Ensemble forecasting: status and perspectives

摘要

One of the main challenges for Numerical Weather Prediction is the Quantitative Precipitation Forecasting (QPF). The accurate forecast of high-impact weather still remains difficult beyond day 2 and many limited-area ensemble prediction systems have been recently developed so as to provide more reliable forecasts than achievable with a single deterministic forecast. As a consequence the calibration of ensemble precipitation forecasts has become a very demanding task, for improving the QPF, especially as an input to hydrological models. Different calibration techniques are compared: cumulative distribution function, linear regression and analogues method.