Comparison of fallout radionuclide (caesium-137) and modelling approaches for the assessment of soil erosion rates for an uncultivated site in south-eastern Australia

Soil erosion rates are quantified using the fallout radionuclide (caesium-137) approach and models (empirical RUSLE and physically based SIBERIA) for a small catchment in south-eastern Australia. Two hillslope transects (under native grass) were sampled for 137Cs activity and soil redistribution rates were determined using empirical and theoretical conversion methods. These soil redistribution rates were compared with RUSLE predictions for the two transects and SIBERIA model predictions for the entire catchment. The net soil loss rates established in this study were also compared with the results of other studies in the region obtained with a range of different methods. Estimates based on 137Cs using an empirical conversion method compared well with published regional rates derived using rainfall-runoff plots, sediment yields and 137Cs, whereas theoretical 137Cs conversion models were found to over-estimate soil redistribution rates. Similarly, the RUSLE model significantly overestimated soil erosion rates in this study as was the case in other studies in the region. The agreement between SIBERIA and 137Cs, and erosion rates obtained elsewhere in the region, provides confidence in SIBERIA for catchment scale erosion assessments. The results of this study demonstrate the limitations associated with using theoretical 137Cs conversion models in environments for which they are not suited. This study also highlights the need for caution when quantifying soil erosion using both field methods and modelling approaches. The results demonstrate that DEM based erosion models are reliable tools for the prediction of soil erosion on the hillslope and catchment scale.