西藏地区积雪深度模型分析与研究

Analysis and Research on the Model of Snow Depth in Tibet Area

摘要: 本文利用西藏地区35 a地面逐日观测资料对西藏地区积雪深度建立模型：以日降水量、平均气温、最低气温、最高气温和积雪深度、数值预报次日降水量和平均温度为影响因子来建立预测次日积雪深度的多元回归方程(由于缺少数值预报资料，所以用第二天实际观测资料代替数值预报资料)。选取平均积雪最大的五个站点画出散点图并求出相关系数(R)和标准偏差(SEE)进行检验分析，后通过改进方法分季节建立积雪深度模型再进行分析，对比改进前后的拟合效果，发现改进后所得方程相关系数(R)有所增加，标准偏差(SEE)有所减小，这都说明改进后的方法得到的模拟值更接近于实际值。

Abstract: In this paper, by using 35-year daily observation data of ground, it establishes model for snow depth in Tibet. By the use of impact factor such as daily precipitation, average temperature, the lowest temperature, the highest temperature, snow depth, numerical forecast precipitation of the next day, and average temperature, it establishes multiple regression equations that predict snow depth of the next day. Due to the lack of numerical forecasting data, it uses the observed data of the next day instead of numerical forecasting data. It selects five sites in which daily average amount of snow is largest to draw scatter diagram, and obtain correlation coefficient (R) and standard deviation (SEE) to test. Then it establishes the seasonal model of snow depth by improving method and analyses. Comparing imitative effect before and after improvement, it shows that required correlation coefficient (R) of equal has increased and standard deviation (SEE) has decreased after improving. All of the above descriptions’ mean analog value after improving method is much closer to actual value.