion. Coastal Engineering 11, pp 329351. Warren, I R, Larsen, J & Madsen, P A (1985) Application o Shon Wave Numerical f Models to Harbour Design and Future Development o the Model. Int. Conf. on f Numerical and Hydraulic Modelling of Ports and Harbours, Binnimgham, April.

MIKE 21 Wave MOdel*ng

MIKE 21 BW BOUSSINESQ WAVE MODULE
The MIKE 21 BW Module is based on the numerical solution of a new form of the twodimensional Boussinesq equations. The Boussinesq equations include non-linearity as well as frequency dispersion. Basically, the frequency dispersion is introduced in the flow equations by taking into account the effect that vertical accelerations have on the pressure distribution. The major restriction of the classical Boussinesq equations is their water depth limitation. The new form of the Boussinesq equations incorporates a significant improvement of the dispersion characteristics. The maximum depth to deep water wave length ratio, d/L,, is increased from 0.22 to 0.5. With these new equations, the MIKE 21 BW Module is suitable for the simulation of the propagation of wave trains travelling from deep water to shallow water. The model is capable of reproducing the combined effects of most wave phenomena of interest in coastal and harbour engineering. These include shoaling. refraction, diffraction and partial reflection of irregular short-crested and long'crested finite-amplitude waves propagating over complex bathymetries. Phenomena, such as wave grouping, generation of bound sub-harmonics and superharmonics and near-resonant triad interactions, can also be modelled using MIKE 21 BW. MIKE 21 BW includes porosity for the simulation of partial reflection from and transmission through for instance pier structures and breakwaters. Sponge layers are applied when full absorption of wave energy is required (eg for open sea boundaries). Finally, MIKE 21 BW also includes internal generation of uni-directional and directional waves.

Wave breaking, wave-current interaction and wave-ship interaction is presently under investigationldevelopment and is not yet included in MIKE 21 BW.
Application Areas

MIKE 21 BW can be applied to the study of wave dynamics in ports and harbours and in small coastal areas. Wave disturbance in harbours is important for ship berthing as well as for cargo loading and unloading. The disturbance inside harbour basins is one of the most important factors when engineers are to select construction sites and determine the optimum harbour layout. The transmission of waves from the sea into a harbour protected by breakwaters is a process which involves mainly shoaling, refraction, diffraction and reflection processes.
8

- Spectrd Density

0.0

0.1

0.2
1/s

0.3

0.4

0.5

2 ,

- Surface Elevation

1
I

I

I

1202

1204

1206

1208

1210

Frcquency spectrum and rime series of sutface elevarionsfrom a JONSWAP spectrum can be used as boundary data in MIKE 21 B W.

A Short Lkscriptmn

21

MIKE 21 Wave ModeIGna

Distwbonce Coeff.
O A h

l.m -

n

o.00 0.80 0.70 0.80 0.u

- )no - 0.90 - aao - 0.70
- a50

1.10 1.10

0.25 - CUO 0.20 - a25 0.15 - azo o.m - at5 0.M 0.10 Mi 0.05 t d

-

Wave disturbance coeficients obtainedfrom MIKE 21 B W using unidirectional irregular waves as boundary data.

Instantaneous suflace elevations obtainedfrom MIKE 21 BW using unidirectional irregular waves as boundary data.

Danrk Hydraulirk lnnnuf I Danish nydraulic InRitUte

ANNEX 1
Database User Guide

RED SEA HINDCAST STUDY
DATABASE USER GUIDE
1
INTRODUCTION
The Red Sea hindcast data are accessed through a user-friendly windows program called Map. This user guide describes how to install and operate this system.
2

SYSTEM REQUIREMENTS
The system requires a PC with the following minimum configuration :
e e
e

e

e

Windows 95 or Windows NT. 80486 CPU. 8 megabytes of memory. 50 megabytes free harddisk space for a hard disk installation or 1 megabyte for a CD-ROM installation (see below). Graphical resolution of 800 x 600,256 colours. CD-ROM drive.

3

INSTALLATION
The system and the associated data are distributed on a CD-ROM.

You can select between two types of installations, a hard disk installation and a CDROM installation.
The hard disk installation is the normal installation, where the entire system is copied onto the harddisk. This type of installation requires aproximately 50 MB of free harddisk space. If the CD-ROM installation is used, the data and programs are not copied onto the harddisk, but left on the CD-ROM. The harddisk is only used for storing setup and temporary data. This type of installation requires approximately 500 KB of free harddisk space.

To install, do the following :
1.

Insert the CD ROM in your drive. Note the drive letter, e.g. D:. Execute the setup program : D:\Bin\Setup.exe. You will be asked to enter the following information :

2.

Red Sea Hindcast Study

1

Database User Guide

Source Directory : The directory in which the system resides on the CD-ROM, typically D:\Redsea. Target Directory : The directory where you want to install the system, typically C:\Redsea. Installation Type : Select either Hard Disk or CD-ROM installation, see above. Disable Copying of Files : Check this field to disable copying of files to your hard disk. You can use this if you copied or moved the files yourself and only want to change settings in the registry.

3.

0
4.

Using the normal Windows 95 utilities, create an icon (shortcut) for the Map.exe program in the bin directory under the installation directory. You can also add a the program to the Windows 95 program menu.
The system can now be executed by double clicking on the Map icon.

See chapter 6 for information on how to un-install the system. Please ensure that you are using small fonts in your windows display setup. This setting can be changed in the windows control panel, display settings. Please note that in the file pltdev.dat in the menu directory, the screen size must not be less than 210 x 155.If this is the case, please edit the file to correct. See section 7.4 for further information.
4

OPERATION
Startup The system is started by double clicking on the Map icon on your desktop, or by selecting the appropriate startup menu button.

