Development of Some Pedotransfer Functions to Predict the Saturated Hydraulic Conductivity of Gravel Soils Using Partial Least Square Regression Method

Saturated hydraulic conductivity (Ks) is one of the most important physical properties of soils which is expensive and time-consuming to directly measure. Hence, indirect methods, such as pedotransfer functions (PTFs), were developed to predict the Ks. Previous studies showed that most of the PTFs common in the literature can not suitably predict the Ks. Hence, this study was conducted to develop some new PTFs. In this study, some physical properties of 49 gravel soils, including Ks, bulk density and particle size distribution, were measured in a land in the campus of ferdowsi university of mashhad. The measurements were performed in a regular quadrangular grid with 4 meters distances. To measure the Ks, inverse hole method was used. To derive some PTFs, 8 arbitrary sets of independent variables were selected. For each set, the best subset of independent variables was selected using best subset regression method. Then, this PTF was found using partial least square regression method. To evaluate the validity of the derived PTFs, we used cross-validation method. The results showed that the PTF that used d50, geometric mean and standard deviation of the particle size distribution as independent variables could more precisely predict the Ks. For this PTF, R2, RMSE, MAE and R2pred are 0.4, 0.245, 0208 and 0.3 respectively.