Extended analysis of unsaturated hydraulic functions using the RETC code

الملخص EN

Ten descriptive models ; two describe the water content as a function of matric potential θ (ψ) ], four describe unsaturated hydraulic conductivity as a function of water content [k (θ)], and four describe unsaturated hydraulic conductivity as a function of matric potential [k (ψ)] ; were used by the RETC optimizing program to analyze seven-observed different data sets for predicting or estimating unsaturated hydraulic characteristics.

The seven data sets are : 1) all laboratory measured θ — W data, 2) all laboratory measured θ — W data with one k — θ matching data point, 3) all laboratory measured θ — W data with one k — ψ matching data point, 4) all laboratory measured θ — W data with all k — θ data, 5) all laboratory measured θ — W data with all k — ψ data, 6) one θ — W matching data point with all k — θ data point and, 7) one θ — W matching data point with all k — ψ data; where 6, ψ, and k indicate volumetric water content, applied suction or pressure head, and k is field determined instantaneous-unsaturated hydraulic conductivity.

Highly significant correlation values, R > 0.99, were obtained from fitting of van Genuchten and Brooks and Corey models to laboratory measured # ψ) relationship.

Predicted k (θ), and k (ψ) functions from soil water retention compared very well with observed data with linear correlation exceeded 0.800 and regression coefficient approached unity on 1 : 1 line.

When one k (θ), or k (ψ) matching point was used in the analysis besides 0 (W) data, the k (θ), and k (ψ) functions were more precisely predicted.

The linear correlation between matched and observed k (θ), and k (ψ) functions exceeded 0.959 with a regression coefficient of 1.0158 on 1 : 1 line.

Estimated k (θ), and k (ψ) functions obtained from analyzing both the retention and conductivity data matched the observed values with correlation coefficient of 0.937 and a regression coefficient of 0.9888.

Predicted θ (ψ) relationships obtained from analyzing observed k (θ) besides one θ (ψ) Matching point underestimated observed data especially near saturation suggesting the susceptibility of the prediction to high water content values.

Predicted 0 (ψ) values obtained from analyzing the k (ψ) besides one 0 (ψ) matching point compared precisely with observed θ (ψ) values.

When all predicted θ (ψ) data were plotted against measured θ (ψ) data on 1 : 1 line, it produced correlation coefficient value of 0.846 and a regressing coefficient value of 0.9308.