O B; B R
Abstract
Soil water retention curve (SWRC) is important in the studies of soil and water relationship, soil conservation, irrigation scheduling, drainage, solute transport, plant growth and crop water stress. The performance of 10 models (Simons et al., Libardy et al., Campbell, Farrell and Larson, van Genuchten, ...
Read More
Soil water retention curve (SWRC) is important in the studies of soil and water relationship, soil conservation, irrigation scheduling, drainage, solute transport, plant growth and crop water stress. The performance of 10 models (Simons et al., Libardy et al., Campbell, Farrell and Larson, van Genuchten, Brooks and Corey, Driessen, Exponential Bruce-Luxmore, Power Bruce-Luxmore, and Rogowski) were evaluated using SWRC 3.0 software. After statistical evaluation of the results, appropriate models were identified. Seventy three soil samples of tea fields in Lahijan city in Gilan province were selected and separated into soil textural classes. The soil moisture content at eight different matric potentials were measured. Using the statistical index of Root Mean Square Error (RMSE), Akaike Index (AIC) and coefficient of determination (R2), the accuracy of each model quality was identified. Results showed that in all textural classes the Simons et al. Model with a minimum RMSE (0.0136 to 0.0307 cm3 cm-3), the lowest value of AIC (-1280), and the highest coefficient of determination (0.96 to 0.98) had the best performance, while the power Bruce–Luxmor Model with the maximum RMSE (0.0372 to 0.0881 cm3 cm-3), the highest value of AIC (-879), and the minimum value of R2 coefficient (from 0.81 to 0.92) had the weakest performance. Libardi Model after Simons et al. Model had acceptable performance and better results than the other models. Overall, the Simons et al. and Libardi et al. models were recommend for determination of SWRC in Lahijan area soils.