CMIP5全球气候模式对青藏高原地区气候模拟能力评估

青藏高原是气候变化的敏感和脆弱区，全球气候模式对于这一地区气候态的模拟能力如何尚不清楚。为此，本文使用国际耦合模式比较计划第五阶段（CMIP5）的历史模拟试验数据，评估了44 个全球气候模式对1986~2005 年青藏高原地区地表气温和降水两个基本气象要素的模拟能力。结果表明，CMIP5 模式低估了青藏高原地区年和季节平均地表气温，年均平均偏低2.3℃，秋季和冬季冷偏差相对更大；模式可较好地模拟年和季节平均地表气温分布型，但模拟的空间变率总体偏大；地形效应校正能够有效订正地表气温结果。CMIP5 模式对青藏高原地区降水模拟能力较差。尽管它们能够模拟出年均降水自西北向东南渐增的分布型，但模拟的年和季节降水量普遍偏大，年均降水平均偏多1.3 mm d-1，这主要是源于春季和夏季降水被高估。同时，模式模拟的年和季节降水空间变率也普遍大于观测值，尤其表现在春季和冬季。相比较而言，44 个模式集合平均性能总体上要优于大多数单个模式；等权重集合平均方案要优于中位数平均；对择优挑选的模式进行集合平均能够提高总体的模拟能力，其中对降水模拟的改进更为显著。

Evaluation of CMIP5 Models over the Qinghai-Tibetan Plateau

The ability of climate models in reproducing climate over the Qinghai-Tibetan Plateau, where the natural environment is sensitive and vulnerable to climate change, remains unclear. Here, we examine the performance of 44 models participating in the Fifth Phase of the Coupled Model Intercomparison Project (CMIP5) over the Qinghai-Tibetan Plateau by comparing their outputs with ground observations of surface air temperature and precipitation for the period 1986-2005. The results show that CMIP5 models underestimate annual and seasonal temperatures, with an average of -2.3℃ for the annual mean and larger cold biases in autumn and winter. CMIP5 models can reasonably capture the climatological spatial patterns of annual and seasonal temperatures, but they overestimate spatial variability. The ability of CMIP5 models in reproducing annual and seasonal temperatures can be improved through topographic correction. Comparatively, CMIP5 models perform poorly in reproducing annual and seasonal precipitation. They can capture the climatological spatial pattern of annual precipitation that mainly features a northwest-to-southeast increase, but they overestimate annual and seasonal precipitation amounts, with an average of 1.3 mm d-1 for the annual mean mainly derived from spring and summer. Moreover, the simulated spatial variability of annual and seasonal precipitation is greater than that in the observation, particularly in spring and winter. In general, the ensemble mean of 44 models shows a better skill than most of individual models; the 44-model ensemble mean with the same weights performs better than the median of 44 models; and the ensemble mean of the chosen models with a demonstrable ability can further improve the skills of climate models, particularly for annual and seasonal precipitation.