0

4.1

4.2

Basic Operation When Map has been started, you typically do the following :
0

Type the password in the dialog that appears and press enter or click on the OK bunon. A window is now opened displaying the entire Red Sea. Select the desired area by pressing one of the buttons in the Area menu.
2 Database User Guide

0

Red Sea Hindcast Study

0

0

Select a point by moving the mouse pointer to the desired location and clicking with the left mouse button. The UTM coordinates of the mouse pointer is continuously shown in the second pane of the status bar and the UTM coordinates of the selected point are shown in the third pane of the status bar. Select one of the buttons in the data menu for displaying the desired information. The data are presented in tabular or graphical form.

The following sections further explain the function of the various parts of the systems.
4.3

MenuBar
File Open Opens a file for display. Files with extension ".ct2" can be opened. Change password. A dialog appears where the current and the new password can be typed. This is a list of the 4 last files that have been displayed. Exit the program. Load a map showing the entire Red Sea. Display the Jizan map. Display the Rabigh map. Display the Jeddah map. Display the Yanbu map. Display the Duba map. Show grid in UTM coordinates (default). Show grid in geographical coordinates. Show grid in model coordinates. No grid is shown. -. Set the UTM zone. A dialog is opened where a new UTM zone can be typed. Show or hide the toolbar. Show or hide the status bar. Turn on or off hardcopy of plots. When this option is turned on, a plot is sent to the printer instead of the screen. This option is automatically turned off every time a plot is sent to the printer. See also the section on
3
Database User Guide

-______--Password Recent Documents

_-----____
Exit

Area Red Sea Jizan Rabigh Jeddah Y anbu Duba Grid UTM GEO Model None Set UTM Zone View Toolbar Status Bar Plot Hardcopy

0

Red Sea Hindcast Study

printing and exporting data. Zoom In Zoom Out Previous Zoom Refresh Select Point Zoom in on a sub area by pressing the left mouse button and dragging the mouse. Zoom out, so the entire map is shown. Displays the previously zoomed area. Redraw the map. Opens a dialog, where a point can be selected by typing the UTM coordinates of the point. Show contour plot instead of box plot. Show overview window. A small window is opened showing the entire map. When the user zooms, the zoomed area is shown framed in the overview window. You can move the frame and resize it to change the zooming. Show the color scale. Change the font used for information on the map. The function of this menu is discussed in chapter 5 below. Shows the version number for Map.

---_______
Contours Overview

Color Scale Font Data Help About Map
4.4

Tool Bar
The tool bar is the area under the menu bar. It contains various graphical buttons that act as a shortcut to some of the buttons in the menu bar. The tool bar contains the following buttons : Button Open File : Help : Hardcopy: Zoom In : Zoom Out : Overview : Shortcut File -> Open. Help -> About Map. View -> Hardcopy View -> Zoom In. View -> Zoom Out. View -> Overview.

4.5

Status Bar

The status bar is the lowermost part of the window. It contains 6 fields where various status information is shown. The status bar contains the following fields :

Status Pane :
Pointer Tracking : Selected Point :

This shows the immediate status of the program and also displays help texts for the buttons in the menu bar. Shows the UTM coordinates of the pointer. Shows the UTM coordinates of the selected point.
4

0

Red Sea Hindcast Study

Database User Guide

0
4.6

Caps Lock : Num Lock : Scroll Lock : Selection of Point

Indicates if the Caps Lock key is activated. Indicates if the Num Lock key is activated. Indicates if the Scroll Lock key is activated.

Most of the queries into the database require that a point is selected. This is done moving the mouse pointer to the wanted point and clicking with the left mouse button. A red cross is drawn to indicate the point and the UTM coordinates are displayed in the third pane of the status bar. A point can also be selected by manually typing the UTM coordinates, see the section on the pop-up menu.

4.7

Pop-Up Menu
The pop-up menu appears when the right mouse button is pressed and held while the pointer is within the map. The menu contains the following buttons : Zoom In Zoom Out Previous Zoom Refresh Select Point Shortcut to View -> Zoom In. Shortcut to View -> Zoom Out. Shortcut to View -> Previous Zoom. Shortcut to View -> Refresh. Shortcut to View -> Select Point.

0

Red Sea Hindcast Study

5

Database User Guide

0

5 5.1

DATA PRESENTATION Introduction.
This chapter discusses the menus under the Data button in the menu bar and gives examples of the tables presented when each button is pressed. Some of the tables may contain empty fields, which indicates that the corresponding value is exactly zero. The notation O+ is used to indicate a value that due to the precision would be shown as 0.0. In most cases numbers are shown with one decimal, therefore values below 0.05 are shown as O+.

5.2

Winds
The menu accesses normal and extreme wind data

5.2.1

Normal
The Directional Dist. of Ws button displays the following tables :
Wind S p e d Directional DisLribution *rea : REDSEA Position : Reference Point E II5.48E 26.00Nl Biiqht : 10 In) Durntlon : 10 nln. Iweragc
I Wind Speed I Ids)
I

-

.......................................... ..........................................
Direction ICOming from. relative to north,
30 .......60 ....1.0 .....0.......2.0 ....3 3 0 . Omni ............2 .....8..210 ..0..7 ...... . . 90 . 150 1 .. 2 4 . 300 .. $
( I

...... ......

