Research On The Density Optimization Of Automatic Weather Station

Automatic weather station is applied to the meteorological service because of its accurate meteorological data, high spatial resolution and low business costs. Reasonable and moderate density of automatic weather station can make sure that the interpolation of the observation of meteorological elements is fully accurate and representative in all regions.In order to ensure a certain accuracy of interpolation in these areas, automatic weather station network must meet the appropriate density. For example,if automatic weather station network is too close, it can lead to the waste of human, material and financial resources, if automatic weather station network is too thin, it will lead to distortion of data analysis, errors of judgment, incorret forecasts and other adverse consequences. So, the main research content of this paper is to achieve the necessary interpolation accuracy in less station network density and much more sufficient amount of information.Given the above, to work out the problem of reasonable density of automatic weather station, spatial structure function is used as the criterion for the design as well as interpolation method and SVM.Considering of the characteristics of China’s terrain and population distribution, this paper selects the data of daily air temperature, daily relative humidity and monthly precipitation from2002to2010over three plain areas in China(Yangtze River Plain, the North China Plain, Northeast China Plain), the spatial structure function is used to estimate maximum admissible spacing between stations. The article adds different stations with the changing of interpolation accuracy(l=16km、l=12km、l=8km). The meteorological data of the adding staions is obtained by the best interpolation method of triangulation of linear interpolation, cubic curve interpolation, biharmonic spline interpolation and kriging interpolation.Finally, this paper uses the method of SVM to process the data of simulation station and optimize the model of distribution of the automatic weather stations. The final goal of this paper is to achieve the purpose of ensuring the amount of information and cutting the expenses.