Analysis Of Spatial And Temporal Variation In Snow Cover On The Qinghai Tibet Plateau

Snow cover is a very active natural factor on the earth’s surface,and the change of snow cover can reflect the climate change.An appropriate amount of snow will promote the balance and stability of the ecological environment,while excessive snow will increase the risk of snow disasters and other natural disasters.In addition,the snow cover on the Qinghai Tibet Plateau also has a direct impact on the river flow of inland rivers and the lives of residents along the river.Therefore,the study of snow cover on the Qinghai Tibet Plateau has important practical significance.Firstly,based on the MOD10A2 data of the Qinghai Tibet Plateau from 2012 to 2022,this thesis uses descriptive statistical methods to simply analyze the spatial and temporal distribution characteristics of snow cover.The analysis shows that there are significant spatial differences in the distribution of snow cover on the Qinghai Tibet Plateau,with the number of snow cover periods in most pixels fluctuating between years.Over the past 11 years,the snow cover rate on the plateau has shown a significant periodicity and a weak downward trend.Based on the analysis results,it is determined to use cluster analysis to make a more precise and accurate division of the Qinghai Tibet Plateau,and further study the spatial distribution characteristics of snow cover on the plateau.The time series model is used to study the characteristics of snow time change in various snow covered areas,and predict and analyze them.Secondly,the clustering analysis algorithm is introduced to further accurately study the spatial distribution characteristics of snow cover on the plateau.Based on the annual snow cover period data of 12322194 pixels on the plateau,in order to obtain more accurate and reasonable clustering results,K-Means,Mini Batch KMeans,and BIRCH algorithms are used to cluster the pixel points.ANOSIM analysis is used to test the clustering results,and the clustering results are compared with historical research.It is found that pixel points on the Qinghai Tibet Plateau can be divided into five categories each of which can be regarded as a snow covered area with common characteristics.Among them,the stable snow area accounts for 8.13%of the study area,and is located in the high-altitude area in the northwest of the plateau.The average snow cover days is about 320 days.The long day snow area accounts for 12.62%,which is distributed around the annual snow covered area and Nyenchen Tanglha Mountain in the southeast of the plateau.The average snow covered days is about 224 days.The medium day snow area accounts for 18.56%,which is located in the central mountainous area of the plateau.The average snow covered days is about 160 days.The short day snow area accounts for 27.74%,which is mainly distributed in the central hinterland of the plateau,and the average snow covered days is about 80 days.The unstable snow area accounts for 32.95%,which is distributed in the northern basin of the plateau,the valley of southern Tibet and the dry and warm valley of the southeast.The average snow cover days is about 24 days.The pixel points in the various snow covered areas obtained in this thesis are densely distributed at the cluster center and sparsely distributed at the edge,and there are significant differences between classes.Finally,the time series analysis model is introduced to study the temporal variation characteristics of snow cover in various snow cover areas.Based on the clustering results,the snow coverage rate of each type of snow cover area is calculated with an 8-day cycle,and its periodic change characteristics are analyzed by establishing a piecewise function model.The Mann-Kendall trend test is used to judge the long-term change trend of snow coverage rate,and a linear regression model is established as the trend component of the series.Make a white noise test on the residual part of the trend component removed.If it passes the test,further judge its normality.If it fails the test,fit the AR model to the series,and make a normal white noise test on the model residual.Based on the modeling results of each component,establish a time series addition model for the snow coverage of five types of snow covered areas,and use the model and the Monte Carlo simulation of the normal white noise series to obtain the snow coverage prediction value.The analysis shows that the snow cover rate in various snow covered areas is based on a one-year cycle.The overall fluctuation range of snow cover rate in the stable snow area and the unstable snow area are not significant,maintaining a high coverage level of 0.75-1 and a low level of 0-0.15,respectively.The other types of snow covered areas are larger and the long day snow area is greater than the medium day snow area is greater than the short day snow area.In addition,the snow coverage rate in various snow covered areas has a slight downward trend in recent years,with significant decreases in the stable snow area and the unstable snow area.The time series model established for predicting snow cover rate data in various snow covered areas has shown good performance.The model established in this article can be used to predict and analyze the snow cover rate in different snow cover areas for different problems,effectively addressing the impact of snow cover changes.