The researchers developed a method to improve forecasts of solar radiation, which is important for predicting solar power generation. They used a combination of forecasts from different weather prediction centers to reduce errors and increase reliability. By analyzing the spread of these forecasts, they were able to detect potential errors in advance. The method was particularly effective in winter months and could help predict when forecasts might be inaccurate. This approach could be valuable for improving regional weather forecasts and ensuring a stable energy supply from solar power.