I 22.0 I 20.0 I 18.0

I

o+ ........ ................................. ....... ..................................
0.2 0.2
Ot

I I I I I I I I

16.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0 1.1 -

24.0 , 22.0 I 20.0 I
18.0 I 16.0 I

I I I 0.1 I
0.1 0. 0.1 0.5 0.
Ot

0.1

I 12.0 I 10.0 I B.0 I
I..O
6.0 4.0 2.0

0.
Ot

Ot

o*
0.1

01 .
0.1
O+ 0.3

0.1
0.1
0.4

O+

1.0

0.2
0.4

0.1
0.1

01 .
0.7 0.3

I

I I

2.0 2.1 0.2

0.1 0.3

0.1
0.4

0.4
1.7

0.5

0.5

0.3 0.1

0.2

o+

0.1

0.2

0.3

0.8

o+

o+

0.8 I 0.9 2.6 I 2.9 0.9 8 . 4 I 9.5 2.8 16.8 I 20.5 6 . 4 21.1 I 29.6 6 . 9 12.8 I 26.0 1 . 7 2.6 I 9.4 0 . 0.2 I 0.9

0.3

calm :

0.0

0

Wind Sooed Directional Lxcssdancs Area : REDSEA Position : Reference Point E II5.4BE 26.OONl Helght : 10 In) hlrat1.n : 10 mi". average

-

.......................................... ..........................................
I Wind Speed I Imlal I > I > I > I >
22.0
20.0 18.0

I_____________(

I Direction IComIng tram. relative t o north1 I 0 30 60 90 120 150 1 8 0 210 240

270 100 330 .................................. I .................................. I
I I 0.11
0.9

(*ani I

I

I

I
0.1

I I I

16.0
01
0.1 0.7 1.7

I > 14.0 I > 12.0 I > 10.0 I > 8.0

o*
0.

0.4

Ot
D. 0.1

01

01

1.1

I > I > I > I >

0.1
O*

6.0

0.1
0.7

4.0 2.0

3.7
5.8 6.0

0.2
0.7 0.9

1.1

1.2 1.3

0.1 0.4
0.5

0.3

0.3 0.9

0.1 0.6
0.1

0.1 0.2
0.5 0.7
0.7

0.1
0.2 0.6
0.9 0.9

0.4 0.5

1.2 1.2

0.1

0 1 4.1 . 0.5 10.5 2.1 1 7 . 4 2 . 9 1 9 . 1 65.2 I 9 9 . 1 I 2.9 19.1 65.8 I 1 0 0 . 0 I

I 0.11 I 1.0 I 1.5 I 3.9 I 11.9 I 1 3 . 4 I 28.1 I 33.9 I 4 9 . 8 I 63.6 I 62.6 I 8 9 . 6 I

0

Red Sea Hindcast Study

6

Database User Guide

The Seasonal Dist. of Ws button displays the following tables :
Ylnd Speed
-ea
: REDS=

-

Seasonal DiILIibYLion

Po8ILIon : Reference Point E 05.48E 26.0ONJ "Clpnt : 10 I W Duration : 10 m i " . A"vez.,e wind Speed
~~

. .

I Period
RpPJ""

Irn/SI ......I J a n - m r ......
.......................I

Jul-ssp oct-Dec

22.0 20.0
18.0

16.0
14.0

12.0 10.0

. .

8.0 6.0 ,.o 2.0 1.1 -

-

... ..

24.0 22.0 20.0
18.0 16.0

0.2

14.0 12.0

1.0 3.7
9.9 19.0

01 . 1.8 3.5
9.8

0.3 2.1
9.0

10.0
8.0 6.0 4.0

27.0
27.2 11.1
(1.9

20.9 28.7 24.5
9.6

2.0

1.1

21.2 32.6 26.6 7.6 0.5

0.5 2.5 9.3 20.8 30.2 25.8 9.5 1.3

wind Spasd Seasonal Excecdance %re* : REDSEA Position : Reference Point E 135.48E 26.00NI Height : 10 Irnl DYI.ti0" : 1 0 mi". Werage

-

......................... ........................
I Wind Speed I ln/rl

1 I > 22.0 I > 20.0
I > 18.0 I > 16.0

I Period I Jan-nar Rpr-J"" J1SP "-.
I I I

ocr-rmc

.I

I I

I
I I

I > 14.0 I > 12.0 I > 10.0 I > 8.0 I > 6.0 I > 4.0 I 2.0 I > 1.1 I Calm

I I
I I
I

0.2 1.2
4.9
14.e

0.1 1.9
5.4

0.3
2.1

I 0.5 I
3.0

I

15.2
36.1 64.8

11.4

I
I

33.8 60.7
88.0

32.6 65.3
'11.9

,

89.3
98.9 100.0

I
I
I

99.1
100.0

99.5
100.0

12.3 I 33.1 I 63.31 89.2 I 98.7 I 100.0 I
I

The Monthly Dist. of Ws button displays the following tables :
Wind W e e d Mea : Posirion : Heiqht : rmr.cion :

REDSEA

- Month Distribution
I mnlh I

.......................................... ..........................................
I wind speed
JUl aw sep OFf ,..... .,.Jan..............................., ...... .I... ...."ax .Rpr .May ................... Dec .. Feb .. .. ..

Reference Point E 135.48E 26.0ONI 10 lm) 10 nin. Rvaraae

I Irn/Sl

JY"

"0"

I An".

I

0

.......................................... ..........................................
Wind Speed

.. ..

