Dependences of wind forecasting bias on the land surface type, radiation budget and synoptic forcing

Abstract:

Prediction of air pollution events and assessment of wind energy rely heavily on the accuracy of atmospheric models in forecasting wind speed and its response to land types, radiative forcing and weather systems. In this study, we investigate the dependence on land surface types, radiation budget and synoptic forcings of wind bias simulated by the Weather Research and Forecast model designed specifically for forecasting solar irradiance and wind energy. We divide the surface-based observational sites over New York State into five groups of different land types (i.e., inland, lakeside, river-valley, seaside, and offshore) and examine the differences in wind diurnal cycle. Preliminary results illustrate that inland sites have the smallest diurnal cycle of wind speed perturbation, whereas seaside and offshore sites have the most obvious wind diurnal cycle. The model systematically overestimates the surface wind speeds by about 2 m/s and shows a better performance in reproducing the diurnal cycle of wind at the seaside sites than the inland sites. We further seek to examine if radiation budget and/or synoptic forcing are responsible for systematic wind bias under different levels of wind speed. The result shows that the model cannot simulate the phenomenon that winds of low and high speed occur during nighttime and daytime, respectively, indicating that wind forecast bias is related to the modeling bias in the diurnal cycle of radiation budget driving planetary boundary layer development. The influences of synoptic forcing will be investigated by examining the relationship of wind bias to weather system indicators.