Comparison of Spatial Interpolation Methods on Meteorological Element of the Beijing-Tianjin-Hebei region

摘要

Spatial interpolation methods in common use primary include inverse distance weighted (IDW), spline function (Spline), Kriging and overall polynomial interpolation. Based on the data of the meteorological monitor data at 90 stations in the Beijing-Tianjin-Hebei region from 1961 to 2007, which involve precipitation, mean temperature, duration of sunshine of each month, the calculation according to different interpolation methods were generated. In order to investigate the optimization methods, the results were used to compare with the meteorological monitor data at the other 10 stations that did not participate in former interpolations. After the error analysis, spline method Tension is the most accurate model in the results of the monthly precipitation interpolation, IDW secondly. For example, the relative error of the stations that did not participate in the analysis is 8.61 in Tension while 9.03 in IDW on precipitation interpolation in January. So the Tension is the best method to calculate the precipitation in the region based on the monitor data. Overall polynomial interpolation has the highest accuracy in the results of the monthly average temperature and monthly duration of sunshine interpolation. For example, in overall polynomial interpolation the relative error of the monthly average temperature and monthly duration of sunshine interpolation is 10.82 and 2.02 in January. Obviously, overall polynomial interpolation can be used to obtain the raster data around the station. It can be improved to consider the region as a combined part which can be divided in to mountainous region and plain. In this way, local polynomial interpolation can be used to compare with other interpolation method. By choosing the most precise method that reflects the actual distribution, estimated value of each pixel was obtained in the study area which render agroclimate data basis of various kinds model on growth of plants and other applications and is useful to the development of precision agriculture.