Evaluation of multiple regional climate models for summer climate extremes over East Asia

• Published in: Climate dynamics Vol. 46; no. 7-8; pp. 2469 - 2486
• Main Authors: Park, Changyong, Min, Seung-Ki, Lee, Donghyun, Cha, Dong-Hyun, Suh, Myoung-Seok, Kang, Hyun-Suk, Hong, Song-You, Lee, Dong-Kyou, Baek, Hee-Jeong, Boo, Kyung-On, Kwon, Won-Tae
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.04.2016; Springer; Springer Nature B.V
• Subjects: Analysis; Atmospheric models; Climate change; Climate models; Climatology; Earth and Environmental Science; Earth Sciences; Extreme weather; Geophysics/Geodesy; Marine; Oceanography; Precipitation; Regional analysis; Regional planning; Summer; Temperature; Temperature distribution; East Asia; Korea, Rep; China, People's Rep; Climate extremes; Model evaluation; CORDEX East Asia; Regional climate models; Temperature and precipitation
• ISSN: 0930-7575; 1432-0894
• DOI: 10.1007/s00382-015-2713-z

In this study, five regional climate models (RCMs) participating in the CORDEX-East Asia project (HadGEM3-RA, RegCM4, SNU-MM5, SNU-WRF, and YSU-RSM) are evaluated in terms of their performances in simulating the climatology of summer extremes in East Asia. Seasonal maxima of daily mean temperature and precipitation are analyzed using the generalized extreme value method. RCMs show systematic bias patterns in both seasonal means and extremes. A cold bias is located along the coast, whereas a warm bias occurs in northern China. Overall, wet bias occurs in East Asia, but with a substantial dry bias centered in South Korea. This dry bias appears to be related to the colder ocean surface around South Korea, positioning the monsoonal front further south compared to observations. Taylor diagram analyses reveal that the models simulate temperature means more accurately compared to extremes because of the higher spatial correlation, whereas precipitation extremes are simulated better than their means because of the higher spatial variability. The latter implies that extreme rainfall events can be captured more accurately by RCMs compared to the driving GCM despite poorer simulation of mean rainfall. Inter-RCM analysis indicates a close relationship between the means and extremes in terms of model skills, but it does not show a clear relationship between temperature and precipitation. Sub-regional analysis largely supports the mean–extreme skill relationship. Analyses of frequency and intensity distributions of daily data for three selected sub-regions suggest that overall shifts of temperature distribution and biases in moderate–heavy precipitations contribute importantly to the seasonal mean biases.