Evaluation of Different Models for Predicting Soil Water Retention from A Semiarid Region of Duhok

Abstract

A study was carried out to examine the capability of bi and uni- models to explore water retention data for two different textured soils from a semiarid region Duhok-Iraqi Kurdistan. A soil hydraulic model, can be used to derive the parameters of soil hydraulic properties for describing soil water movement. The program "SWRC Fit," which performs nonlinear fitting of soil water retention curves for the two examined soils using six models, was employed for this purpose. The six models are the Brooks Corey, van Genuchten, Kosugi, Fredlund and Xing model as uni models and Durner and Seki as a bimodal model used for this purpose. The program can be carried out straightly from a web page at http://purl.org/net/swrc/; The program was used for determining Fifteen parameters of soil hydraulic for two different textured soils. As related to Berderash sandy loam soil the LN uni-model revealed better performance than the FX, BC and VG models and at same time the FX, BC and VG models exhibit similar fitting precision on average. Whereas bi-models is shown that the Durner’s bi- model (DB) revealed very well fitting performance than that of the bi-log-normal distribution model (BL) models, regarding to Zawita clay soil it revealed that the uni -models, of BC, VG, FX and LN revealed good fitting performance with similar fitting precision on average. Whereas in bi- models it was noticed that the log-normal distribution model (BL) gave very well-suitable performance than that of the Durner’s b- model (DB). After comparing between the uni-models and bi-models of the six mentioned soil hydraulic models from accuracy and coefficient determination (R2), for two different textured soils under the study it can be illustrated: firstly, for fine texture soil (clay soil) the most precision fitting performance was found by model (BL), whereas for light texture soil (sandy loam soil) the most very well-fitting performance was noticed by (DB) model, and secondly the bi-models hydraulic were better than uni-models in fitting performance.