Area

: REDSU

- Month

Excecdance

PoIiClOn : Reference Point E 1 3 5 . 4 8 1 26.00NI Height : 10 l m l DuraLlOn : 10 .n i. A""crsge

................................................. ................................................
I Wind 1 1d.l
Speed

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......... .......... .........
I > 22.0 I

I Month I Jan I
0.6

Feb

m i

Rpr

Hay

JY"

JUl

A"q

. . . . . . .Nov. . . I . . . .OCL. . . . . . .Sep. . . . . . .Dec . . .. .
I I I

I > 20.0

I > lS.0 I > 16.0 > 14.0 I > 12.0 I 10.0 I > 8.0

,
I

,

4.5
14.7

0.3 1.2 6.2
11.4

0.3 1.5 4.0
12.4

0.3 3.8

8.2
19.1

1.9 6.3 i5.7
40.6 70.0

0.1 1.7
10.9

I
0.4

1vm. I I I I I 01 I .
1.01

0.3

0.8

0.1

0 . 7 ,

2.3
12.0

3.4
11.6

1.5
1.5

3.0
14.6

1.7
8.6

I I > I >

, 6.0
4.0

I >

cam

2.0 1.1

41.4 26.9 73.3 56.5 94.5 87.2 85.8 I 8 9 . 6 9B.I 9 8 . 0 99.3 99.7 98.7 99.7 99.2 91.2 I 99.1 100.0 100.0 1 0 0 . 0 100.0 100.0 1 0 0 . 0 100.0 100.0 lOO.0 100.0 100.0 100.0 I 100.0 I I
36.8 62.8
61.5
64.1

33.5 58.8 86.4 99.3

33.8 61.0 89.2 99.6

34.0 62.4 88.4

30.7

32.1

38.5

86.6

91.8

89.5

93.8 99.4 100.0

71.3 92.8

27.2 60.4 89.0

4.31 i3.6 i 30.8 I 60.1 I

3.91 I 33.9 I 63.6 1
13.4

I

I
I

1

0

Red Sea Hindcast Study

7

Database User Guide

0

5.3

Currents
This button gives access to normal and extreme current data.

5.3.1 Normal
A dialog is shown where the start and end dates of the period requested can be given.

Press the enter key or click on the OK button and a plot will be displayed. This plot shows the bottom, depth averaged and surface current speeds as well as direction. Please note that this may take a some time.

5.3.2 Extreme
The Extreme button displays the following table :

0

Are*

A.c.e Soir0

Extreme Current Spasd and IUsociafcd Direction : RE s z Ou Dlrectlon : 330 1Deg.i Poairion : Reference Point F I,S.IBE

26.00")

..I ..

I Recurrence I CYrIenr Speed Irnfll I Period iyearil I Surface Depth-Aueraqed

I

..,........., ........... ....... . ... ... .

On-Boffom
0.0 0.0

I
I I

I 1

I 10 I 50
I 100

I 0.3 I 0.3 I 0.3 I 0.3
A

0.1

0.1 0.1 0.1

0.0
0.0

I
I

Warning :

secondary f l o w direction may exist approximately

180

deqreci from direcrion shorn.

5.4

Water Level
This button gives access to normal and extreme water level data.

5.4.1 Normal
The Tide Level Plot button opens a dialog identical to the one used for normal currents. When this dialog is closed, a plot is displayed showing the tide level. This may take some time. The Reference Levels button displays the following table :
Normal water Levois Area : REOSU Posit1.n : Reference P o i n t E 135.48E 26.00NI

....................
I W T

: I m s : I wm : I L ( m: I "JL :
I "LYN I WL"

0 . 3 1m1 I 0.3 (ID, I
0.2

(4I
Irnl

I mns :

: :

I U T i ....... ....... ......

I im1 I 1m1 I 1m1 I -0.3 mi I - 0 . 4 in1 I
01 .
0.0 -0.1 -0.2

0

Red Sea Hindcast Study

9

Database User Guide

Extreme

The Surge vs. . . eturn Period button displays the following table : R
Excreme water Leuels Area : REDSEA ~~~~t~~~ : ~~feience point L 1 3 5 . 4 8 ~2 6 . n n ~ 1

5.5

Waves

Ths button gives access to normal, extreme and fatigue wave data.
5.5.1

Normal
The Directional Dist. of Hs button displays the following tables :
Area
I I I
I I
"S

S i p i f i s a n t Wave Height - Directional Distribution : REDSEA Position : Reference Point E 135.48E 2 6 . 0 0 N I
Irnl

_.........._.._.....~~~~~~~~~~~~~~~~~~~.........~~~~.-.-..----------------------.
1......~...~~~~~~~~..........------------....._....~~~~~~~~......~~...------------. I %. 5 , .. - L . . -. O ,
5.0 4.5 4.0

I Direction [waves coning from. islarive t o north1 I 0 30 60 90 1 2 0 150 180 210 2 4 0 2 7 0

300

330 I m

i

I
I I I I I 1

3.5
3.0
2.5 2.0 1.5

,
I

1.0 0.5 0.1

-

5.5 I 5.0 I

I I 3.5 I
4.5 4.0

I I
I

3.0 I
2.5 I 2.0 I
0.

rlln

-

1.5 I 1.0 I 0.5 I

0.

o*
0.3
0.2

o+
0.1 0.1
0.1 0.1
0.1

0.2
0.9 0.5

0.2

0.1 0.7 0.5

0. 0.1 0.4

o+
01 0.2

0.4

0.1

0. 0.1 0.2 0.1

0.1

0. 1.6 I 0.2 1.1 I 0.411.21

0.11 0.5 I

0.1
0.5 ,.6

0.2
0.6 0.5

2.3 26.7 I 5 . 7 11.3 I 3.6 1 . 7 I
I

4.6 11.8 29.6 40.6 11.1

0.1

Height
:rice

-

Direcrianal Excecdance

Point E 135.IBE 26.00N1
300 330 I

lirecrion IYaYes coming from. relative t o north1 0 30 60 90 120 150 180 210 2 4 0 2 7 0

m i

01 .
0.6

0.1

Ot
0. 0.1 0.2 1.2

2.2
6.6

0.6 2.2
6.8 18.6

0 .

