DEM Dividing Regions Optimization Of China Land's Accumulated Temperature Spatial Interpolation And Precision Analysis

Accumulated temperature is the sum of the daily average temperature during a certain period of time. Being an important factor which affects the growth of plants, accumulated temperature is an index used to study the demands of energy for plant growth and evaluates heat resources. It is important for agricultural production and ecological restoration。As important climatic resources, it differs from other resources because of its great regional and temporal variety. It is practically meaningful to learn the spatial variety and distribution. Mainly making advantage of the data from 2384 meteorological stations located all over the country and 1km×1km DEM data, the article implements the interpolation of perennial average accumulated temperature based on explore research of the spatial distribution of accumulated temperature and dividing accumulated temperature regions. And then get the countrywide accumulated temperature distributed diagram.Spatial autocorrelation analysis implies that the distribution of accumulated temperature has comparely strong positive correlation on the whole region and applies with certain rules. To specify, the distribution of accumulated temperature has obvious character of spatial clustering. The strong spatial autocorrelation of accumulated temperature tells the fact that accumulated temperature strongly depends on large- scale and mesoscale geographic factor such as climate,terrain,physiognomy,atmospheric circulation. Meanwhile, it is proved feasible to acquire the accumulated temperature of one unknown location by the means of spatial interpolation.Spatial correlation analysis implies that accumulated temperature is notably connected with latitude and altitude. It is most closely correlated with latitude and less correlated with altitude while barely correlated with longitude. After the region dividing, accumulation temperature has decreased correlation with latitude and increased correlation with altitude in most accumulation temperature region. This indicates that the latitude zonality of accumulation temperature is , to some degree, demolished when the study area dwindles. At the same time, the vertical zonality caused by microscale terrain is better revealed.When geographic position and terrain are introduced to interpolation model as variables, the general tendency of the accumulated temperature's spatial distribution is better revealed. Meanwhile the local detailed characters of accumulated temperature's distribution are displayed and the holistic precision of the interpolation is improved.In this article, when dealing with interpolation of accumulated temperature, multiple regression method achieves the highest precision while ordinary Kriging method achieves less precision and IDW method the least.Through the dividing of accumulated temperature region according to clustering method, the interpolation precision achieved using IDW,ordinary Kriging,multiple regression method is, to various degree, improved。Dividing region by clustering is improved quite fitful to deal with accumulation temperature interpolation.Regarding large-scale study area in which there is great variety of spatial distribution of the element to be interpolated, it helps to improve the precision of interpolation if the regionalization is implemented according to the character of the data and choose corresponding interpolation model or method according to its own characters of every Interpolation region. Regionalize by clustering helps improve the prognostic precision of different interpolation method.Concerning complanate hypsography such as plain and hill, the interpolation errors of accumulated temperature don't have obvious change trends and regularity when the number of meteorological stations increase. Namely, increasing station density can't effectively ameliorate interpolation error. In the regions where terrain is complex the interpolation errors present degressive trend with the increasing number of meteorological stations. In other words, the number of stations will markedly effects the result of interpolation. Therefore meteorological stations should be increased in northwest where terrain is complex and meteorological stations are sparse.The stability of IDW, ordinary Kriging and multiple regression can be ordered as: multiple regression > ordinary Kriging > IDW.