Assessment of Wind Forecasts from a Numerical Weather Prediction Model for Indian NPP Sites

This paper evaluates the performance of a Numerical Weather Prediction Model, National Center for Medium Range Weather Forecast-NCMRWF Unified Model-Global (NCUM-G), for wind forecasts at selected Indian Sites viz. Tarapur Atomic Power Station (TAPS) and Kudankulam Nuclear Power Plant (KKNPP). The study has been performed for a duration of three months during which predicted and measured hourly wind data was inter-compared. Forecast accuracy is evaluated using standard statistical methods for linear and circular variables. The performance of the model in forecasting wind speed and direction is seen to be season (month) dependent. The measured and predicted wind data is found to be in good convergence, though the distribution of individual differences between measured and predicted wind data indicated random error. The model performance can be further improved by data assimilation, proper choice of parameterization schemes, using high resolution input data as well as increasing model resolution.