全球海气耦合模式对我国极端强降水模拟检验
Evaluation of Extreme Heavy Precipitation in Coupled Ocean-atmosphere General Circulation Models
-
摘要: 以1961—1999年我国地面观测逐日降水资料作为观测基础, 初步分析了18个全球海气耦合模式对我国20世纪极端强降水的模拟能力。分析模式对不同级别降水的模拟发现, 各模式模拟的我国1~10 mm小雨日数普遍明显偏多; 10~25 mm中雨日数的模拟结果总体上也以偏多为主, 虽然部分模式能够模拟出我国南方存在的高值中心, 但位置偏北至长江中下游地区; 25~50 mm大雨日数在我国南方明显偏少, 并且大值中心的位置基本都没能模拟出来; 50 mm以上暴雨日数的模拟结果也明显偏小, 除MIROC3.2(hires) 外大部分模式在长江以南地区的结果都未超过2 d; 大部分模式不能正确模拟出我国东部地区大雨日数变化趋势的空间分布。进一步分析各模式对极端强降水的模拟发现:各模式极端强降水阈值明显低于观测; 半数左右的模式模拟出了1961—1999年西北西部极端降水增加的趋势, 个别模式趋势系数的大小与观测相当, 大部分模式对东北和长江中下游地区的模拟结果呈与观测反向的变化趋势, 没有模式能够模拟出我国东部地区存在的东北—华北与华中—长江中下游—华南存在的极端强降水日数增加-减少-增加-减少的空间分布; 大部分模式模拟的极端强降水日数标准差与观测结果比较接近, 这可能主要是由于对观测和各模式使用了同样的判定极端强降水发生的方法。总的来看, 全球海气耦合模式对我国极端强降水的模拟能力还有待进一步改进。Abstract: Based on the surface observation of daily precipitation from 1961 to 1999 in China, the simulation capabilities of 18 coupled ocean-atmosphere general circulation models for extreme heavy precipitation in China are evaluated.The analysis of annual rainy days with different categories precipitation in models shows that the annual rainy days with 1—10 mm daily precipitation in China are overestimated in all of the 18 models.The annual rainy days with 10—25 mm daily precipitation are also overestimated in most areas of China in most of the models and the high value center locates in South China in observation northward to the mid-low reaches of Yangtze River Basin in some models although those models are able to capture the high value centers to some extent.The annual rainy days with 25—50 mm daily precipitation are obviously underestimated especially in the southern China in the models and the locations of high value centers are missed in most of the models.The annual rainy days with daily precipitation greater than 50 mm are also obviously underestimated in all of the models, the days in the models except MIROC3.2(hires) are less than 2 days in south of the Yangtze River.And most models are not able to simulate the right spatial distribution of heavy rain trend in eastern China.The analysis of extreme heavy precipitation in models shows that the thresholds of extreme heavy precipitation are generally underestimated in all of the models; similar increase trend in west part of Northwest China to observation are obtained in about half of the models, and the trend coefficients in some individual model are close to the observation.While opposite linear trend in northeast area and mid-low reaches of the Yangtze River is produced by most of the models, and the " +-+-" spatial mode in Eastern China for the extreme heavy precipitation can not be caught hold by any model; close standard deviations of annual rainy days with extreme heavy precipitation with observation are obtained by most models, which is maybe mainly due to the same analysis method among the observation and the models.In conclusion, it is very necessary to further improve the simulating capabilities of the coupled ocean-atmosphere general circulation models for extreme heavy precipitation.