我国西北地区地形复杂、环境恶劣、气象观测水平有限,在海拔较高的山区缺乏长期观测资料。再分析气象数据集(产品)基于气象卫星、数值模拟等技术,结合站点实测资料制作而成,具有时空分辨率高、可靠性好等优点,越来越多地被应用于大气水文相关的研究工作中。选择西北典型的高寒山区流域为研究对象,首先分析3套具有代表性的再分析气象数据产品(CMADS、ITPCAS和TRMM 3B42)降水资料在研究区年际和年内的分布情况,然后利用3套降水资料驱动SWAT水文模型,将模型模拟的结果与水文站实测径流资料进行比较,并进行参数率定和敏感性分析,从而分析3套数据集在研究区对降水的模拟效果。最终获得以下结论:3套降水资料在流域内的降水量分布情况存在较大差异,CMADS和ITPCAS数据集在降水总量上大于TRMM 3B42数据集,且CMADS数据集的年际降水和年内降水分布与冰川分布较为一致;3套再分析数据产品中,CMADS数据集驱动水文模型的模拟效果最好,逐日、月的NSE系数达到0.53、0.71,ITPCAS和TRMM 3B42数据集的NSE系数较低,径流模拟效果较差。
In the northwest China,the terrain is complex and the environment is bad,the level of perennial meteorological observation is limited,and the long-term observation data are lacking especially in high altitude mountainous area. Combined with rigorous screening of the measured data,reanalysis datasets( products) based on the data from meteorological satellite and the technologies numerical simulation were obtained. Due to its high temporal resolution,higher reliability,reanalysis datasets are widely used in research work.In this paper,we chose a typical alpine mountain watersheds located in the northwest as the object of study,and three representative reanalysis meteorological data products( CMADS,ITPCAS and TRMM 3B42) were used to simulate the annual and inter-annual distribution of precipitation in the study area. Firstly,we analyzed the time and space distribution of precipitation data from three datasets in study area. And then we used the precipitation data to drive the distributed hydrological model SWAT. The results of the model simulation were compared with the runoff data measured by the hydrological station,the simulated results of precipitation data based on the three datasets from CMADS,ITPCAS and TRMM 3B42 in the study area were analyzed. The results show that: The time and space distributions of precipitation data in the basin reflected by the three datasets exist remarkable differences,and the total precipitation data from CMADS and ITPCAS datasets are much larger than that from TRMM 3B42 dataset in the annual scale,the annual and inter-annual precipitation distribution from CMADS datasets is more consistent with the distribution of glacier than those from ITPCAS and TRMM3B42 datasets; For the three reanalysis data products,the simulated results of CMADS dataset driven hydrological model are better than the simulated results of ITPCAS and TRMM 3B42 datasets,and the month and day NSE efficiency coefficient of CMADS datasets is respectively reached to 0. 53、0. 71,the simulated results of IT