o*
0.3 0.5

0.1
0.1

o+

o*
0.1
0.4

0.2
(. I

Of

0.1

0.1
0.3
0.4

1.7

0.3

0.1 0.1

0.1

0.8

0.3

1.4

0.9

0.1 0.3 0.4

0.3
0.8

0 . 6 17.7 2.9 44.a 8.6 75.7

18.2
88.8 99.9

1.112.180.4

0.1

Red Sea Hindcast Study

IO

Database User Guide

The Seasonal Dist. of Hs button displays the following tables :
S i g n i f ~ c a n tWave Height Ssaronal DlstrlbUtlon Area : REDSEA Position : R e f e r e n d Polnl E 135.48E 26.00Nl

. ....................... .......................
HI

-

<mL)
6.0

1 Period

.........I ...... ..... .. . . . . ./

Jan-Har Apr-Jun Jul-Jsp

..... ..... .....

5.0 1.5 4.0 1.5 3.0 2.5 2.0 1.5 1.0 0.5 0.1 Calm
5.5

.. .

OCL-DCC

I 5.5 I 5.0 I
1.5 I 4.0 I

3.5 I 3.0 I 2.5 I 2.0 I
1.5 I 1.0 I

0.2 0.8 2.1
5.6

0.3 1.1

2.9 5.3
11.0

11.5
27.7

39.3
12.7

0.5

I I

30.3 37.8 10.9
" 7

0.4 3.2 12.5 11.3 11.2 11.5

1.0

4.2 12.2
29.1

44.2
9..

.-

--

01 .

......... .........

Significant Wave Height Scasonal Exceedance Area : REDSEA Poaition : Reference Point E I35.48E 26.00Nl

-

........................ ........................
I
I
"S

Id
5.5 5.0 4.5
4.0

I Period

...........I ........... ) . I > I > I >
I >

I Jsn-nar Apr-J"" JYI-slp O C L - D e C
I I I

I > I > I >
I >

........................ ........................

I > I > I > I > I

1.5 3.0 2.5 2.0 1.5
1.0 0.5

I I
I I

0.2 1.0

0.3
1.5 4.4 9.7
0.4

3.1

10 .
5.2

I I I
I

8.7 20.2
47.9 87.2

20.7
51.0
88.8

3.5 16.1
47.4 88.5

11.4
46.4 90.6

01 .

I
I

99.9

calm

0.1

99.7 0.3

100.0

100.0

The Monthly Dist. of Hs button displays the following tables :
Significant nave Height - Month Dlsfrlbution Area : REDSEA Posltlon : Reference P ~ i n tE (35.18E 26.00Nl
I " 3

........................................... ..........................................
I I ...... ...... I 5.5 6.0
I
Irnl

Month

Jan

Feb

Mar

am

way

Jun

Jul

A w

rep

oct

No"

Dec I Ann.

I I

I
I I I

I I
I I

I
I I

0

....... ......
Slgnificanf w Area : RE

4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 01 . Calm
5.0 4.5

5.5 5.0 4.5

I

I
I

I

I
0.3
0.3
0.8 2.2

I
0.1 I 0.5 I

4.0

1.0
3.9 5,4 9.7 2 0 . 6 32.5 31.6 4 0 . 1 15.1 7 . 3 0.8
12.0

I

3.5 3.0 2.5
2.0

1.5 1.0
0.5

8.1 3.0 1 1 . 6 11.5 1 1 . 5 2 6 . 9 21.4 3 4 . 3 3 5 . 9 4 4 . 0 38.3 1 8 . 1 10.2 9 . 7 0.1 0 . 1

0.9 0.8 5.7

0.6

3.4

2.4 4.7 8.8

1.1

0 . 1 0.4 0.6 0.9 0 . 6 2.4 5.8 1.3 5.1 3.2 1 3 . 4 16.1 8.0 15.5 7 . 4 3 7 . 7 3 4 . 5 3 2 . 4 2 6 . 9 37.3 2 5 . 2 3 8 . 8 38.1 35.9 1 9 . 8 38.5 5 0 . 9 4 3 . 5 10.5 11.5 9.4 13.5 2 . 7 12.7 12.8

1.7

11.4

I 1.5 I 4.2 I 13.5 I 24.6 I
I

1.6 I
4.6 I 11.8 I 29.6 I 40.6 I 11.1 I
0.1 I

L
I

Height
\

- Honlh
Fcb

Exseedance

....... ......
I HI 1.1

POllltiO"

: RL

~ n c ePoint E 135.48E 26.00Nl

I
[

.

......................................... ........................................
(onth
Jan

...... ......
5.5 5.0
4.5 4.0 3.5

Mar

M r

Hay

Jun

JuI

A w

Sen

OCL

Nov

Dec I
I I I I

&.

I
I

I >

I > I > I > I > I > I > I > I > I >
I > I >
I

I
I I I
0.1

0.3
0.9 1.0 7.4

0.3
1.1
3.2

1.0

I
1.1

3.0 2.5
2.0
1.5

0.9

3.4

4.3
12.7 24.2

6.2
11.7

e.,

19.0
45.9 81.8 99.9

1.0 0.5

45.6

0.1

89.7 9 9 . 9 100.0

52.0 90.3

....... ......

calm

0.1

01 .

