A comparison of spatial interpolation methods for surface temperature in Thailand

Abstract

This paper aims to compare performance of two type of spatial interpolation methods for the estimation of surface temperature. Eleventh years of monthly mean surface temperature data at 120 stations located in Thailand have been used for this study. Inverse distance weighting (IDW) and ordinary kriging (OK) were used to interpolate surface temperature. Leave-one-out cross-validation (LOOCV) had been use to analyze the spatial error of interpolated data. The result of a Wilcoxon signed rank test indicated that the performance of ordinary kriging was significantly better (p value = 0.002) than the inverse distance weighting. The summary statistics show interpolation errors of the kriging for monthly mean surface temperature varying within 0.68-0.84.