Study On Spatio-temporal Distribution Of Snow Cover In Northeast China And Its Simulation On Snowmelt Runoff

Northeast China is the mian production base for agriculture, forestry and animalhusbandry, and also is among three biggest snow cover area that seasonal snowfall isprevalent and steady. The winter in this area is cold and permanent, and lasts as longas half year in the northern part. On the one hand, snow cover is important waterresource and recharge, seasonal snowmelt runoff could cover10%-15%of discharge;on the other hand, snow cover is among major source of natural distaters, heavysnowfall in the winter could result in snow storm and damage and sudden snowmeltrunoff improve the chances of flood. Therefore, study on snow dynamics andsnowmelt ruoff simulation has realistic meaning for regional water sourcemanagemetn, agriculture and husbandry development, and distaster prevention.Due to high cloud coverage of daily MODIS Terra MOD10A1and MODIS AquaMYD10A1, the daily and multidy combination is designed and inplemented, and dailyproduct MODISDC and multi-day product MODISMC of Northeast China and fourbasins are generated, after which cloud cover fraction decreased9.96%and46.69%respectively. The average combination day of MODISMC in Northeast and basin is3-4day and2-3day, respectively. Five accuracy index are established and used forevaluating product quality based on confusion matrix of snow cover classification ofstation and satellite, and results show that the overal accuracy of MODISDC andMODISMC is47.51%and76.52%, higher than that of MOD10A1(34.45%) andMYD10A1(30.75%).Using flexible multi-day snow cover combiantion MODISMC, snow coverfraction (SCF), snow cover day (SCD) and snow cover index are calculated in Northeast China. SCF has increased by0.22%every year since HY2003, and thesnow cover area has increased by22.57km2every year. The snow cover lasted forlongest time in Changbai Mount and Xiao and northern Da Hinggan Moutain, morethan six months; followed by Xiao Hinggan Moutain that has snow cover duration of3to5months. Both SCF and SCD in Northeast China exert an increasing trend inpast12months, the snow cover distributed most widely and last for longest time inHY2013, and the snow distribution in HY2003and HY2008is opposite to HY2013.There exist a negative correlation between air temperature and SCD (r=-0.60, P<0.05)and a signifiacnt positive correlation between precipitation and SCD (0.78, P<0.05).The spatial and temporal pattern of snow cover have been studied from threeaspects, including latitude, topograph and vegetation type, and result shows that: SCDrelated to latitude positively (r=0.91, P<0.05), and snow changes with the latitudeobviously; elevation is most important factors among three topographic factors.Calculating the correlation coefficient between NDVI and SCD, the influence ofvegetion on snow cover varies as the elevation varies, a trasition elevation zone isfound between1200-1600m, snow cover related closely with vegatation whenelevation is less than1200m, and vegetation has little effects on snow cover whenelevation is higher than1600m.Four basin including Liaohe Basin (LH), Nenjiang Basin(NJ), upperSonghuajiang Basin(SU) and downstream Songhuajiang Basin(SD) are selected andthe MODISMC are generated for each basin. SCF, SCD and snow cover index arecalculated for comparing spatial and temporal changes of four basins, and followesthe order decreasing from SD, SU, NJ to LH. NJ show the fastest increasing amongthe four basins, snow cover area inceased9.44km2per year and SCD increased4.48day per year. Using classification and regression tree (CART), the influence ofvegetaion, topographic and climate has been studied and ranked, and it is found thattopography and vegetation is major factors that influences snow cover changes inNortheast China. Erdao-Songhuajiang Basin located in SU basin are selected as studyobject, the hydrological and climate data are collected from March to October of2010. Snow delpetion curve is derivied from MODISMC, air temperature and precipitationare interploated by the method of krigring for each elevation zone, Snowmelt RunoffModel(SRM) uses daily runoff from hydrological station and run successfully, andthe result show that R2=0.57and Dv=25.59%. The simulated runoff could reflect microdevelopment of snowmelt runoff and the runoff features.