1.7 6.2 1.8 6.1 3.8 5.71 13.1 19.2 I 22.3 9.8 21.6 11.2 5 2 . 6 50.8 54.7 36.7 5 8 . 8 36.4 $3.7 I 92.7 89.5 88.5 9 0 . 5 8 6 . 5 97.3 8 7 . 3 8 7 . 2 I 84.1 99.2 1 0 0 . 0 100.0 100.0 100.0 1 0 0 . 0 100.0 100.0 100.0 I 0.8 I 16.9 28.9

5.0 10.4 20.1

49.5

0.1 2.6 15.9 50.5

0.4

0.6

0.9

0.6

I 1.51

0.6 I
2.21

I

6.81
18.6 I 48.2 I
B8.B

I

99.9 0.1

I
I

0

Red Sea Hindcast Study

11

Database User Guide

The Hs vs. Most Probable Period button displays the following'two tables :
Ilea" Upcrossing Period Area : KEOSU

Poalfion

: ReferenCC~POint E 135.4BE 26.00Nl

.

1

I
I I
5.2 5.0
4.8 6.2

I I

I
I 6.1 I 5.8 I
5.5

..._________________~~~~~~

3.6 3.5

4.8 4.0 3.6

4.8 4.3 3.6
3.5

3.5

..l 1.2 3.6 3.3

4.7 4.2

6.1 I 5.8 I 5.5 I 5.1 I 4.7 I
4.3 I 3.7 I 3.4 I

I

5.1 I 4.7 I 4.2 I

3.7 3.3

3.7 I 3.4 I

The Wave Rose button displays a wave rose for the normal wave data. Note that the wave rose plots are given only for the offshore boundary in the local areas.
5.5.2

Extreme
The Significant Wave Height (Hs) button displays the following table :
Significant wave Height l H * l Are. : KEDSU Poalilon : Refezence Point E 135.4BE 26.00NI

r

Red Sea Hindcast Study

12

Database User Guide

The Peak Wave Period (Tp) button displays the following table
Peak Y D V e Period IT*) Area : DUBA Fosltlon : 7 6 4 O i O E . '30206538

............................................ ...........................................
I XCCYIIenCe I rlirsction I w a Y e l coming fro., m 1 a t i w t o north1 I Period (Years1 I 0 30 60 90 1 2 0 150 1 8 0 210 240 270
1
10

I

...........I................/ I

300 3 3 0 .............................I ............................

1
I

50

I 100 ..........I................5.6 .......7.2........1 ..............2.0............ ........ ........ . 1.5 .. 6.6 .. 8.1 10.1 9

I I I

1.9
2.2

1.6
1.8

1.3
1.1
1.5

2.3 2.4

1.9

1.4 1.6 1.1 1.8

1.5
1.7 1.8 1.8

4.6
5.2

5.4
6.1 6.5

5.8
6.5 6.9 1.1

5.9
6.6
7.0

6.6
7.5
1.9

8.2 9.3
9.9

5.5

hvli I 1 1 . 9 I 8.2 I 8 . 9 I 9.3 I 9.5 I 9.9 I
I 10.1 I

The Maximum Wave Height (Hmax) button displays the following table :
m"e Height I m x , Area : DUBA Posltlon : 764010E, 302065%
HaXxmm

............................................ ...........................................
I Directlo" i m v e s coning from. r 1 t v e.i. t o north1 I 0 30 60 PO 120 150 180 210 240 270
300

I ResYrrence I Period I Y e a i a l ............................................I................ 1 I 1 I 10 I 50

330 I m
1.1 I

I. .

i I
I

I

I
I I

0.4
0.5

0.3
0.4

0.2
0.2

0.2
0.2

0.2
0.1 0.2

0.6

0.4

I 100 0.6 0.2 2.6 ..........I.............0.2..........3.2 ........ ..............0.4......... ........... ....... . .. .. 4 . 3

0.3 0.3

0.2

1.2 1.5 1.7 1.8

1.8 2.2 2.5

2.0 2.5 2.8 3.0

2.2 2.7
3.1

2.9 3.6
4.1

4.4 5.6 6.3 6.6

4.4

5.2 I 5.9 I 6.2 I

5.6 I
6.3 I

6.6 I

The Most Probable Associated Period (Tmax) button displays the following table :
Host Probable assOcIafed pericd I m x l and Range -ell : WnA P o s i r i ~ n : 174417E, 30390038
I RecYrrence I Dlr.cfAo" I 0 1 Period
I I 1 1
I Y * Y e I cominq

......................................... .........................................
from, ralsflvs
1.8
LO

north1
l.B

.____________I_________

30 50 90 150 ...................120..........180 .................... .......... ..

10
50

I I I

2. . 2.2-2.7 2.7
2.4-3.0

I
I I

I
I

I 100

I I

I I 2.7-3.3 .....................2.4-2.9 .........2 ......... ........................ ......... 0 - ........ .... 2 . 0 - 2 5 ..2 . 5 3 . 2 - 4 . 0 ................................ ................................

2.9 2.6-3.2 3.0 2.7-3.3

2.4 2.2-2.7 2.7 2.4-3.0 2.9
2.6-3.2

2.1
1.9-2.4

2.9

5.1
4.5-5.6 5.7 5.1-6.2 5.1 5.5-6.1

2.4 2.2-2.6
2.6

1.7-2.0 2.0 1.8-2.1 2.2
2.0-2.4

2.3-2.8
2.5

3.0

2.3

1.7-2.0 2.0 1.8-2.3 2.2 2.0-2.4 2.3

2.6-3.2 3.3
2.9-3.6 3.5

3.2-3.9 3.6

6.2
5.5-6.9

210 270 300 330 I m i .......240....................... ........ ....................... ..

