Development of Pedotransfer Functions for Estimating Soil Retention Curves and Saturated Hydraulic Conductivity in Jordan Valley

Two hundred disturbed and undisturbed soil samples were collected from three depths: 0-15, 15-30 and 30-60 cm. Particle size distribution, organic matter, alkalinity, electrical conductivity, and saturated hydraulic conductivity were determined for each soil sample. The soil water retention curve for each soil sample was measured at matric potentials of 0, -20, -40, -60, -80, -100, -330, -500, -1000, -3000, -5000, -10000, and -15000 hPa. The developments of PTFs for soil hydraulic properties were done using backward multiple regression analysis. The performance of all developed PTFs showed that as more input variables were included, mean error (ME), mean absolute error (MAE), and root mean square error (RMSE) were decreased; intercept and slope also of the linear regression analysis were become closer to zero and one, respectively. The developed PTFs were evaluated using 30 independent soil samples which were not used in PTFs development. The regression coefficient, ME, MAE, RMSE, intercept and slope were relatively close to the developed PTFs. Additionally, the correlation between predicted and measured properties were not significantly different at 0.05 level. It was noticed that the developed PTFs performed well in predicting soil hydraulic properties. As a result, this suggests that the developed PTFs can be used to estimate soil hydraulic properties using the basic soil properties instead of using available PTFs which is estimation and performance should be assessed as the soils used in its development were relatively dissimilar from our soils.