Solar Energy, Vol.188, 339-352, 2019
Solar irradiance modelling using an offline coupling procedure for the Weather Research and Forecasting (WRF) model
The research proposed herein seeks to improve solar irradiance magnitude and variability results produced by the Weather Research and Forecasting (WRF) model using a novel offline coupling procedure (OCP). The OCP includes simulations in clear sky conditions, where the effects from varying atmospheric composition depend on the broadband clear sky model used, on cloud attenuation and on decomposition techniques to accurately separate the global irradiance into direct and diffuse components. Furthermore, shadowing and slope effects from orographic features and other obstacles are included with much greater detail. Benefits of the offline coupling procedure were quantified by comparison against local solar radiation measurements over a period of one year. A baseline test to explore the different configuration options was implemented. It considers two aerosol databases, three broadband clear sky models, two cloud attenuation corrections based on clear sky index (either global or by components) and four decomposition models. Given the amount of results from baseline tests, the relative root-mean-square error (rRMSE) was used as the criterion to identify the most suitable OCP configuration, which was subsequently used in the performance analysis. The current baseline test comprises seven sites selected from the Baseline Radiation Network (BSRN) in different terrain complexities and atmospheric conditions. Statistical indicators associated with the annual global horizontal irradiance reveal that OCP improved WRF results by 88.4% in terms of the relative mean bias error (rMBE) and by 5.0% in the value of the Nash-Sutcliffe efficiency (NSE). A standard OCP configuration to model global horizontal irradiance only is proposed. The rMBE of annual hourly global horizontal irradiance oscillated between -3.6% and 3.9%, whilst the NSE varied between 0.608 and 0.939. The analogous quantities for diffuse horizontal irradiance were -19.3% <= rMBE <= 4.6% and 0.371 <= NSE <= 0.717, while for direct normal irradiance -6.5% <= rMBE <= 23.2% and 0.101 <= NSE <= 0.656.