| description abstract | s an important global data resource, reanalysis is widely applied for climate impact studies of the past several decades. For the first time, monthly mean temperature and monthly total precipitation derived from the newest generation reanalysis product?the ECMWF twentieth-century reanalysis dataset (ERA-20CM)?is quantitatively evaluated based on probability density functions and 702 meteorological stations during the period of 1960?2009 across China. This study attempts to investigate how well each member ensemble prediction of ERA-20CM performs for different regions. Generally, all ensemble predictions in ERA-20CM are able to recreate the real conditions on a comparable level. More than 90% of the observed probability for temperature and more than 80% of the probabilities for precipitation could be captured by ERA-20CM over China. However, the performance changes significantly from region to region because of different topographical features and climate characteristics. The Tibetan Plateau is the most difficult to model for all member ensembles. The Jianhuai region is the area with the best performance for both temperature and precipitation. Although the best and worst ensembles for temperature and precipitation for each region were selected according to the skill scores, the differences among the 10-member ensemble predictions are negligible. This evaluation would be helpful for the potential users of reanalysis data, such as ERA-20CM for local climate impact assessments in China. | |
