Wind/PV: errors, correlations & forecasts

There is an increasing interest in using PV to generate electricity as the cost of PV modules decreases. This study will present results on forecast skill, systematic errors and correlations of NWP data for wind and PV. Thirty years of radiation data from the MERRA2 and ERA-Interim reanalyses were compared with hourly data from a representative selection of Irish meteorological service synoptic weather stations. Standard skill scores were calculated for relevant parameters: 10-metre wind speed and direction, incoming shortwave radiation, 2-metre temperature. Situations where the reanalysis radiation skill scores were poor were investigated by running WRF forecasts on a domain centred over Ireland. Forecasts were initialised at 12UTC and run for forecast time horizons from one to five days. Time series analysis allowed the correlation between skill scores for wind and PV parameters to be investigated. Finally, a few popular post-processing techniques were applied to investigate the potential to remove systematic errors.