Study On Snow Cover Variations And Snowmelt Runoff Modeling In The Yarlung Tsangpo-Brahmaputra River Basin

The snow cover over the Qinghai-Tibet Plateau(QTP)and its surrounding areas is very sensitive to changes in climate.Due to the complexity of the climate and geographical environment in this large region,the response of snow cover to climate change should exhibit spatial differences,but currently,they remain unclear.In this study,the spatiotemporal variations of snow cover from 2002–2015 in the Yarlung TsangpoBrahmaputra River Basin(YBRB)were analyzed using an op-timized high Asia Moderate Resolution Imaging Spectroradiometer(MODIS)snow product,a watershedwide SWAT hydrological model were established,the actual snow area decay curves of different areas in the watershed based on snow cover fractional and snow water equivalent data released by NSIDC were fitted and applied to the SWAT model in order to to optimize the snowmelt runoff simulation.Finally,we compared and analyzed the difference in the contribution of snowmelt runoff in different regions,the specific contributions of temperature and precipitation to snow cover changes in different parts of the basin were investigated.The results were as follows.The snow cover was unevenly distributed in the YBRB,and the snow cover rate is generally high in the east and low in the middle.The seasonal variation of the snow cover rate in different regions in the basin is significantly different,and the snowmelt period among the sub-regions is also not synchronized.The fractional snow cover(FSC)was high in the east and low in the west,translating to an average of 8.4% in YBRB.The FSC increased with elevation,and 86% of the snow cover was distributed in the 4500–6000 m altitude zone.The seasonal variations of the FSC differed obviously across regions,while the snow cover in the mid-downstream regions not only exhibited higher FSC but also lasted longer.Moreover,the climate in the maximum snow cover range(Max SCR)in the YBRB generally tended to be warmer(0.45 °C/10 a)and wetter(31.2 mm/10 a).However,the magnitude of climate change varied greatly across regions,causing the FSC to also change in response.On an annual scale,the FSC in the Max SCR of the YBRB showed a slightly increasing tendency(3.76%/10 a)during 2002–2015,but the changes exhibited large spatiotemporal heterogeneities.The FSC increased significantly in the source and midstream regions from winter to spring(10.5%–18.0%/10 a),mainly due to changes in precipitation.Slight changes in the FSC were observed in the upstream regions(-1.7%–1.6%/10 a),mainly due to temperature changes.Moreover,the FSC increased in the downstream region(4.4%/10 a)but was controlled by both temperature and precipitation changes.Typically,significant regional differences were observed in the FSC response to changes in temperature and precipitation,and the ranges of temperature-and precipitation-dominated areas also changed constantly across seasons.A SWAT hydrological model for the YBRB was established,and the snow area decay curve in SWAT model was revised based on snow observation data.The default snow area decay curve parameters of SNOCOVMX and SNO50 COV used in the SWAT model do not conform to the actual conditions of snow areas in the YBRB,and the characteristics of snow decay between different subregions in the basin are also big differences.The mean value of SNO50 COV in the source region(0.688)is significantly lower than other areas(0.760?0.805),and the mean value of SNOCOVMX(32.53?41.89)in the upstream subregion II-III is significantly lower than the mean value of other subregions(63.46?73.09).The optimized snow area decay curve parameters can be substituted into the SWAT model to obtain a better simulation effect.The PBIAS of the 5 stations in the YBRB are all controlled within ±10%,and the NSE is above 0.75.This shows that the optimized snow parameter scheme based on remote sensing snow data is helpful for accurately simulating the runoff process of the YBRB.According to the revised snowmelt parameters of each subbasin,the snowmelt runoff simulation calculation is carried out.The results show that the snowmelt runoff generation process in each sub-region in the basin has different characteristics: Snowmelt exists throughout the year in the source area,the peak of runoff is between April and May,and the low period is between November and January;the runoff process in Zone II?V in the upper and middle reaches is similar.Snowmelt runoff is mainly in the spring snowmelt period from February to May,followed by October to December,and there is basically no snowmelt runoff from June to August;The snowmelt runoff in the downstream area is mainly from March to May,but the change process of snowmelt runoff from October to July of the following year is gentler than that of other sub-regions.The peak of snowmelt runoff in the first three subregions(I?III)from upstream to downstream in the basin is in April,and the peak of snowmelt runoff in the last three subregions(IV?VI)is in March.Snowmelt runoff also has significant spatial differences in different river sections,it contributes the highest proportion of regional discharge(a multi-year average of 21.96%),and the lowest in upstream region III(1.91%).In the mid-stream and downstream region,although the total amount of snowmelt runoff recharge is the largest,the contribution rate is only 3.61%.The contribution rate of snowmelt runoff in the source and upstream area showed were increasing during 2001-2014,with the highest growth rate in the source area(4.72%/10a),but a downward trend in the middle and downstream areas(-0.34%/10a).These findings help reveal the different performance of the snow–water cycle mechanism in the QTP and its surrounding areas under climate change.