Comparison of Reanalysis, Analysis and Forecast datasets with measured wind data for a Wind Power Project in Jhimpir, Pakistan

Abstract

This study aims to evaluate reanalysis, analysis and forecast data of wind speed and direction with the surface measured data at a site in Jhimpir region of Pakistan. Five datasets (reanalysis, analysis and forecast); NCEP-NCAR, 20th Century Reanalysis v2c (20C), CFSR, GFS and NCEP-FNL have been used for this study. The comparison has been made on six hourly and daily bases. The statistical parameters used are Mean Bias Error (MBE), Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and correlation coefficient (R). The Weibull probability distribution has been drawn, the shape and the scale factors have been calculated. The CFSR data shows significantly good results for lower temporal resolution (daily) with a high correlation coefficient for a wind speed of 0.898 while at a higher temporal resolution of 20C shows significantly lower values of MBE of -0.87 m/s and CFSR shows lower values of RMSE of 2.09 m/s. CSFR wind direction data show good agreement with measured data in terms of statistical parameters The wind industry in Pakistan is under development stages with the possible potential of the offshore wind farm and expansion of land wind farms in other areas of the country, this study will be an initiative toward the initial wind resource assessment for potential sites in the wind corridor.