Hydraulic Conductivity Estimation Using Conveniently Available Soil Characteristics

Soil saturated hydraulic conductivity is one of the most important physical characteristics of soils that affects water movement in soil. . The aim of this study was to determine the most important parameters in prediction and modeling of saturated hydraulic conductivity from conveniently available parameters, using the decision tree and error estimator cross validation and re- substitution. In this study, 72 soil samples with six different textures were collected from the village of Morgmalek and Shahrekord District. Conveniently available soil properties were introduced into software in 4 scenarios (the first scenario: pH, EC, % sand, % clay, OM%, CaCO3, mean weight diameter of dry aggregate (MWD dry), mean weight diameter of wet aggregate (MWD wet), BD; the second scenario: CaCO3 , OM%, % sand, % clay, BD, % gravel, electrical resistivity, dielectric constant, root penetration resistivity; the third scenario: pH, EC, Geometric mean diameter )dg(, Geometric standard  deviation )σg(, OM%, CaCO3, mean weight diameter of dry aggregate (MWD dry), mean weight diameter of wet aggregate (MWD wet), BD; and the fourth scenario: CaCO3, OM%, dg, σg BD, % gravel, electrical resistivity, dielectric constant, root penetration resistivity). Saturated hydraulic conductivity was measured with single ring. The results showed that moisture followed by structural features such as (σg) in the first scenario, and OM and BD in the second scenario, BD and MWD in the third scenario, and OM and BD in the fourth scenario were the most important parameter affecting saturated hydraulic conductivity. Correlation between predicted data by decision tree and measured data in the second and fourth scenarios were 0.83 and 0.82, respectively, and 0.79 in the first and third scenarios. All four scenarios were successful in modeling with respect to error rate and %RMSE. However, the %RMSE in the second and fourth scenarios was lower and the correlation coefficient was higher than the other two scenarios. The RMSE values in the four scenarios were 0.79, 0.83, 0.79, and 0.82, respectively.