Abstract Solar energy is key factor in the demand for clean energy development and management. In particular, global horizontal irradiance (GHI) and direct normal irradiance (DNI) are the foremost important solar resource components that need to be well characterized in order to seek an efficient operation of photovoltaic and concentrated solar power plants, respectively. The objective of the present work is to assess the quality of short-term (24 h) forecasts from a global Numerical Weather Prediction (NWP) model regarding the GHI and DNI components for solar energy applications. Forecast accuracy for the Integrated Forecasting System (IFS), the global model of the European Centre for Medium-Range Weather Forecasts (ECMWF), was verified through the comparison of the predicted hourly values with the corresponding ground-based measurements in southern Portugal. In this study, results from one year of IFS data are analysed, yielding a general good agreement between model and four ground-based measuring stations. High correlations occur particularly for GHI whilst DNI simulations are predominantly hindered by cloud and aerosol representation (i.e. the radiative effects of clouds tend to be underestimated by the model and the radiative effects of the aerosols are overestimated by the model under very clear atmospheric conditions), being closely linked to the parameterization of absorption and scattering phenomena as function of cloud and aerosol type and dimension. Relative differences of annual availabilities for GHI are found between ∼0.16% to ∼2.12% whilst for DNI values ranging from ∼7% to ∼12% are found. The respective correlations coefficients are around 0.95 for GHI and between 0.65 and 0.77 for DNI. Regional irradiation maps of GHI and DNI are presented, showing that NWP predictions are an important tool for the operation of electricity generation systems based on solar energy.