Snowmelt Snow Depth Changes Impact Factor And Simulation Studies

Snow as an important freshwater resources, not only the most active surface of multipleattributes of natural factors and the response factor is most sensitive to changes in the environment.The snowmelt brought ablation as an important freshwater resource reserves, plays a veryimportant role in the rational use of water resources in this arid area in Xinjiang. Snowmeltprocess is a complex physical processes, snow depth changes as one of the key feature of thesnowmelt process variable, in addition to the temperature factors, but also with many other factorsare closely related.In this paper, the Ili N River basin for the study, analyzed snow depth changes in differentperiods of the key factors affect snowmelt snow depth exploration and research, the use ofprincipal component analysis of these the factors extracted principal component, principalcomponent will be extracted as a the next snow depth simulation model input factor, to simulatethe snowmelt period of snow depth, and a principal component analysis and neural networkcombined snowmelt period snow depth simulation model. The important conclusions of theresearch are as follows:(1).In N Ili River, the snow began to accumulate from the end of October each year, the peak inFebruary, followed by a gradual melting, disappeared at the end of March-early April. SnowDepth in a plot, the snowmelt the period presented a variation of the "n" type.(2).Established the snowmelt period of snow depth and hydrometeorological factors multiplelinear regression equation by T-test, is not obvious that the various parameters in the regressionequation significantly. Need to use other methods to further exclude collinearity of meteorologicaland hydrological factors, and then to simulate snow depth.(3). In this study, principal component analysis principal component extraction snow depth ashydrological factors affecting the snowmelt period, extracted three principal components, heatingredients, water vapor content and air flow ingredients. Further analysis showed that thedominant heat ingredients (air temperature, net radiation, surface temperature), followed by watervapor component (water vapor pressure, relative humidity, rainfall), and the minimum impact onthe air flow component (average wind speed).(4). Based on principal component analysis of the BP neural network and RBF neural networksnowmelt period snow depth model comparative analysis can be seen, based on principalcomponent analysis to establish the RBF neural network structure is simple, good stability, and ismore suitable for the simulation snowmelt snow depth changes.