Wind energy production backcasts based on a high-resolution reanalysis dataset

Numerical weather prediction reanalysis data has been widely used in wind energy studies. Examples include large-scale wind resource analysis and wind energy backcast simulations for electricity market integration studies. The success of such studies is highly dependent on the accuracy of the datasets used. The relatively coarse horizontal spatial resolution of many reanalysis datasets such as ERA-Interim (c. 80 km) limits their usefulness in such studies. The advent of high-resolution, country-specifc datasets, such as Met Eireann's MERA reanalysis, allows for more detailed backcasts to be developed, with corresponding improvements in accuracy. The 2.5 km horizontal resolution of MERA, in combination with its previously-reported low bias compared to ERA- Interim 10 m wind speeds, makes it ideal for wind energy production estimation, as the spatial resolution is sufficient to resolve some terrain effects. In this study, we investigate the accuracy of wind energy production backcasts for a wind farm location in Ireland derived from MERA data. The results of various bias correction schemes are introduced, with the overall results showing good prediction accuracy even when relatively simple corrections such as the Kalman filter are applied.