Soil loss tolerance in calcareous soils of a semiarid region: evaluation, prediction, and influential parameters

Predicting soil loss tolerance ( T ‐value) as a first and crucial step in assessing soil erosion using pedotransfer functions (PTFs) could save time and cost. Therefore, this study aimed to evaluate T ‐value and its influential parameters for calcareous soils of the Dorudzan Watershed, Fars Province, Iran, and to develop PTFs for its prediction using easily measureable soil properties. T ‐value was determined in 60 soil profiles based on the soil thickness method. Various basic soil properties were used to develop five PTFs using stepwise multiple linear regression. T ‐value varied from 3.5 to 22.5 with a mean value of 10.4 t ha −1 yr −1 , which is approximately equal to the proposed value by US Department of Agriculture (11 t ha −1 yr −1 ). T ‐value in Inceptisols located in the dam shore and agricultural lands was significantly higher than that of Entisols located in the forest and rangelands. Organic matter (OM) with a correlation coefficient of .77 with T ‐value was the most influential variable in prediction of T ‐value. The PTF 5, which used OM, soil aggregate stability, and permeability, with the highest R 2 (0.72 and 0.79), the lowest mean error (−0.44 and 1.75 t ha −1 yr −1 ), the lowest root mean square error (2.98 and 2.75 t ha −1 yr −1 ) and the minimum Akaike information criterion (51.1 and 17.2) provided the best predictions for T ‐value. However, when considering the acceptable indices and easier measureable soil properties (OM, soil erodibility, and pH) in PTF 3, we concluded that it was the best model to predict T ‐value. Generally, the application of thickness method is recommended to determine T ‐value in semiarid regions.