Evaluation of various wind products around the Yangtze River Estuary

The Yangtze River Estuary is a vulnerable area ecologically with complicated hydrological conditions whose changes depend to a considerable extent on regional wind variation. This work performed a comparison of 10 m vector wind off the Yangtze River Estuary derived from Japanese 55-year Reanalysis (JRA-55), National Centers for Environmental Prediction (NCEP) Climate Forecast System, version 2 (CFSv2), Cross-Calibrated Multi-Platform Ocean Wind (CCMP), and European Centre for Medium Range Weather Forecasts (ECMWF) Interim Re-Analysis (ERA-Interim). To evaluate various wind products, we collected wind observation from five buoy stations. These buoy observations are located in the area off the Yangtze River Estuary and provided by the Shanghai Meteorological Bureau, which cover the year 2012 and have been subjected to quality control. Our work shows that ERA-Interim has the lowest RMSE (1.87 m/s) for wind speed followed by CCMP with RMSE of 1.95 m/s. CCMP has the highest correlation coefficient. For wind direction, ERA-Interim and CCMP are also the best ones with the lowest errors, RMSE of 67.47 degree and 67.50 degree for ERA-Interim and CCMP, respectively. Although the wind speed of JRA-55 is the one with the lowest bias (0.41 m/s), its RMSE is larger than those of CCMP and ERA-Interim. Our results also revealed that the CFSv2 product with the resolution of 0.2 degree has larger errors than the 0.5-degree product for wind speed. Our comparison also demonstrated that the accuracy of various wind products substantially depends on the magnitude of wind speed. For intermediate wind speed (4-12 m/s), all wind products have the lowest errors (RMSE and Bias) in magnitude except for 0.2-degree CFSv2 product which has its lowest RMSE at high wind speed (>12 m/s). For low wind speed (<4 m/s), most wind products have negative biases implying an underestimate of wind speed. At higher wind speed, however, these products show positive biases except CFSv2 which has positive biases for all wind speeds. The results indicate that wind products with finer spatial resolution does not necessarily have better performance than coarser resolution products in representing the local wind around the Yangtze River estuary.