Dec 29, 2020

Attribution of Extreme Precipitation with Updated Observations and CMIP6 Simulations

by
Siyan Dong, Ying Sun, Chao Li, Xuebin Zhang, Seung-Ki Min, and Yeon-Hee Kim
,
AMS Journal of Climate

Study key findings & significance

  • The study presents a formal detection and attribution analysis of the observed changes in precipitation totals from heavy precipitation days and the contributions of these extreme precipitation days to annual precipitation totals. 
  • The study uses extreme indices from the updated observations and the newly available CMIP6 simulations.
  • The influence of anthropogenic forcing can be detected in these extreme indices globally and in the Northern Hemispheric continental regions. 

Abstract

While the IPCC Fifth Assessment Working Group I report assessed observed changes in extreme precipitation on the basis of both absolute and percentile-based extreme indices, human influence on extreme precipitation has rarely been evaluated on the basis of percentile-based extreme indices. Here we conduct a formal detection and attribution analysis on changes in four percentile-based precipitation extreme indices. The indices include annual precipitation totals from days with precipitation exceeding the 99th and 95th percentiles of wet-day precipitation in 1961–90 (R99p and R95p) and their contributions to annual total precipitation (R99pTOT and R95pTOT). We compare these indices from a set of newly compiled observations during 1951–2014 with simulations from models participating in phase 6 of the Coupled Model Intercomparison Project (CMIP6). We show that most land areas with observations experienced increases in these extreme indices with global warming during the historical period 1951–2014. The new CMIP6 models are able to reproduce these overall increases, although with considerable over- or underestimations in some regions. An optimal fingerprinting analysis reveals detectable anthropogenic signals in the observations of these indices averaged over the globe and over most continents. Furthermore, signals of greenhouse gases can be separately detected, taking other forcing into account, over the globe and over Asia in these indices except for R95p. In contrast, signals of anthropogenic aerosols and natural forcings cannot be detected in any of these indices at either global or continental scales.