ASSESSMENT OF METHODS FOR PREDICTING SOIL ERODIBILITY IN SOIL LOSS MODELING

Indirect methods are frequently used to determine soil erodibility (K-factor) because its direct assessment in the field is expensive. Predictive models, however, cannot be applied indiscriminately. The coherence between model outputs and soil features of specific localities must be checked as the models were developed for reference areas with specific features that may not correspond to those in the study area. Accordingly, the present study applied indirect methods to estimate soil erodibility in an area showing soil erosion and evaluated the results obtained for soil characteristics, observed K-factor values and bibliographic data. The soil in the Guariroba stream subwatershed (study area) is mostly loamy sand (Neossolo Quartzarenico ortico, an entisol). However, loam (Neossolo Quartzarenico hidromorfico, an entisol) and clay (Latossolo Vermelho distrofico, an oxisol) soils are also found in the area. The statistical evaluation of the models for K-factor estimation showed that the Sharpley & Williams (1990) model was the most suitable for the study area.