Publications

Nyssen, J.,  Poesen, J.,  Moeyersons, J.,  Lavrysen , E.,  Mitiku , H. & Deckers, J. 2002. ‘Spatial distribution of rock fragments in cultivated soils in northern Ethiopia as affected by lateral and vertical displacement processes’. Geomorphology 43: 1-16. Elsevier.  (PR).

Article dans une revue scientifique / Article dans un périodique

Although, in semi-arid environments, rock fragments at the soil surface and within the topsoil play an important role in desertification control, little is known about their spatial distribution. Therefore, this study analyses spatial patterns of rock fragment cover along catenas, and the vertical variations in volumetric rock fragment content in soil profiles in the highlands of Tigray, northern Ethiopia. Natural and anthropogenic processes inducing these patterns are assessed. Volumetric rock fragment content (RV) was analysed in 10 soil pits. All rock fragments were extracted, and their volumes determined by pedostratigraphic unit, size and lithology. The rock fragment cover (RC) was determined by the point-count method using vertical photographs of the soil surface. The following processes contribute to the vertical variability of RV: (1) in Vertisols, upsqueezing as a consequence of swell-shrink cycles (argillipedoturbation) is responsible for high RC at the soil surface; (2) large rock fragments ( > 7.5 cm) are rapidly brought to the soil surface by kinetic sieving through tillage, even in the case of continuous fine sediment deposition, what may result in a rock fragment rich subsoil, underlying a thick soil layer (up to 80 cm) without large rock fragments and a topsoil with a high RC at the surface; (3) Skeletic Regosols at the foot of cliffs show no systematic vertical rock fragment distributions. As to rock fragment cover along the catena, some fundamental differences appear between the basalt and limestone substrate. On the basalt catena with slope gradients between 0.06 and 0.42 m m
1, RC is high everywhere (57–85%) and is unrelated to slope gradient. Vertical processes such as kinetic sieving through ploughing and argillipedoturbation determine the rock fragment distribution at the soil surface. In limestone areas, argillipedoturbation is less active and RC is positively correlated with slope gradient (R2 = 0.74; n=6; P < 0.05), when only the rock fragments (0.5–2.0 cm across) are considered. Including the larger (>2 cm) and more kinetic sieving sensitive rock fragments lowers R2 to 0.46 (n=6; n.s.). Preconditions for the present spatial distribution of rock fragments at the soil surface in the study area are the occurrence of two geomorphic processes in the past: (1) mudflows depositing diamictons including much coarse debris and (2) intense water erosion, which occurred after deforestation and exposed rock fragments at or near to the soil surface. Generally, the balance between lateral and vertical movements of rock fragments now controls the spatial distribution of RC. With respect to the lateral displacement processes: (1) lateral transport over the soil surface by trampling, tillage and concentrated runoff, especially on steep slopes, and (2) rockfall from the cliffs. Vertical supply of rock fragments to the soil surface is caused by (1) selective runoff erosion and the development of erosion pavements, (2) tillage induced kinetic sieving, bringing preferentially large rock fragments (>7.5 cm) to the soil surface, and (3) argillipedoturbation in Vertisol areas.