Combining topographic and soil attributes for spatial prediction of model parameters of soil shrinkage curve

This study was conducted to investigate the spatial variability of model parameters of soil shrinkage curve in Cherlgerd district, Chaharmahal-va-Bakhtiari province. Ninety nine locations were selected and soil clods were collected from the 0-10 cm layer and brought to the laboratory. Soil shrinkage curve was measured, according to Archimedes's principle, by covering the undisturbed clods with an Acrylic Resin which was slowly permeable to water vapor and impermeable to liquid water. Then, the void ratio (e) values were plotted vs. the moisture volume ratio (υ) values. The shrinkage curve data were modeled using Peng and Horn (2005) model and the model parameters (er, es, α, n and m) were optimized for each soil. Spatial structures of the model parameters were identified by the means of Moran's I index. Using ordinary-kriging (OK) and regression-kriging (RK) methods showed that for spatial prediction of er, es and α, the RK method was superior when compared to the OK method. However, the RK was not successful for spatial prediction of n and m; this could be due to low coefficient of determination of the regression equations relating parameters n and m to auxiliary soil and topographic attributes Overall, the results of this research indicated that soil shrinkage model parameters possess spatial variability in the study region and the obtained spatial information could help sampling strategy for modeling soil shrinkage in future. Moreover, the results might be useful for precise spatial management for engineering purposes and agricultural and tillage fundamental applications in the region.