Characterizing The Convergence And Robustness Of The Kernel Density Mapping Method For Site-Adaptation Of Global Horizontal Irradiation In Western Europe

The estimation of the annual solar resource at a given location is a crucial part of the financing of new photovoltaic power plants. The standard practice is to use satellite-derived databases of ground irradiance estimations. Although these databases have a long-term history that properly captures the climatological variations of a given site, they are not calibrated with ground measurements and therefore may suffer from significant biases. In this paper, we focus on Western Europe and propose to use ground measurements of global horizontal irradiation (GHI) combined with a Kernel Density Mapping algorithm to calibrate several satellite databases (Solargis, HelioClim-3 and CAMS Radiation); we then show that their biases can be reduced from roughly ±3 % down to 0.5 %. More importantly, we perform a sensitivity analysis on the minimal amount of ground measurements required to perform a proper calibration and show that 12 months is the perfect duration to achieve optimal calibration performance. Additionally, we compare the calibration performance on several different satellite databases and show that a combination of several databases can lead to lower overall uncertainty.