Comparison of Some Approaches to Determine Spatial Dependence of Soil Properties

Abstract

Knowledge of variability and spatial structure of soil properties is essential for optimal design for collecting soil samples and effectively applying management decisions in the field. The objective of this study is to compare some approaches for characterizing, and comparing spatial dependence of isotropic second-order stationary processes. The evaluated approaches are the nugget to sill ratio (NR), normalized (by fitted sill) semivariogram, correlograms, and two integral scales. Soil samples, collected at a regular 50 m × 50 m grid from 0-15 cm depths, were analyzed for sand and clay, bulk density (b), saturated hydraulic conductivity (Ks), wilting point, available water content (AWC), pH, electrical conductivity (EC), nitrate-nitrogen (NO3- N), and chloride (Cl) were determined. Geostatistical software (GS+, Gamma Design Software, Plainwell, MI) was used to estimate the variance structure of various measured soil properties. Analysis include using data on the spatial variability of various properties from four published studies. NR displayed spatial dependence ignoring the influence of range, normalized semivariogram and correlogram provided the visual comparison, and both integral scales incorporated the influence of range and provided single number spatial dependence summaries. Either of the integral scale formulations can be used to characterize the spatial dependence of soil properties from agricultural fields.