Abstract Global horizontal irradiance (GHI) is simulated using a three-dimensional atmospheric meteorology-chemistry model and a triple-nesting configuration over the Middle East with a focus on the hot desert climate of Qatar. The model performance was assessed with measurement data of solar radiation from a ground monitoring station in Doha (Qatar) collected over a three-month period, of representative and distinct meteorological regimes. We have examined the ability of the model to reproduce GHI values under two different shortwave downward radiation parameterizations, and assessed the sensitivity of our results to the presence of aerosols. The introduction of an advanced treatment of aerosols greatly improves the model performance in predicting GHI. Explicitly including aerosol processes and its emissions in the model significantly reduces the relative root mean square error for GHI from 25% to 13% in May and from 20% to 12% in August. A significant improvement of the systematic bias was achieved (from up to 30% to approximately 2%) when aerosols are fully considered during all three seasons. The RRTM (Rapid Radiative Transfer Model) shortwave radiation scheme performs somewhat better than the Goddard scheme both with and without aerosols. This work suggests that GHI predictions in regions that experience high aerosol loadings can benefit significantly from a detailed and explicit treatment of aerosols and their physicochemical processes. This offers a novel approach to better manage the fluctuating nature of solar radiation originating from variable weather and air pollution conditions.