Validating surface downwelling solar irradiances estimated by the McClear model under cloud-free skies in the United Arab Emirates

•McClear, a physical model, uses inputs from MACC to estimate the solar irradiances.•We validate the solar irradiance estimates of McClear under cloud-free skies.•We compare MACC and AERONET AODs to understand the errors of McClear.•McClear performs notably better in DNI estimations when compared to other models. McClear, a fast model based on a radiative transfer solver, exploits the atmospheric properties provided by the EU-funded MACC project (Monitoring Atmospheric Composition and Climate) to estimate the surface downwelling solar irradiances for cloud-free instances. This article presents the first validation of the McClear model for the specific climate of the United Arab Emirates where skies are frequently cloud-free but turbid. McClear accurately estimates the global horizontal irradiance measured every 10min at seven sites. The bias ranges from −9Wm−2 (−1% of the mean observed irradiance) to +35Wm−2 (+6%). The root mean square error (RMSE) ranges from 22Wm−2 (4%) to 47Wm−2 (8%) and the coefficient of determination ranges from 0.980 to 0.990. Estimates of the direct irradiance at normal incidence exhibit an underestimation that is attributed to the overestimation of the aerosol optical depth in the MACC data set and not accounting for the circumsolar radiation in McClear. The corresponding bias ranges from −57Wm−2 (−8%) to +6Wm−2 (+1%). The RMSE ranges from 62Wm−2 (9%) to 87Wm−2 (13%) and the coefficient of determination ranges from 0.830 to 0.863. When compared to two other models in the literature, McClear is better able to capture the temporal variability of the direct irradiance at normal incidence. The validation results remain comparable for the global horizontal irradiance.