Comparison of Reanalysis, Blending Satellite Data and Buoy Surface Wind Data over the Persian Gulf Basin for 2009

Currently, continuous measurements of wind speed and direction in the Persian Gulf Basin are limited to the locations of few fixed buoys. The main reason for doing this research was the lack of sufficient observations of surface wind data over the Persian Gulf Basin .Surface wind data are essential in maintenance of coastal structures, oil industry, transportation and fishery in the region. In this study, two sources of offshore wind speeds and directions including ERA-Interim reanalysis data with three resolutions (0.75, 0.25 and 0.125) and satellite data collected from National Climatic Data Center-Blended Sea Winds (NCDC-BSW) are compared against the corresponding verifying buoys observations. If gaps exist in buoys dataset, the nearest synoptic station data are used instead to generate relevant wind roses and distributions. Observed buoys data are obtained from three buoys located in different parts of the Persian Gulf‘s Northern Coast. Examining the observed buoys data show that the wind circulates from Northwest in Bushehr synoptic station to West in Asaluyeh and finally, to Southwest in Qeshm. This circulation that is also confirmed by sea level pressure and streamline maps could be associated to the effect of intense Shamal winds which blow most of the times during the year in the region. The results of comparison of the reanalyzed and satellite wind data against the corresponding buoys data showed that if the wind speeds in the reanalyzed and satellite wind data are in the range of 3.6-6.6 m/s, then lower RMSEs and also better correlations are calculated both for satellite and reanalyzed datasets. Generally, as the wind speeds deviate from their respective mean values then, the errors in both datasets will increase. Overall results for wind directions are much better. NCDC-BSW showed acceptable
correlation and fine P values with observations collected by buoys, nevertheless, the results for reanalyzed datasets were not as satisfactory, and attempting different resolutions of this dataset did not improve the results significantly, as the main resolution of this data set is 0.75 and other resolutions are just retrieved by the interpolation of the original data.