Premium
Future changes due to model biases in probabilities of extreme temperatures over East Asia using CMIP5 data
Author(s) -
Seo YeWon,
Yun KyungSook,
Lee JuneYi,
Lee YangWon,
Ha KyungJa,
Jhun JongGhap
Publication year - 2017
Publication title -
international journal of climatology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.58
H-Index - 166
eISSN - 1097-0088
pISSN - 0899-8418
DOI - 10.1002/joc.5233
Subject(s) - climatology , ensemble average , environmental science , climate model , coupled model intercomparison project , climate change , probability density function , general circulation model , extreme value theory , mean radiant temperature , atmospheric sciences , heat wave , statistics , mathematics , physics , geology , oceanography
This study examines the performances of 31 global climate models in the Coupled Model Inter‐comparison Project 5 (CMIP5) in terms of probability density functions (PDFs) for maximum ( T max) and minimum ( T min) air temperatures over East Asia in the present and CMIP5‐model projected future changes. In general, most of models well reproduce warm‐season peak for both T max and T min but exhibit large inter‐model spread for simulating cold‐season peak, especially for T min. Minimum values of T min and T max are more strongly dependent upon model selection than maximum values of them. For the last 25 years of the 21st century, under the Representative Concentration Pathways 4.5 scenario, models project shifts toward warmer values in the PDFs of T max and T min and broadening in the shape of PDFs. Models with warm biases in PDFs tend to show larger shifts in temperature changes, but seasonal mean temperature biases do not affect to future changes. It is notable that the broadening of PDFs in the future influences temperature extreme events. Using the changes in probabilities of heat waves as one of extreme temperature events by comparing multi‐model ensemble (MME) and models with good performance of PDFs, this study shows that MME tends to overestimate its duration. Our findings suggest that future changes in temperature extremes projected by models are strongly come from the biases detected in those models when simulating present extreme temperature PDFs. Therefore, correcting the intrinsic biases of models rather than seasonal mean correction is necessary to reduce the uncertainties in predicting future changes in temperature extremes.