Developing Pedotransfer Functions for Saline and Saline-Alkali Soils



Soil moisture retention curve is one of the soil hydraulic prosperities which its direct measurement is time consuming and expensive. Therefore, indirect methods such as developing pedotransfer functions have been used to predict this characteristic from soil readily available or easily measurable data. In this study, multiple linear regression method was used to develop point pedotransfer functions (PTFs) for saline and saline-alkali soils of Iran. For this purpose, 68 soil samples with EC values greater than 4 dS/m of which more than half of them had ESP values greater than 15% were selected. Using Cross validation method, the random splitting of data into the development and validation subsets was repeated 10 times. A ratio of 3:1 was used to split data into development and validation sets in each replication. In the SPSS software, parameters such as geometric standard deviation (?g), geometric mean diameter (dg), sodium adsorption ratio (SAR), electrical conductivity (ECe), carbonate calcium (CaCO3), bulk density (BD), organic matter (OM), clay and sand content were applied as the independent variables, and volumetric water content was determined at matric potentials of 10, 33, 100 , 300, 500, 1000, 1500 kPa. The derived PTFs were compared with the H3 model of Rosetta software for 10 splits of validation data set. Comparison of the mean RMSE and R2 values showed that the developed PTFs resulted in more accurate estimation than the Rosetta software at matric potentials of 100, 300, 500, 1000, 1500 kPa. Whereas, Rosetta model resulted in slightly better estimation than derived PTFs at matric potential of, 33 kPa. For the PTFs developed in this study, the RMSE, ME and R2 values ranged from 1.61 to 4.17 (, -0.181 to 0.66 and 0.52 to 0.76, respectively. While for the Rosetta model, RMSE, ME and R2 values ranged from 2.82 to 4.27 (, -2.65 to 0.094 and 0.37 to 0.74, respectively.