5.7
5.1-6.2 6.3 5.1-6.9

5.8
5.2-6.4

6.2
5.6-6.8

6.8
6.2-7.5

3.9
3.5-4.3
4.4 3.9-4.8

I
I

I I I

6.5
5.8-7.1

6.9
6.2-7.6

7.6
6.9-8.4

I
1

6.8 6.2-7.5 7.6 6.9-8.4

I
I I I I I

I .......6 ......................... ........ 4 - 7 . 8 ....................7.6-9.3 .... 6 . 9 - 8 . 4 .

6.8 6.1-1.5 7.0 6.3-7.6

6.9 6.2-7.6 7.1

7.1 6.7-8.2
1.6

8.2
7.4-3.0

8.1 7.6-9.3

4.7 4.2-5.2 4.8 1.3-5.1

I
I

8.2
7.4-9.0 8.4

I

I I

The Maximum Crest Elevation (Emax) button displays the following table :
H a x i m u crest E e . ' n 1vrO Above MSL Area :m Position : 164010E. 3020653N
I RecYrle"ce
I b X ,

............................................ ...........................................
I Period l Y e a r a l I
I I I I D'rocrlo" i*ave* 0 30 60

.................................................I................ I I 1 I 10

coming from, I.<l.fi". t o north1 90 120 150 180 210 240 210

300
2.1 3.5
3.9
4.1

I

0.2
0.3

0.2
0.2

0.1
0.1

0.1
0.1

I

50

I 100

0.4 0.4

0.2
0.3

0.2
0.2

0.1
0.1

0.1 0.1 0.1
0.1

0.7

0.9 1.1 1.1

1.1 1.4

1.2
1.6 1.8

1.3
1.7 1.9

1.8 2.3
2.6 2.7

1.6
1.6

1.8

2.0

I 3 3 0 I hvli I I 2.6 I 2.7 I 3.2 I 3.5 I 3.7 I 3.9 I 3.8 I 4.1

,

The Animations of Storms button displays a dialog where a storm can be selected. An animation of Hs is then shown. The animation is repeated in an endless loop. Press ESC to stop the animation. The Storm Plots button displays the storm selection dialog and then displays a plot showing the Hs, Tp and the mean wave direction.

0

Red Sea Hindcast Study

13

Database User Guide

0

5.5.3 Fatigue
The Wave Height vs. Direction button displays the following table :

..... . ....

0 179 94935

The Wave Height vs. Period button displays the following table :
ratigue : W*Ye Height " "pcrorrinq Period I Area : REDSER Position : 235079E, 2 6 3 0 0 4 2 8 Depth : - 5 0 0 . 0 0 1m1

- A"nnua1
1.75

Number

or

waves

................................................ ...............................................
I Wave Height I Period 18) I 0.25
0.75
1.25 2.25 2.75 3.25 3.75 4.25 4.15

0

0

Red Sea Hindcast Study

14

Database User Guide

a

5.6

Joint Statistics This button gives access to joint statistics data through two buttons :
The Hs vs. Tp button displays the following table :

a

The H vs. W button displays the following table : s s
mea
J o i n t OscUrence of Hs v s . Wind Speed

: D m Position : 764010E. 30206538 Death : -500.00 1m1

I
D* (1.2 1.2 4.3 6.1

0.1 0.8

0.2

0.

1.6
0.5

0.3 0.1
Ot

.. . .

0.6

.. .. . .

a

0

Red Sea Hindcast Study

15

Database User Guide

5.1

0

Storm Summary

5.8

Miscelaneous

The following table is displayed :

.............................. .............................
....1.... ...... ...I .. .

I D m
IAizT.
I Water T.
IC1 IC)

I winter
119.4 I 21.6

spring S m e r I...................
28.5 24.2 32.3 25.8

rail
23.3 22.7

hnual I
25.9 23.6

I
I

I TNB" .........I IAirT. IC1 I Water I. I C 1 I 5.lI"ltY lPPtl I Density 1g/11 I BOtt. h t l .

.............................. ............................. winter . . . . . . . .Sprin9. .m. . .ail. .h.u.l.I. . . . .I/ . . . . . . . . e. . . . . . n.a . . , ... S . r . . .
I
122.3 I 24.4 I
28.6
26.8

