Precipitation representation over a two‐year period in regional reanalysis

Abstract

Precipitation is a critical aspect of climate. In Europe, extreme precipitation events are costly and represent a threat to life. Evidence suggests the frequency and intensity of these events is increasing in Europe and therefore long‐period datasets that can represent such events accurately are required. Precipitation is challenging to represent in gridded models and therefore is not well trusted in relatively low‐resolution global reanalyses. For the first time, the European Reanalyses and Observations For Monitoring (EURO4M) project has produced high‐resolution atmospheric regional reanalyses of Europe that improve representation of precipitation. These are based on operational forecast systems at the Met Office and the Swedish Meteorological and Hydrological Institute (HIRLAM). The improvement of quality of precipitation in the regional reanalysis datasets over their parent global reanalysis is demonstrated here, together with a discussion on the difficulties of validation of gridded precipitation. It is shown that regional reanalyses show particular improvement in representing high‐threshold events. It is also shown that higher resolution, time‐varying data assimilation and direct assimilation of precipitation all contribute to improving representation of precipitation. Resolution is of particular importance when representing extreme events.