Predicating the Spatial Variability of Some Soil Properties by Using Geostatistic Methods in Darreh Viseh, Karaj

Document Type : Research Paper

Authors

Abstract

Soil is a heterogeneous environment whose properties have spatial and temporal changes. Soil and vegetation have reciprocal relationship and influence each other. Thus, investigation of spatial variability of soil properties with other environmental change is essential. In order to analyze spatial variability of soil properties in Dareh Viseh rangelands, 78 soil samples from 0-30 cm soil depth were taken and transferred to laboratory. Soil properties including clay, silt, sand, and organic matter were measured in laboratory. After normalizing the data, the semi-variograms were obtained and evaluated. Then, kriging, inverse distant weighting, and radial basis function methods were evaluated for the obtained data. To compare these methods, the cross validation was used by statistical parameters of Mean Absolute Error (MAE) and Mean Bias Error (MBE).  The results showed that the kriging method provided a reasonable prediction for silt, sand, and organic matter of soil, while the radial basis function provided better estimation for prediction of clay in the study area.

Main Subjects