A Comparison Between Empirical and Fractal Models Fitted to the Measured Soil Moisture Characteristic Curve Data

Soil moisture characteristic curve (SMCC) is an important hydraulic properties in modelling of water movement and solute transport in unsaturated zone of soil, which its direct measurement in a laboratory is expensive, time-consuming and laborious. Therefore, the objective of this study was to compare the fitting capabilities of several SMCC fractal (Tyler and Weatcraft, 1990; Rieu and Sposito, 1991; Perfect, 1999 and Bird et al., 2000) and empirical models (Brooks and Corey, 1964; Campbell, 1974 and van Genuchten, 1980) to the observed data and to select the appropriate models. For this purpose, fifty-four soil samples were randomly collected from Ghorveh-Dehgolan plain and some important physical and chemical properties of soils were measured in the soil laboratory. The SMCC fractal and empirical models were fitted to the measured data based on the least esquare error approach using the solver toolbox of EXCEL software. In order to evaluate fitting quality of the proposed models, three statistical parameters including coefficient of determination (R2), root mean square error (RMSE) and Akaike information criterion (AIC) were used. The results indicated that all models did well with R2 ranging from 0.75 to 0.99. Bird et al., (2000) and van Genuchten (1980) models are selected as the best models, respectively, based on the highest R2 and the lowest RMSE and AIC. Rieu and Sposito (1991) model was the weakest, although the average R2 was greater than 0.75 and the average RMSE was smaller than 0.071 cm3/cm3.