Evaluating ensemble post?processing for wind power forecasts

Capturing the uncertainty in probabilistic wind power forecasts is challenging, especially when uncertain input variables, such as weather, play a role. Since ensemble weather predictions aim to capture system, they can be used propagate this through subsequent forecasting models. However, systems are known biased and underdispersed, meteorologists post-process ensembles. This post-processing successfully correct biases variables but has not been evaluated thoroughly context of forecasts, generation forecasts. The present paper evaluates multiple strategies for applying We use Ensemble Model Output Statistics (EMOS) method evaluate four possible strategies: only using raw ensembles without post-processing, one-step strategy where post-processed, we ensembles, two-step both Results show that final improves forecast performance regarding calibration sharpness, whilst does necessarily lead increased performance.