Applying Pedo-transfer Functions to Simulate Spatial Heterogeneity of Cinnamon Soil Water Retention Characteristics in Western Liaoning Province

摘要

In this study, by measuring soil water retention characteristics and fitting them to van Genuchten Equation, three types of pedo-transfer functions (PTFs) were estimated using linear regression (MLR3) and nonlinear regression (ENR3) based on three textural classifications, and using linear regression (MLR7) based on seven textural classification. And their abilities to quantify soil water retention characteristics and spatial heterogeneity in field scale were evaluated as well. By comparisons between fitted and PTFs-estimated parameters, the parameters from MLR7 and ENR3 were closer to fitted values than ones from MLR3. The results showed that some complex mathematical approaches such as nonlinear regression, and detailed soil classification may obtain more accurate PTFs. The analysis on spatial heterogeneity of retention characteristics using geostatistics indicated that PTFs-estimated parameters' semivariances were fitted a spherical model. Great autocorrelation range and proportion of structural variance for MLR3 and MLR7 shows linear regression could express suitably the spatial heterogeneity, but detailed soil textual classification didn't improve structure of variability.