Collaborative Forecasting Platform: Results after one year of operation for wind and solar forecasting in Belgium

Transmission System Operators (TSOs) rely on accurate power forecasts for secure and efficient system operation, especially with weather-dependent renewable generation and crossborder power flows. Forecasts are usually sourced from third-party service providers. As renewable energy capacity grows, forecast errors increase balancing and operational costs, making accuracy increasingly important. Traditional forecast procurement through bilateral contracts is limited by high costs, low transparency, weak incentives for continuous improvement, and barriers for innovative or smaller forecasters. We present a collaborative forecasting platform where TSOs and other forecasting consumers create open forecasting and data-analytics markets. Forecasters, from commercial to individuals, compete continuously and are rewarded based on objective performance metrics. The platform aggregates independent forecasts into ensembles that outperform single forecasts, supports joint funding by multiple forecasting services consumers, and enables a data analytics market. Developed as open-source software, it was tested on Belgian day-ahead offshore wind and solar forecasts, improving point accuracy (RMSE) by 6–8% and probabilistic (MWI) by 7–14% over current operational forecasts.