Abstract :
[en] Because of high inter-annual variability, representative soil erosion measurements are best carried out over time scales of a few decades. In this paper, a novel approach was tested that relies on diachronic, high-resolution soil mapping. For this purpose, a 116-ha, cropland-dominated watershed was selected in central Belgium. A first augering soil survey was carried out in 1956, with a density of 1.8 observation points /ha. A second survey was carried out in 2015 with a density of 4.5 observations/ha. The results were interpreted in terms of changes in depth of transitions between successive horizons. The results indicate that a great proportion of the soils classified as non-eroded in 1956 were strongly eroded in 2015. Strongly eroded soils represented 52% of the watershed area in 2015 but only 8% in 1956. Some colluvial areas have been subjected to erosion and then covered again by colluvium over the last 60 years. Only the flatter areas close to the watershed boundary appear to remain in the “non eroded soils or soil subject to weak erosion” class. A mean net erosion rate of 37 t/ha.yr is computed within the watershed with this method. The diachronic comparison of the pedological maps over a 60-year time interval in a context of intensive farming is a potential source of information for the calibration of landscape evolution spatial modelling. This method joins the few ones allowing a quantitative spatialization of erosion and deposition phenomena at the catchment scale.
Scopus citations®
without self-citations
2