The increasing integration of renewable energy sources requires reliable forecasts across multiple time horizons to ensure the security of supply, reliability and system efficiency. While short-term or day-ahead forecasting approaches and climate projections on decadal timescales are well established fields of research, seasonal forecasting remains a less explored research field with limited and strongly regime-dependent predictive skill.

This paper presents an approach for forecasting photovoltaic (PV) power generation in Germany with a lead time of up to six months. The methodology is based on seasonal ensemble climate forecasts provided by the German Meteorological Service (DWD) for key meteorological variables, including global radiation, temperature, and wind speed.

Based on these climate variables, models are developed to derive PV generation forecasts that provide both deterministic and probabilistic information. A particular focus is placed on the use of ensemble methods for uncertainty quantification, as well as on identifying suitable spatial and temporal aggregation levels. Forecast performance is assessed against climatological baselines using established verification metrics and validated against observed PV generation data at both plant level and aggregated transmission system operator level.

Preliminary results indicate forecast skill in extended winter and summer periods, while reduced predictive performance is observed during transitional seasons. For those same seasons, previous results on wind power forecasting within the same research project already prove to have substantial forecast skill.

The aim of this work is to systematically investigate the potential of seasonal climate forecasts for PV generation planning in Germany, thereby contributing to the integration of renewable energy sources into current and future energy systems.