中文摘要：

近几十年来,中国区域气候经历了以变暖和东部降水呈现＂南涝北旱＂分布特征为主的变化.这里基于一个区域气候模式（Reg CM4.0）历史模拟和归因试验的对比,探讨了温室气体排放和自然变率分别对上述特征形成及各大水文流域气候变化成因的可能贡献.试验中Reg CM4.0的分辨率取为50 km,积分时间为1961~2005年,两个试验的驱动场分别来自于BCC_CSM1.1全球气候模式的历史试验和归因试验结果.分析结果表明,1961~2005年观测中出现的气候变暖现象,在所有流域均主要是由于人为温室气体排放引起的,在大部分流域自然变率亦有所贡献,中国区域平均的变暖中,人为温室气体排放所产生的作用达到80%.中国东部降水的＂南涝北旱＂分布变化,则可能主要是自然变率起了主导作用,人为温室气体排放在一定程度上减弱了这种变化的强度;中国西北地区的降水增加,可能主要源于人为温室气体排放的贡献,自然变率的作用与其相反,是导致降水减少的.同时指出,分析结果中关于气温的结论相对可靠性较高,降水的结论尚存在较大不确定性,未来需开展进一步的深入研究.

英文摘要：

China has experienced warming and a pattern of ＂southern flood and northern drought＂ in precipitation in recent decades. Attributable contributions of greenhouse gas emissions and natural climate variability to these phenomena are discussed for China and the main hydrological basins, based on a comparison of historical and attribution simulation tests using a regional climate model（Reg CM4.0）. Two tests are performed at a resolution of 50 km, driven by the results of BCC_CSM1.1 historical and attribution simulation tests of CMIP5 in 1961–2005, respectively. The forcing of greenhouse gas emissions and natural climate variability are included in the historical test, and greenhouse gas emission is excluded in the attribution test. Results show that the simulated trends of temperature and precipitation from Reg CM4.0 are more reasonable than from BCC_CSM1.1 for 1961–2005, and the warming during this period was mainly caused by the emission of greenhouse gases, although natural climate variability also led to regional warming in most areas except the Tibetan Plateau. The warming magnitude from greenhouse gas emissions was much larger than from natural climate variability over most basins and China as a whole, with 80% of the warming attributable to forcing by greenhouse gases. The ＂southern flood and northern drought＂ pattern of precipitation in Eastern China may be largely due to natural climate variability, and greenhouse gas emission to some extent weakens the strength of this change; increased precipitation in Northwest China is mainly induced by increased greenhouse gas emission, while natural climate variability, in contrast, leads to reduced precipitation. The assessments of temperature are more reliable than those of precipitation, because model limitations and the complex relationship between precipitation and external factors result in large remaining uncertainties in precipitation assessments. It should be pointed out that attribution studies of climate change on the regional scale are very