Seasonal forecasts of 100m Wind and solar radiation for Germany

As Germany accelerates its renewable energy transition aiming for 80% renewable electricity by 2030, the expansion of wind and solar capacity, the phase-out of fossil fuels, and improvements to grid infrastructure are crucial for ensuring a sustainable and secure energy system. In this context, seasonal forecasts of wind and solar radiation have the potential to support energy reserve management, planning for variable renewable supply, and improving the long-term resilience of the energy system. This study focuses on the development and evaluation of seasonal forecasts for 100m wind speed and solar radiation across Germany. We apply the statistical-dynamical downscaling method EPISODES (Kreienkamp et al., 2019) to hindcast data (1990–2020) from the German Climate Forecasting System Version 2.1 and investigate the predictability and forecast skill of 100m wind speed and solar radiation forecasts on lead times ranging from one to six months. The analysis focuses on the summer season, when solar energy production is highest, and the winter season, when wind energy production peaks. Despite the overall rather low forecast skill of seasonal forecasts for Germany, we find that skillful wind forecasts for the winter season are possible with a reasonable correlation to observations. Furthermore, the forecast model is able to predict solar radiation in summer over southern Germany, a region that contains most of the solar plants in Germany, relatively well. We further employ a statistically selected subsampling approach (Dalelane et al., 2020 and Dalelane et al. 2025, in preparation) to generate a smaller ensemble based on large-scale teleconnections in the North Atlantic and apply it to the forecasts. With this approach, we find a substantial increase in forecast skill for both wind and solar radiation in both summer and winter compared to the full ensemble. Our findings show that skillful seasonal forecasts in winter and summer are possible despite the limitations and challenges of seasonal prediction. In the future, we plan to use multi-model approaches and teleconnection indices to further explore potentials for more skillful seasonal prediction of wind and solar radiation and publish skillful forecasts on the DWD climate prediction webpage (http://www.dwd.de/climatepredictions). This user-oriented website consistently evaluates and displays subseasonal, seasonal and decadal climate predictions at high resolution for Germany.

Kreienkamp, F., Paxian, A., Früh, B., Lorenz, P., & Matulla, C. (2019). Evaluation of the empirical–statistical downscaling method EPISODES. Climate dynamics, 52, 991-1026.

Dalelane, C., Dobrynin, M., & Fröhlich, K. (2020). Seasonal forecasts of winter temperature improved by higher‐order modes of mean sea level pressure variability in the North Atlantic sector. Geophysical Research Letters, 47(16), e2020GL088717.