【作者】 刘珂；

【导师】 许吟隆；

【作者基本信息】 中国农业科学院 ， 气象学， 2011， 硕士

【摘要】 20世纪以来,在全球气候变暖的大背景下,极端气候事件频繁发生,中国区域的气候也发生了巨大的变化。从1951年至2001年,年平均地面气温升高1.1℃,增温速率达0.22℃/10a,与平均气温增加对应的是各种极端天气气候事件也发生了变化。在气候变化研究中,利用气候模式模拟是一种有效的方式,至今已经有许多的学者对全球气候模式(GCM)和区域气候模式(RCM)的模拟结果进行了分析,分析表明GCM和RCM对中国区域的气候具有一定的模拟能力,但存在一定的偏差。针对单个GCM在空间分辨率和系统偏差方面模拟效果的局限性,本文将IPCC AR4发布的20个GCM按分辨率分为高/中/低分辨率的模式集合,同时利用高分辨率的区域气候模式(PRECIS)的模拟结果,对比分析了GCM集合结果和RCM在中国区域的的模拟结果。文中分析的气候要素包括:年/夏季/冬季平均气温,年/夏季/冬季平均降水量,日最高/低气温,极端年较差,霜冻日数,高温热浪指数,连续干日数,极端降水贡献率,连续5日最大降水量。每个指标主要从空间分布,时间变化趋势特征进行分析,并初步得到以下结论:1.平均气温方面,高分辨率的气候模式能够更准确地模拟出中国区域的平均气温的空间变化特征,尤其是在西藏高原地区;而高分辨率的RCM明显改进了GCM的模拟效果。在平均气温的变化趋势上,低分辨率的GCM模拟的平均气温的变化趋势与观测最接近;2.与气温相联系的极端气温事件方面,高分辨率的GCM和RCM对极端气温事件的效果相对较好。在极端低温和极端高温事件中,气候模式对极端低温事件的模拟效果好于极端高温事件。在对极端气温事件的变化趋势上:极端低温事件,高分辨率的气候模式模拟的极端气温的变化趋势稍好于低分辨率的气候模式;极端高温事件,中等分辨率的GCM对高温事件的变化趋势模拟效果较好,而其他模式模拟的效果差;3.平均降水量方面,各气候模式集合对平均降水量的模拟效果较平均气温较差,各类GCM在西藏高原南部和西部的降水量偏多,在华南地区平均降水量偏少,高等分辨率的GCM相对于低分辨的GCM,模拟效果改进主要在西藏高原地区,而在华南地区,则未见改进;RCM相对于GCM,在西藏高原和华南地区,模式模拟的降水量分布特征有了明显的改进,这与RCM中加入了中尺度物理过程和更精确的地形因子有关。在平均降水量的变化趋势上,中等分辨率的GCM模拟的平均降水量的变化趋势最接近观测;4.极端降水事件方面,RCM相对于GCM模拟的极端降水事件与观测更接近。各GCM集合中,空间分辨率越高,模拟的极端降水事件的空间分布特征越好。极端降水事件的变化趋势:在无雨事件的变化趋势上,低分辨率的GCM集合模拟的变化趋势与观测最接近;在极端降水的变化趋势上,高分辨率的RCM模拟的极端降水事件的变化趋势更接近与观测;5.模式在模拟的平均气温和平均降水的变化趋势上,GCM模拟的气候要素的变化趋势的年际变率偏小,而RCM模拟的气候要素变化趋势的年际变率偏大。更多还原

【Abstract】 Since the 20th century, in the context of global climate warming, extreme weather events occur frequently, the regional climate in China has undergone tremendous changes. From 1951 to 2001, annual mean surface temperature has increased 1.1℃, temperature growth rate is about 0.22℃/10a, corresponding to the average temperature increasing, a variety of extreme weather events have also changed. Using model simulating climate and forecast is an effective way, so far, there have been many results of the global climate models and regional climate model simulation were analyzed, analyses show that global climate models and regional climate models for the regional climate of China have certain simulation capability, but there are some deviations. For the limitions of the individual global climate model, including systematic bias, low spatial resolution, and mesoscale physical processes can not be described, the paper using the results of model collections of different resolution form 20 global climate models distributed by IPCC AR4, and result of high-resolution regional climate model simulated, then compared the results of global climate models and regional climate model in China. The climate factors analyzed including: annual/summer/winter mean temperature, annual/summer/winter precipitation, daily maximum/low temperature, extreme annual temperature range, frost days, heat wave during index, consecutive dry days and extreme precipitation contribution, 5 consecutive daily maximum precipitations. Mainly from the spatial distribution analysis and the time trend analysis of each characteristics, and the following are the preliminary conclusions:1. In terms of average temperature, high-resolution climate models can smulate more accurately spatial distribution of the average temperature in China, especially in the Tibetan Plateau; high-resolution regional climate model significantly improved the effect of the low-resolution global climate models; for the trend of average temperature, the result of low-resolution global climate model simulated and the mean temperature trends of observed is the nearest.2. For the extreme temperature events associated with the temperature, high-resolution global climate models and regional climate model on the effect of extreme temperature events are relatively good. Between the extreme cold and extreme heat events, the climate model simulations of extreme low temperature events better than the extreme high temperature events. In the trend of extreme temperature events, for the extreme low temperature events, the trend of extreme temperature of high-resolution climate model simulated are slightly better than low-resolution climate models; for the extreme high temperature events, the medium-resolution global climate model is better; the other models have bad effects on the trend of extreme high temperature events.3. In terms of average rainfall, the climate model ensemble simulation results on the average precipitation less than the average temperature observed, the result of various global climate model in south and west of the Tibetan plateau is more than normal rainfall, less than normal average rainfall in South China, higher resolution global climate models relative to the low-resolution climate models are batter, specially in the Tibetan Plateau, but in South China, there is no improvement, regional climate model relative to the global climate models, in the Tibetan Plateau and southern China, the simulated distribution of precipitation is more similar to observation, which joined mesoscale physical processes and more fine topography information. In the trend of the average precipitation, the result of medium-resolution global climate model simulated is closest to the average trend of precipitation observed.4. In extreme precipitation events, respecting to global climate models, regional climate model simulation is closer to observations of extreme precipitation events. Collection of global climate models, the higher spatial resolution, the better the spatial distribution of the simulation of extreme precipitation events. The trends of extreme precipitation events: in the trend of consecutive dry days, low-resolution simulations of global climate models is closest to the trend of observations, in the trends of extreme precipitation, high-resolution regional climate model simulations is closer to the observed.5. In the trend of temperature and precipitation change, the result of the global climate model simulate is smaller than observation, but the regional climate model is bigger.更多还原