Abstract
This study was conducted to derive the relationships of soil shrinkage parameters and indices with soil and environmental variables in calcareous soils. Ninety nine undisturbed clods were collected from surface soils in hilly regions of Cherlgerd, western Iran. Soil shrinkage curve was measured based on Archimedes' principle, by covering the clods with an acrylic resin. The shrinkage curve data were modeled using Peng and Horn (2005) model. The model's fitting parameters and several shrinkage indices (i.e. relative void ratio changes, mean slopes at various shrinkage zones, coefficient of linear extensibility, and total and relative shrinkage capacities) were predicted using multiple linear regression models by including soil properties (pedotransfer functions, PTFs) and by combination of soil properties and environmental variables (soil spatial prediction functions, SSPFs) as inputs. The results showed that, on average, the structural, proportional, residual and zero shrinkage zones comprised 17.2, 66.2, 15.2 and 1.4% of total shrinkage for the studied soils. The shrinkage capacity (ShC) and relative shrinkage capacity (Δetotal-rel) varied, respectively, in the ranges 0.204–0.641 and 0.288–0.589 in the studied soils. While clay fraction increased the ShC and Δetotal-rel, organic matter had a diminishing effect on the Δetotal-rel. An extended structural zone was observed in fine-textured soils, presumably due to greater aggregation. Volume change in the structural shrinkage zone was greater in weakly-structured calcareous soils because carbonates would minimize resistance of aggregates against the shrinkage forces. PTFs could explain 12–48% of variability of the model's parameters, and the inclusion of topographic attributes (i.e. SSPFs) significantly increased R2 values. Developed PTFs could explain 11–41% of variability of the shrinkage indices. The particle size fractions and relative bulk density were identified as most important soil properties for the prediction of shrinkage indices. Overall, the use of SSPFs by including topographic attributes such as dispersal area, elevation, surface curvature and plan curvature and normalized difference vegetation index (NDVI) could improve the performance of the prediction functions for soil shrinkage indices.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have