1.RCE-TEA, Institute of Atmospheric Physics(IAP),Chinese Academy of Science(CAS);2.Laboratory for Atmospheric Chemistry LAC,Institute of Atmospheric Composition and Environment Meteorology,Chinese Academy of Meteorological Sciences CAMS,China
This study focuses on the bias characteristics of multi-model ensemble mean in precipitation basing on the grid datasets of the fifth phase of the Coupled Model Intercomparison Project (CMIP5) models and the observation (CRU TS 4.0). Three bias correction methods are tested and then a projection on precipitation is made for coming 30 years. Results shows that the model precipitation is overestimated in northern Asia and underestimated in the South, with 30%-40% more precipitation in Qinhai-Tibetan plateau, Inner Mongolia and Mongolia, and 20%-30% less in southern coast of China mainland and Vietnam, and 30%-40% less in South Asia than the observation. The similar bias pattern is found in 2006-2015 precipitation projected under RCP4.5 with CMIP5 historical climate simulation, implying that the bias pattern is almost stationary, belonging the model climate drift which can be removed through the difference between a period-mean projection and the historical simulation averaged over a period. Bias correction test confirms that the logarithm regression (LR) is better in the North than year-to-year increment regression (YYIR), whereas the YYIR is better in the South than LR. The combination of the YYIR and LR is applied in the bias correction of the 2021-2050 precipitation projection under RCP4.5. The projection shows a change in precipitation pattern versus 1976-2005, such as 6%-10% increase in warm season (May-October) precipitation in Arabian Peninsula, northwest part of South Asia, southern Iron, northern part of Central Asia, the Far East of Russia, the east part of Northwest China, the northeast part of China and the coast belt around the Bohai Sea; the precipitation would decrease by about 10% in the south part of Central Asia and northern Iron, about 3% reduction in Southwest China mainland. As for the cold/cool season (November to next April) the precipitation projected would increase in the North, especially in high latitude area, and decrease in the South, such as about 6%-10% decrease in South Asia, about 10% drop in Northwest China, North China, Northeast China, Mongolia and high-latitude Asian continent. Hence, climate warming would expand the area of South Asian monsoon westward leading more frequent rainfall with flood in the dry land of Northwest South Asia, like Pakistan, whereas severe drought would likely appear in the area around southern Caspian Sea. It must be noticed that the weak decrease of the precipitation in South China mainland in the projection may imply a severe drought for coming 30 years due to the increase in surface evaporation with temperature warming. So, a preparation should be made against the potential risk of the drought for coming 30 years.