Optimization of rain gage stations number based on interpolation methods and principal components analysis in Iran

Optimization the number of synoptic stations in the estimation of rainfall is an important step in terms of reducing the maintenance cost and saving in the data collection. The main objective of this study was to determine the optimal number of synoptic stations to estimate the amount of rainfall in Iran country. Accordingly, the amount of rainfall of synoptic stations related to a common 14-year period was received from the National Weather Service and the performances of five different interpolation methods were evaluated. Based on the results of radial basis function (RBF), with a margin of error of 0.63, this method was selected as the most appropriate method in fitting the data. Studies show by eliminating the synoptic stations in PCA method the estimation error of RMSE increases from 0.48 to 0.52 related to the case that all synoptic stations were used and in the radial basis function interpolation method decreases from 0.63 to 0.55 which indicates the suitability of this method in the optimization of synoptic stations. The results indicate that removing 34 and 22 points from the network of synoptic stations in Iran respectively in the PCA method and interpolation method of radial basis, the resulting error is acceptable.