Site adaptation with machine learning for a Northern Europe gridded global solar irradiance product

Gridded global horizontal irradiance (GHI) databases are fundamental for analysing solar energy applications' technical and economic aspects, particularly photovoltaic applications. Today, there exist numerous gridded GHI databases whose quality has been thoroughly validated against ground-based irradiance measurements. Nonetheless, databases that generate data at latitudes above 65˚ are few, and those available gridded irradiance products, which are either reanalysis or based on polar orbiters, such as ERA5, COSMO-REA6, or CM SAF CLARA-A2, generally have lower quality or a coarser time resolution than those gridded irradiance products based on geostationary satellites. Amongst the high-latitude gridded GHI databases, the STRÅNG model developed by the Swedish Meteorological and Hydrological Institute (SMHI) is likely the most accurate one, providing data across Sweden. To further enhance the product quality, the calibration technique called "site adaptation" is herein used to improve the STRÅNG dataset, which seeks to adjust a long period of low-quality gridded irradiance estimates based on a short period of high-quality irradiance measurements. This study introduces a novel approach for site adaptation of solar irradiance based on machine learning techniques, which differs from the conventional statistical methods used in previous studies. Seven machine-learning algorithms have been analysed and compared with conventional statistical approaches to identify Sweden's most accurate algorithms for site adaptation. Solar irradiance data gathered from three weather stations of SMHI is used for training and validation. The results show that machine learning can substantially improve the STRÅNG model's accuracy. However, due to the spatiotemporal heterogeneity in model performance, no universal machine learning model can be identified, which suggests that site adaptation is a location-dependant procedure