Performance evaluation of reanalysis models for upsampling of solar irradiance and wind speed data

The transition to renewable energy sources is a cornerstone of modern energy system transformation. Yet, the integration of solar and wind power remains challenging due to limited availability of high-resolution time series data. Accurate modeling and forecasting of renewable energy generation are essential for grid stability, and the effective deployment of technologies such as water electrolysis. However, most available datasets are provided at coarse temporal resolutions, which restricts their utility for dynamic system analysis and hinders the development of robust operational strategies. This study addresses this critical gap by evaluating the performance of reanalysis models and introducing a statistical upsampling approach that generates data with 1-min resolution from reanalysis data with 1-h resolution using a first-order Markov chain Monte Carlo method. The method is validated against measured data. The normalized root-mean-square error of the simulated data is less than 10% and the mean intra-hour standard deviation differs from 4.8% to 9.1% compared to measured values for solar irradiance and wind speed. The generated data facilitates the optimization of system operation, and enhances the planning of sector-coupled infrastructures. Ultimately, this approach enables reliable integration of renewable energy despite limited data availability and thus contributes to the advancement of the sustainable energy transition.