Estimation of non-stationary return levels of extreme temperature by CMIP6 models

Abstract

This study aims to investigate the effects of climate change on return level of extreme maximum temperature (EMT) events in Iran. To this end, the CRU gridded dataset was used to collect EMT for the 1901–2014 period and future data were projected from four available CMIP6 models, where the BCC-CSM2-MR performed best under the latest Shared Socioeconomic Pathways-Representative concentration pathways (SSPs-RCPs) emission scenarios for the 2015–2100 period. The non-stationary state of the distribution was considered under three models GEV0 (location and scale parameters are constant), GEV1 (nonstationary of location), and GEV2 (nonstationary of scale) based on the evaluation criteria . The findings indicate that, when using a non-stationary approach and considering the SSP5-8.5 scenario for a 2-year return period, the return level of extreme temperature increased by up to +4°C compared with the stationary approach, while considering a non-stationary approach without climate change, the increase in the return level of extreme temperature was much smaller(up to +0.7°C). MCMC and DE-MC showed no significant differences and demonstrated that all stations are non-stationary in terms of the location parameter (GEV1). HIGHLIGHTS Temperature frequency was analyzed based on the Bayesian method (MCMC and DEMC) under three generalized extreme value models.; The impact of the non-stationary approach was investigated on the return level of maximum temperatures during historical and future periods.; Intra-period trends were assessed in the extreme temperature due to climate change (CMIP6).;

Keywords

• climate change
• extreme temperature
• iran
• non-stationary analysis
• return period