Application of Snowmelt Runoff Model (SRM) in ungagged Manasi River Basin, Northwest China

Thursday, 18 December 2014
Xiaona Chen1, Shunlin Liang1,2 and Anming Bao3, (1)Beijing Normal University, Beijing, China, (2)University of Maryland College Park, College Park, MD, United States, (3)Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi, Xinjiang, China
The climate change will have great impacts on snowmelt runoff, including prolong of snowmelt period, advance of peak in snowmelt discharges, and the notable increase of snowmelt volume in melt-season. These changes have great significance and caused widespread concerns in arid region because the snowmelt runoff is one of the most important water resources which seriously related to the agricultural and socio-economic development. However, limited by the poor geographical environment and spare distribution of in situ observations, the snowmelt runoff simulation are still a challenge in some ungagged catchment. The Snowmelt Runoff Model (SRM) is one of the few models in the world that requires remote sensing derived snow cover as model input. Based on the SRM hydrological model, this study simulated the snowmelt process in melt-season of 2007 in Manasi River Basin and try to make contribution to the understanding of snowmelt runoff process in such ungagged catchment. The conclusions are as follows: the SRM model represented a certain semi-distributed model with physical mechanism, has a good applicability in scarcely meteorological and hydrological suits distributed catchments. The correlation coefficient between computed runoff and measured runoff could reach 0.93 with 2.57% volume derivation in ideal situation without the effects of instantaneous precipitation during March-Jun, 2007. As one of the most important input variables, the daily snow cover extent derived from MOD10A1 are effectively make up the lack of in suit snow observations. The classification accuracy of MOD10A1 reached 0.81 in 95% confidence level on 3×3 pixels statistical scale verified by the Landsat 5 TM images. Meanwhile, temperature and precipitation played important roles on snowmelt runoff simulation, the correlation coefficient between daily temperature and daily measured runoff is 0.46 in 95% confidence level. The lase rate of discharge determined the basic shape of computed runoff, while the precipitation and lag time inputs determined the peak size and peak position of computed runoff. There are also issues unsolved in this study, the threshold value of NDSI used in the MOD10A1 need to be further testified in the arid region, the effects of instantaneous precipitation on snowmelt runoff simulation also need to be further discussed.