Evaluating and Developing Pedotransfer Functions to Predict Soil Saturated Hydraulic Conductivity at Landscape Scale in Central Zagros

Soil saturated hydraulic conductivity (Ks) is an essential property in managing irrigation projects, water and solute transport modeling and for predicting unsaturated hydraulic conductivity functions. Laboratory and field measurements of Ks are time-consuming and costly especially in the large scales. Moreover, Ks is affected by different factors and thus its predicting and modeling by derived pedotrasfer functions (PTFs) or those in the literature can be useful for many purposes. This study was performedto evaluate the already existed in literature and derived PTFs for predicting Ks in Farsan and Koohrang regions, Chaharmahal-va-Bakhtiari province in the central Zagros. Undisturbed samples (volume of 100 cm3) were collected from the surface soil at 100 locations which were well distributed in the region and in different land uses. The Ks of the undisturbed samples was measured using the constant-head method in laboratory. Easily-available soil properties were also determined and PTFs for Ks prediction were derived by multiple stepwise regression. The measured Ks values were also compared with the predictions of several PTFs in the literature. Results showed that PTFs in the literature could not accurately predict the Ks in the region. Among the literature PTFs, powerful neural network model of Rosetta (SCC) was the best. The worse PTFs for Ks prediction in the studied structured soils were Brakensiek et al. (1984) PTFs. However, the derived PTFs predicted the Ks in the region better (R2=52%, RMSE=0.5 mm/h) than the literature PTFs. The findings revealed that logarithm of Ks negatively correlates with relative bulk density (relative compaction), effective bulk density, clay to carbonate content (Clay/CaCO3) and carbonate to organic matter content (CaCO3/OM) ratios and bulk density and clay interaction [ln (BD×Clay)] and has positive correlations with OM, ln(OM/Clay), very fine sand content and soil porosity. The studied soils had good structure and the structural properties had significant role in the derived PTFs. Therefore, it is suggested to include additional structural properties such as aggregate's mean diameter as a predictor for better Ks prediction in future.