Evaluation of Multisource Datasets in Characterizing Spatiotemporal Characteristics of Extreme Precipitation from 2001 to 2019 in China

Jiayi Lu,Kaicun Wang,Guocan Wu,Yuna Mao

DOI: https://doi.org/10.1175/jhm-d-23-0162.1

2024-01-01

Journal of Hydrometeorology

Abstract:The spatiotemporal characteristics of extreme precipitation intensity are crucial for hydroclimatic studies. This study delineates the spatiotemporal distribution features of extreme precipitation intensity across China from 2001 to 2019 using the gridded daily precipitation dataset CN05.1, constructed from an observation network of over 2400 stations. Furthermore, we evaluate the reliability of 12 widely used precipitation datasets (including gauge-based, satellite retrieval, reanalysis, and fusion products) in monitoring extreme precipitation events. Our findings indicate the following: 1) CN05.1 reveals a consistent spatial distribution characterized by a decline in extreme precipitation intensity from the southeastern coastal regions toward the northwestern inland areas of China. From 2001 to 2019, more pronounced declining intensity trends are discernible in the northern and southwestern regions of China, whereas marked increasing trends manifest in the northeastern and the Yangtze River plain regions. National mean extreme precipitation indices consistently exhibit significant increasing trends throughout China. 2) Datasets based on station observations generally exhibit superior applicability concerning spatiotemporal distribution. 3) Multisource weighted precipitation fusion products effectively capture the temporal variability of extreme precipitation indices. 4) Satellite retrieval datasets exhibit notable performance disparities in representing various intensity indices. Most products tend to overestimate the increasing trends of national mean intensity indices. 5) Reanalysis datasets tend to overestimate extreme precipitation indices, and inadequately capture the trends. ERA5 and JRA-55 underestimate trends, while CFSR and MERRA -2 significantly overestimate the trends. These findings serve as a basis for selecting reliable precipitation datasets for extreme precipitation and hydrological simulation research in China.