Forecasting of Wind Resources Using the Weather Research and Forecasting Software

The renewable sources particularly wind resources are highly intermittent in nature. The WRF is the state-of-the-art mesoscale software to predict wind resources. In this study, the prediction skills of WRF forecasting software is tested for the wind speed of the wind farm in Pakistan. The WRF excellently interpolated the weather data to the terrain description of the wind farm's area, which lies in the complex terrain of the Jhimpir wind corridor. Wind speed predicted by the WRF is then compared with the observed speed of the met mast in the wind farm. The WRF predicts wind speed with the quite precision, for June the mean bias error of -0.85 ms
^-1
, mean absolute error of 1.30 ms
^-1
, and root mean square value of 1.63 ms
^-1
, for July mean bias error of -0.85 ms
^-1
and root mean square error is 1.66 ms
^-1
, for the January errors were high, the mean absolute error of 1.87 ms
^-1
, mean bias error is -0.15 ms
^-1
and root mean square error of 2.42 ms
^-1
is observed. Wind direction is also predicted with great accuracy, with the mean error in the wind direction of 33°. Overall, the WRF overestimate the wind speeds when mean speeds are high.