31.7
29.8

25.7 27.8

27.1 27.2 39.0

I 1 I

4
~~~~~~~~~~~~~~~~~~~~~

!
spring
28.6 27.4

............................ ...........................
I WICH

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ) . . . . . . . . . . . . . . I. .

.

I uinter I 23.0

smer
31.5 30.2

Fell
26.3 28.3

-""a, I . . .
27.4 27.7

I

I Air T. I Water T. I Salinity I Densitv

IC)

I

IC)
IPPfl la/ll

I 24.9 I
I

I
I

38.8

I JEDrVW .............................I..............I . .

I Winter
IC1 IC)

Sprin9
28.7 28.0

Suer
31.3 30.7

rail

I...........

hnnuai I
27.7 28.4 38.1

I AAr T.
I

I
I

I

I 23.8 Water T. I 25.8 5Llll"itY IPPfl I Denrity 14/11 I BOtf. "atl. I

27.0 2'1.1

I I I

I

.............................. .............................
I Jim ........................................I................ I

I winter
IC1

sF.rin.3
30.0
29.2

suer
32.0 31.2

Fall
28.9 28.7

A"""al
29.3 28.7

I AIr I . I Wafer T. I Salinity I Dcnalry I Bolt. HL. .I

IC1

I 26.1 I 25.8 I

I I

I I
I I

IPPfI I
Ig/l> I

37.2

0

Red Sea Hindcast Study

16

Database User Guide

0

6

PRINTING AND EXPORTING DATA Tables
Tables are presented using the standard Windows WordPad program. This means that text can be transferred to other programs via the Windows clipboard and printing is straightforward.

6.1

6.2

Plots
Plots can similarly be transferred to other programs by pressing the Alt+PrintScm key. This places a copy of the graphics in the Windows clipboard, from which it can be pasted to a programs like Paintbrush, Wordperfect or Microsoft Word. Plots are sent to the printer when the plot hardcopy option in the view menu is set. See the technical description in section 7 for details on setting up the printer.

0

Red Sea Hindcast Study

17

Database User Guide

0

7

TECHNICAL DESCRIPTION Directory Structure The installation directoj contains the following subdirectories : Bin Data Maps Menu Temp uniras Programs. The data. Contains sub-directories for each area, data type, etc. The maps used by the program. Menu setup tiles for the plot and tidal programs. Temporary directory used for storing intermediate files. All files are removed from this directory when the map program exits. Directory for plot library.

7.1

7.2

Programs Following is a list of the programs in the bin directory. Main user interface program : Map.exe Setup program : Setup.exe Utility programs : CurrExtr.exe CurrNorm.exe LevExtr.exe LevNormP.exe LevNormT.exe trans8.exe TransTab.exe WaveExtr.exe WavePlot.exe WindExtr.exe WindNorm.exe WRosePlt.exe Plot programs : t0plot.exe t0rose.exe v 12play.bat v 12play.exe Tide calculation programs : tidcpc.exe f.tidcpc.exe

0

Red Sea Hindcast Study

I8

Database User Guide

tidhpc.exe f.tidhpc.exe Utility programs / files : Agx.dli dos4gw.exe Lmgr324a.dll M21Setup.bat utm.dat Vsnl6.dll Windev.exe

7.3

Registry The installation program writes four entries into the registry. The data are written into the key :

HKEY-CURRENT-USER\SoftwareDHI\MapL'.O
The following subkeys are written : InstallDir DataDir MenuDir TempDir Password The name of the installation directory The name of the data directory, normally the "Data" directory under the installation directory. The name of the menu directory. The name of the temporary directory. Encrypted password. Do not delete or change this.

You can use the program regedit or regedt32 to edit the registry.
7.4

Printing Information on printer setup for the plot programs is stored in the file pltdev.dat in the menu directory. This file contains information about the drivers used for the graphics. A typical file is shown below :
H '
I I I 1. entry 2. entry 3 . entrv 9 . entri 5. entry 6 . entry 7 . entry 8 . entry 9. entry
:
:

H
Y #
#

: : : :
:

H H H

:
:

Selection name Uniras device name Color i = 11 or bw I = 0) Screen I = 1) OT hardcopy I = 01 length in x-direction Im]. Length in y-direction Iml. NO. of pixels in x-direction. (only scseenl No. of pixels in y-direction. lonly screen1 Driver lonly hardcopy devices).

-.

X-window max is set to screensize 'lwin' '\\OBELIX\laser-hic' 'laser' 'video' 1 1 320 1 0 285 1 0 285 0 0 210.0
240

H
'Desktop' '\\OBELIX\laser-hic' 'laser' 'Video' 1152 8 6 4 'Xll' 197
6 7 3 8 4 6 4 6 'Win' 197 6 7 3 8 4 6 4 6 'Win' 155.0 6 4 0 4 8 0 'Bcvgal2'

Comment lines start with a '#'.
Red Sea Hindcast Study
19

0

Database User Guide

0

The line controlling the hardcopy is the second non-comment line. The file is generated when you run the setup program, but it can be regenerated by running the w@dev.exe program in the bin directory. You must be located in the menu directory when you do this. The program can take some time completing if you have network printers.

-

To change the printer, move the line corresponding to the desired printer so that it
becomes the second non-comment line in the file. Note that the screen size given under Desktop must not be smaller than 210 x 155. If this is the case, edit the file manually to correct this. 7.5 Customizing Plots It is possible to customize the plots created by the system. This is done by running the plot programs in interactive mode :

0

1.

2. 3. 4. 5. 6. 7.

8. 9.
10.

Delete all files in the temp directory. Make the plot you want to customize using the Map progam. Do not quit the map program, as this will remove all files from the temporary directory. Edit the file M21Setup.bat in the bin directory so that it reflects your setup. Open a DOS prompt. Run M21Setup.bat. Change directory to the temp directory. Run either toplot or torose depending on if it is a time series or a rose plot you want to make. Select load specifications to load the .inp file created by the map program. Customize the plot by using the various options. Select Execute to create the plot.

7.6

Troubleshooting The system writes debugging information to various files in the temporary directory. To troubleshoot the system, delete all files from the temporary directory, repeat the operation that caused problems and examine the files that have been created. The most likely files to contain relevant information are files with extension .log and .gpr.

0
7.7

Un-Installation The system can be un-installed in the following way :
0
0

Delete the installation directory, typically C:\Redsea. Delete the key Redsea under HKEY-CURRENT-USER\Software\DHIin the registry. This can be done by using the registry editor (regedit). Be careful only to delete the Map key and nothing else or you may cause irrepairable damage to other software installed on the computer.

0

Red Sea Hindcast Study

20

Database User Guide