Applicability of GLDAS in the Yarlung Zangbo River Basin under Climate Change

The change of runoff has a great influence on global water cycle, and migration or transformation of biogenic matters. As the Tibet's most important economic region, the Yarlung Zangbo River basin is extremely sensitive and fragile to the global climate change. But the river is a typical lack-data basin, where the quantity of available runoff data is extremely limited and the spatial and temporal resolutions are very low. This study Chooses middle reaches of Yarlung Zangbo River basin as the study area, 4 models of Global Land Data Assimilation System (GLDAS) and the water balance equation are used to calculate surface runoff of Nuxia hydrological station from year of 2009 to 2013. Through the analysis of hydrological elements change, the impact of climate factors to surface runoff is discussed. At last, Statistical method is used to compare correlation and error between the 4 models results and in situ runoff observation. The Broke ranking method is applied to evaluate data quality and applicability of the 4 models in the Yarlung Zangbo River basin. The results reveal that the total runoff calculated from 4 models all have similar change cycle around 12 months, and the values all tend to have slight increase as in situ runoff data during research period. Moreover, it can conclude that the runoff time series show obvious period and mutation characters. During study period, monthly mean precipitation and temperature both have obvious seasonal variability, and the variation trend is relatively consistent. Through the analysis of the runoff affecting factors, it shows that the changes of precipitation and temperature are the most direct factors affecting runoff of the Yarlung Zangbo River. Correlation between precipitations, temperature with runoff of Nuxia hydrological station is good, and the correlation coefficients are in the range of 0.727 to 0.924.It shows that climate change controls basin runoff change to some extent. At last, runoff estimated from GLDAS-CLM can better represent runoff of the Yarlung Zangbo River basin than other 3 models with a total ranking score of 2.00. This paper carries out a helpful attempt on hydrological study in lack-data basin. And in the matter of medium and long terms, large and medium scales, the result is benefit to deepen cognition and comprehend on runoff characteristics.