This document discusses wind power forecasting accuracy and uncertainty in Finland. It analyzes forecast errors and imbalance costs based on real data from over 20 wind power sites in Finland from 2010-2012. The key findings are:
1) Aggregating production from multiple sites can significantly reduce forecast errors and imbalance costs, lowering costs by up to 60%.
2) Combining forecasts from different weather models, such as averaging them, can improve accuracy by up to 6%.
3) Quantifying the uncertainty of forecasts through probability intervals is important as wind power penetration increases. Aggregation also decreases the variation and uncertainty of forecasts.