The annual runoff simulation based on Pan-Kriging method of time domain

Abstract

Runoff that is time domain hydrological random variable should be time domain and non-stationary regionalized variables as the factors changes usually shows some trends. Universal Kriging method is one of main methods to deal with optimal valuation of non-stationary regionalized variables. The study builds time domain of universal Kriging forecast model based on geostatistics theory. Analyzing data measures of runoff variation structure, the model is applied to forecast annual runoff on Tuligen River Dahekou station of Luanhe River in the Inner Mongolia, and ordinary Kriging is also applied to forecast annual runoff. Comparative results show that the results with two methods also agree with measured values, and universal Kriging method is better than ordinary Kriging method.