Bias correction of global irradiance modelled with weather and research forecasting model over Paraguay

Abstract The estimation of solar irradiance is performed by means of numerical weather prediction (NWP) models that include all the necessary information to solve the temporal, geographical and atmospheric conditions variability being this the basis of solar energy applications. However, the radiative transfer schemes implemented in meteorological models show systematic errors in the simulation of global horizontal irradiance (GHI). In this contribution, we present a post-process analysis of Weather and Research Forecasting (WRF-ARW) model which combines a Kalman Filter with Model Output Statistics (MOS) for bias correction in order to improve the overall predicted values of GHI simulations over Paraguay. The hourly GHI is simulated at 4 × 4 km 2 of spatial resolution. The annual evaluation of the hourly WRF model without post process shows relative mean bias error (rMBE) of 21% and relative root mean square error (rRMSE) of 81%. The results using several ground stations and combinations of post-process show an annual correction of systematic errors with rMBE of −0.7% and rRMSE of 70%.