Abstract
Soil moisture retention is one of the most important soil physical properties for hydrological analysis and for determining irrigation water requirements for agricultural crops. However, because measuring soil moisture retention at multiple matric potentials is time consuming and labor-intensive and pedo-transfer functions (PTFs) based on soil properties is widely used, soil moisture retention is estimated indirectly. Although PTFs are influenced by various soil characteristics and exhibit considerable variability, this variability has rarely been investigated, particularly for Korean soils. This study was aimed to develop and evaluate the new-PTFs at 10, 33, and 1500 kPa of soil moisture retention (θ 10 , θ 33, and θ 1500 ) for Korean soil using multiple linear regression model for soil samples ( n = 1,302 and 138 for developing and evaluating, respectively) across Korea. The estimated soil moisture retentions values showed positive correlations with the measured values of soil moisture retention of new-PTFs ( R 2 = 0.78, 0.78, and 0.65 for θ 10 , θ 33, and θ 1500 , respectively) compared to previously established model. The RMSE and MAE values for the new PTFsranged from 4.83 to 6.16 and from 3.37 to 4.86, respectively, by showing the lowest values compared to other models. The values of NSE were above 0.5 for all, except for θ 1500 (NSE = 0.21), indicating generally accepted model performance between estimated and measured values. The developed new-PTFs at θ 10 , θ 33, and θ 1500 offer improved prediction of soil water retention for Korean soils than previously established models, providing a valuable tool for soil water management and agricultural applications.