Premium
Reliability of climate models for China through the IPCC Third to Fifth Assessment Reports
Author(s) -
Jiang Dabang,
Tian Zhiping,
Lang Xianmei
Publication year - 2015
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.4406
Subject(s) - climatology , precipitation , environmental science , monsoon , mean radiant temperature , general circulation model , climate model , climate change , spatial distribution , china , atmospheric sciences , geography , meteorology , geology , oceanography , remote sensing , archaeology
Based on observation and reanalysis data, 77 coupled global climate models ( GCMs ) participating in the Intergovernmental Panel on Climate Change ( IPCC ) Third ( TAR ), Fourth ( AR4 ), and Fifth ( AR5 ) Assessment Reports are evaluated in terms of their ability to simulate the mean state and year‐to‐year variability of surface air temperature at 2 m and precipitation over China and the climatological East Asian monsoon for the late decades of the 20th century. Results show that GCMs reliably reproduce the geographical distribution of the variables considered. Compared with observations, however, most GCMs have topography‐related cold biases (although these are smaller than those found in previous studies), excessive precipitation, an underestimated southeast–northwest precipitation gradient, an overestimated magnitude and spatial variability of the interannual variability of temperature and precipitation, and an inadequate strength of the East Asian monsoon circulation. Pairwise comparison reveals that GCMs continue to improve from the TAR via the AR4 to the AR5 for temperature, but have little change for precipitation and the East Asian monsoon. The ability of GCMs varies with season and is affected to certain degree by their horizontal resolutions. Both the arithmetic mean and the median of multiple GCMs are little affected by filtering GCMs in terms of their ability, and the multi‐model mean outperforms most of individual GCMs in every respect.