CMIP6 multi-model evaluation of summer extreme precipitation over East Asia

We provided an overview of individual CMIP6 models and their multi-model ensemble effectiveness in representing both spatial and temporal variation of summer precipitation characteristics across East Asia. The historical runs of the eleven individual CMIP6 models and their ensemble were compared against GPCP gridded observed precipitation for the historical timescale of 1997–2014, with a particular emphasis on the summer monsoon season to evaluate their performance. We quantitatively investigated the individual skill of a selection of models from the CMIP6 experiment and the mean of their ensemble in representing summer precipitation extreme over East Asia using some selected statistical metrics, i.e., normalized mean bias error, normalized root mean square error, which are summarized by the inter-annual variability score (IVS), pattern correlation coefficient (PCC), Taylor skill score, portrait diagram, and the Taylor diagram. Based on the "portrait" diagram and skill score ranking, we indicated that E3SM-1-0, CESM2, BCC-ESM1, Ensmean, and KIOST-ESM were the top five models which performed relatively well in representing both the spatial distribution and inter-annual variability for all five extreme precipitation indices. Additionally, over the mountainous regions of Bangladesh, Nepal, Laos, and Burma, all CMIP6 models overestimated the annual mean of total precipitation on wet days. The vast majority of CMIP6 models were capable of accurately depicting the geospatial extent of all five of the summer extreme precipitation characteristics that were evaluated in the study.