Soil erosion rates under different tillage practices in central Belgium: New perspectives from a combined approach of rainfall simulations and Be-7 measurements

Recent European Common Agricultural Policy (CAP) forces Flemish farmers in the Belgian loess belt to combat soil erosion on high erosion sensitive parcels. A possible measure against erosion is the application of reduced tillage operations. To evaluate the efficiency of reduced tillage practices, erosion rates of a maize field under conventional tillage, non-inversion tillage and strip-till were compared, all with and without wheel track compaction. To assess erosion rates, 20 min rainfall simulations in combination with fallout radionuclide (FRN) Be-7 measurements were used. At the small plot scale (5 m(2)), runoff coefficient, suspended sediment concentration and soil loss were measured directly during rainfall simulation in order to compare the effect of the different tillage treatments and the effect of wheel track compaction. After the rainfall simulations soil samples were taken for Be-7 analysis to explore spatial patterns in soil redistribution within the different plots. Furthermore, a total soil redistribution budget was obtained based on the Be-7 measurements and validated with the direct erosion measurements during the rainfall simulations. Direct measurements showed a significant decrease in runoff coefficient, suspended sediment concentration and total soil loss in the non-inversion and strip-till treatments compared to the conventional tillage treatment. No difference was observed between the non-inversion and strip-till treatment. For all treatments, wheel track compaction increased runoff coefficients and soil loss. The Be-7 measurements confirmed overall trends observed during the rainfall simulations, with high soil losses in the conventional treatment (5.96 +/- 1.37 kg-6.23 +/- 2.36 kg) and a decrease in soil loss with the non-inversion (1.50 +/- 0.34 kg-1.95 +/- 0.54 kg) and the strip-till treatment (-0.19 +/- 0.60 kg-0.10 +/- 2.50 kg). However, Be-7-derived net erosion estimates overestimated total soil loss per plot compared to the direct measurements. The appropriateness of correction factors, like particle size correction and variations in relaxation mass depth, to improve the tracer based assessment of absolute soil loss values was evaluated. Spatial patterns in soil redistribution clearly reflected the sediment and runoff buffering capacity of the inter-row area in the strip till treatment, while higher erosion rates were observed in the plant rows. Yet, sample representativeness to construct the soil erosion budget is a key consideration in light of discrepancies between tracer and rainfall simulation results. The different rainfall simulations support the preference of non-inversion tillage and strip-till over conventional tillage in order to reduce soil loss in the Belgian loess belt. The increased soil loss due to wheel track compaction implies the need to till favorable soil moisture conditions to avoid soil compaction.