Comparative study of solar radiation measurements in Tucuman, Argentina, with reanalyses and satellite-based datasets

Abstract

Reanalysis and satellite-based datasets are easily accessible sources for evaluating global solar radiation at the surface, particularly useful in regions with scarce observational data. This study evaluates their performance considering two reanalyses (ERA5 and MERRA-V2) and one satellite-based (CAMS) datasets, in comparison with ground-based observations at Tucuman, Argentina, over the period 2006–2023. This area has very limited measurements, and those that exist are not readily accessible. The statistical metrics Mean Relative Absolute Error (MRAE), Mean Relative Error (MRE), and squared correlation coefficient (r2), were assessed for hourly, daily and monthly values. Results indicate that CAMS provide the closest match to measurements, followed by ERA5, which in turn outperforms MERRA-V2. All datasets exhibit a tendency to overestimate measured radiation values. Temporal variability analysis reveals a good agreement between datasets, particularly in capturing diurnal and seasonal cycles. Less accuracy is obtained in the interannual variation which is expected considering that it is relatively weak and of climatic origin, linked in particular with cloud cover. This study contributes to the validation of reanalysis and satellite-based datasets in the southern hemisphere. Enhanced validation efforts and adjustments based on ground measurements are necessary to further refine these datasets for practical applications in solar energy optimization in the region here analyzed where measured data are limited.