Geostatistical analysis of a water well field for determination of land management constraints

Abstract

Soil spatial variability and heterogeneity is a tough but very important matter in the field-scale description of soil properties, such as soil electrical conductivity, soil saturated hydraulic conductivity, and soil salinity. Geostatistics is a useful tool to study spatial distribution of soil properties and optimum sampling strategies in field. Estimating soil salinity, EC and Ks is a vital issue in soil fertility and management. Geostatistical methods, kriging and cokriging, were applied to estimate spatial distributions of the variables that were collected from a large size water well field for the surface soil, rather than entire bore-hole profile of the soil. The results suggested that estimation can be improved using cokriging , rather than kriging. Comparing to kriging results, cokriging reduced the mean squared error and improved the estimation of EC by 2-100% depending on cross-correlated variables. Using the cokriging prediction maps of the soil properties, the soil can be managed cell by cell with prescribed appropriate management strategies such as irrigation and manure application to mitigate soil salinity in the region.