Saturday Dec 15, 2018

Oman Sea 22-year Hindcast- 1st Phase

Introduction
Reliable wave information is essential to the design of any coastal facility and to the assessment of coastal processes, such as sediment transport.  A detailed investigation of the wave climate on the Oman Sea coastline of Iran was carried out with the objective of development of a long-term wave climate data for the entire length of the Iranian coastline in the Oman Sea.
 
The Oman Sea coast is subject to a complex wave climate with three distinct components:  (1) monsoon waves generated during monsoon season (June to September) off the southern coastline of Arabian Peninsula in the Indian Ocean and approach from a southerly direction, (2) seas that are generated in the Oman Sea during winter season and approach from a westerly to southwesterly directions; and (3) long-period swells that are generated in the Indian Ocean and approach from southerly to southeasterly directions.

A comprehensive hindcast of wave conditions in the Oman Sea for a twenty-three year (1985 to 2007) time period has been carried out using a state-of-the-art 3rd generation wave model.  Extensive comparisons have been carried out with measured wave data from satellite altimeter, buoy and acoustic doppler current profilers.   

Wave Model

The WAVEWATCH III wave generation model, as developed by the U.S. National Oceanic and Atmospheric Administration (NOAA), was utilized for the wave simulations.  In this study, the main model outer grid extended from 20°E to 123.75°E longitude with a grid resolution of 1.25°, and from 71°S to 26°N latitude with a grid resolution of 1.0°.   In order to get a better definition of waves along the southern coast of Iran, a high resolution nested model was used.  The inner model grid extends from 19.75° N to 27.55° N and 55° E to 73 E with a grid resolution of 0.1°.  The inner model boundary within the outer model extends from 55° E to 73 E along 19.75° N latitude. 
 
Figure 1 WW3 Model Grids 
Figure 1  WW3 Model Grids
 
 Wind Field
A wide variety of wind datasets were considered during the modeling process, as summarized in Table 1.  Various wind datasets were intercompared both statistically and in time series plots.  Based on these comparisons, a combination of NCEP, GFS and Chabahar synoptic station data was used.
Table 1  Wind Data Sources Utilized in the Study 
Table 1 Wind Data Sources Utilized in the Study    
 
QuikSCAT Satellite Corrections to the Wind Data
Global atmospheric model data will tend to under-represent peak wind speeds in storm events due to limited grid resolution.  As well, it is well known that such wind data can have spatial biases.  To address these limitations, spatially varying non-linear corrections (as a function of wind speed) were developed for the global atmospheric winds through grid point by grid point comparisons to the QuikSCAT winds over the Indian Ocean region.
 

Figure 2 Example Seasonal Wind Correction Factors for March, April and May
Figure 2  Example Seasonal Wind Correction Factors for March, April and May

WAVE DATA VALIDATION
Comparisons of the hindcast wave data were carried out to measured wave data in the Indian Ocean and the Oman Sea, including Chabahar.  The comparisons were carried out for the following datasets:

  • Topex/Poseidon and Jason-1 Satellite Altimeter Measurements.  
  • Indian Buoy DS1.
  • The Chabahar Wave Buoy.
  • ADCP Data in Chabahar Bay (AW1, AW2).
Measurements of winds and waves on Indian Ocean are very sparse making simulation of wind and wave fields of Indian Ocean a challenge.   In order to verify the WW3 hindcast, a comparison of the hindcast with deepwater wave data recorded by an Indian wave buoy in Indian Ocean was completed.  The data were obtained from the National Institute of Ocean Technology in India for a wave buoy (DS1) located on the west coast of India at 15.334ºN, 69.357ºE.  The data covered the periods from February 1998 to February 1999, and January to December 2000 and provided an excellent dataset for validation of the Indian Ocean model results.  Figure 3 provides a sample time series comparison of significant wave height, peak wave period, and peak wave direction between DS1 buoy measurements and WW3 hindcast for summer of 1998.  In general, excellent comparisons were achieved.
 

Figure 3 Sample Time Series Comparison for Hindcast (Red line) Against DS1 Buoy Wave Conditions (Black line) 
Figure 3 Sample Time Series Comparison for Hindcast (Red line) Against DS1 Buoy Wave Conditions (Black line)

A Waverider buoy was deployed by Iranian Meteorological Organization outside of Chabahar Bay at a location with an approximate 17 m water depth.  The buoy collected data from May 5th, 1998 to September 2nd, 2000 with occasional gaps.  Figure 4 provides sample time series comparison of significant wave height between Chabahar buoy measurements, WW3 hindcast and ISWM hindcast for summer of 1998.

Figure 4 Sample Time Series Comparison of Wave Height for WW3 Hindcast Against Chabahar Buoy Wave Conditions and ISWM Hindcast
Figure 4  Sample Time Series Comparison of Wave Height for WW3 Hindcast Against Chabahar Buoy Wave Conditions and ISWM Hindcast


Figures 5 provides sample time series comparison of significant wave height between the present ADCP measurements and WW3 hindcast for September and October of 2006. Very good comparisons of wave height were achieved at both AW1 and AW2.

Figure 5 Time Series Comparison of Wave Height for WW3 Hindcast Against AW1 and AW2 data
Figure 5 Time Series Comparison of Wave Height for WW3 Hindcast Against AW1 and AW2 data 

Summary of Validation Results
Figures 6 and 7 provide summaries of the statistical comparisons between the hindcast and measured data.  Excellent agreement was achieved in the Arabian Sea comparisons with the satellite data and Indian buoy DS1.  The offshore satellite comparisons with the Oman Sea (Topex location 3) show a slightly reduced accuracy and greater scatter.  Much of this difference can be attributed to inaccuracies in defining the wind field over the Oman Sea.  The results of the comparisons at Chabahar Bay (Figure 7) are very good, and similar to that of Topex Location 3.  The variability is associated with the difficulty in accurately defining wind fields in the Oman Sea.

Figure 6 Summary of Comparison Statistics for Topex Locations and DS1 Buoy 
Figure 6  Summary of Comparison Statistics for Topex Locations and DS1 Buoy

Figure 7 Summary of Comparison Statistics for Measurements in Chabahar Bay 
Figure 7  Summary of Comparison Statistics for Measurements in Chabahar Bay

The Average Relative Error (ARE) for the hindcast was calculated in identical manner to that outlined in DHI (2005) for ISWM using the relationship proposed by Zhou Liu and Peter Frigaard (2001).  Note that, for example, ARE = 10 % (i.e. accuracy of 90%) means that on average, the predicted wave height deviates from the observed wave height by 10 %.  Obviously a smaller ARE value indicates a better prediction.  ARE values corresponding to various measured offshore datasets were obtained and the results are summarized in Table 2.  An accuracy of significantly better than 80% was achieved.  At the DS1 Buoy location the accuracy is close to 90%.

Table 2  Average Relative Error
Table 2 Average Relative Error  

 
In general, excellent agreement was achieved with wave measurements made in the Arabian Sea.   Slightly reduced accuracy and slightly greater scatter was noted with such comparisons in the Oman Sea and in Chabahar Bay.  This change in accuracy is primarily attributed to the difficulty in accurately defining wind fields in the Oman Sea.  The WW3 hindcast results were exported at all model grid points with 0.1 deg intervals in the Oman Sea.  Hourly time series of several wave and wind parameters as well as directional wave energy spectra are provided.  The files are separated into 1985-1997 and 1997-2007 periods because different input winds were used for each period over the Oman Sea grid.


Last Update : Mar 14, 2012 08:01