The Study On Multivariate Spatial Interpolation Method Of Precipitation In Mountainous Area

The spatial distribution of precipitation is generally realized by its spatial interpolation, because the factors that affect precipitationin the mountains is much and miscellaneous and meteorological stations is little and the distribution is not uniform, general spatial interpolation method often exist deviations. Therefore, the study effecting forthe main factorsof precipitation in the mountains and methods suiting for spatial interpolation of mountainous area is of great significance.Many factors affect the spatial distribution of precipitation in mountainous areas,such as latitude and longitude, topography, vegetation,wind direction, and etc. We base on the annual precipitation data of 10 years and 30 rain-gauge station in Red River Basin and its surroundings, by means of multiple regression analysis method to select the influence factors of precipitation in mountainous area, and study the spatial distribution by using cooperative Kriging method, and compared with the interpolation method; Meanwhile, the precipitation of the study area was analyzed by cluster analysis to partiteand do mixed-interpolation, and then the spatial distribution of precipitation in the study area was analyzed by interpolation. The results show that:(1) Main affecting factors on precipitation in the study area are elevation, latitude, slope and slope direction, main wind direction effect index.Among these factors, the main influencing factors of the seasonal variation are elevation, latitude, slope and aspect;the main influence factors of wet season are elevation, latitude, slope, main wind direction effect index;the main influence factor for the dry season, latitude, elevation gradient and slope direction;(2)The result of the interpolation with cluster analysisand partition is higher than the direct interpolation with non-cluster analysis. For years of average annual precipitation, inarea 1, we use the co-Kriging method with introducing elevation, slope and latitude factors, in area 2, we use the co-Kriging method with introducing elevation, slope and slope aspect, in area 3, we use the inverse distance weighted method which has smallestdeviation and highest accuracy; For wet season precipitation, in area 1, we use the co-Kriging method with introducing elevation, slope and latitude factors, in area 2, we use the co-Kriging method with introducing elevation, latitude and PWEI main wind direction effect index, in area 3, we use the inverse distance weighted method which has smallestdeviation and highest accuracy; For dry season precipitation, in area 1, we use the co-Kriging method with introducing elevation, slope and latitude factors, in area 2, we use the co-Kriging method with introducing elevation, latitude and slope aspect, in area 3, we use the ordinary-Kriging method which has smallestdeviation and highest accuracy.