Using 7Be measurements to estimate the relative contributions of interrill and rill erosion

• Simulated rainfalls were conducted on the cultivated runoff plots.
• 7Be was used to document interrill erosion during a rainfall event.
• Previous model for estimating the interrill erosion rate with 7Be was refined.
• Linear model for estimating interrill erosion contribution with 7Be was developed.

Rapid and reliable methods for estimating the relative contribution of interrill and rill erosion during a rainfall event are needed to provide an improved understanding of soil erosion processes and to develop process-based soil erosion prediction models. Use of the radionuclide 7Be in controlled experiments provides a means of addressing this need and this paper reports an experimental study aimed at refining and testing procedures employed to estimate the relative contribution of the two components of erosion. Four experimental plots (area 5 × 2 m and 10°, 15°, 20°, and 25° slope), filled with a loessial soil, manually tilled, and kept free of weeds with herbicides, were subjected to high intensity rainfall (91.8–120.6 mm h− 1), in order to induce rill development. The evolution of the rill network was documented photographically during the rainfall events and the runoff and sediment output from the plots were collected and measured. The sediment was recovered from the runoff and its mass and 7Be activity were measured. The Yang model, reported previously, was used to estimate the relative contributions of interrill and rill erosion from the 7Be activity of the exported sediment and this model was further refined to take account of the dynamic growth of the rills during the rainfall event. The results from the experiments were also used to develop a simple empirical linear model for estimating the relative contributions of interrill and rill erosion from the 7Be measurements. A comparison of the results provided by the three models showed some differences in the estimates of the magnitude of the relative contributions, although their trend during the event was similar. The estimates provided by the empirical linear model tended to be higher than those obtained using the refined model and lower than those generated by the Yang model, but were closer to those provided by the refined model which was seen as being theoretically the most accurate model. The basis of the empirical linear model was simpler than that of other models, since only a few, easily obtained parameters were required. In practical terms the empirical linear model therefore has important advantages and is judged to provide acceptable results. Further work is required to test its applicability under a wider range of conditions.