The Multimodel SuperEnsemble technique is apostprocessing method for the estimation of weather forecast parametersreducing direct model output errors. It differs from other ensemble analysistechniques by the use of an adequate weighting of the input forecast modelsin order to obtain a combined estimation of meteorological parameters.Weights are calculated by least-square minimization of the differencesbetween the model and the observed field during a so-called training period.Although it can be applied successfully on continuous parameters liketemperature, relative humidity, wind speed and mean sea level pressure, theMultimodel SuperEnsemble also gives good results when applied on theprecipitation, a parameter quite difficult to handle with standardpost-processing methods. Here we present a methodology for the Multimodelprecipitation forecasts with a careful ensemble dressing via theprecipitation PDF estimation.
展开▼
