Experimental landform development by rainfall erosion with uplift at various rates

Abstract

Four runs (1, 2, 3, and 4) of physical analog model experiments, in which a square (ca. 60×60cm) column of a mixture of fine sand and kaolinite is slowly raised at different rates (ca. 5.1, 1.3, 0.5, 0.2mm/h, respectively) under artificial rainfall of about 38mm/h, were conducted to observe how experimental landforms develop in relation to these uplift rates. As a square mound gradually emerges from ground level, fluvial erosion starts at the mound edges and develops into valley systems. This process of fluvial erosion, expressed in a linear relationship between relief (maximum height–minimum height) and mean cell slope (mean value of the highest slope gradient in a 1cm×1cm grid cell), dominates until relief reaches about 60mm, around the time when slope failures (slumps) start to dominate. If fluvial erosion dominates throughout the run (relief stays below 60mm), the uplift rate is considered to be below the lower threshold and landform development is in the “characteristic relief phase.” In all four runs, relief increases above 60mm and slumps become significant as hills grow, indicating that uplift rates in this series are above the lower threshold. In run 1, the uplift at a high rate overwhelmed erosion and a massive mountain-like topography formed despite the occurrence of large slumps. The uplift rate in run 1 is thus above the upper threshold and landform development is in the “mountain building phase.” The mountain is likely to collapse when it grows higher than the limit of mountain growth determined by factors other than uplift rates. In runs 2, 3, and 4, after valley systems develop over the surface, hills grow with the occurrence of slumps and channel profiles seem to become stable at gradients corresponding to the uplift rates. As slopes grow steeper than a certain “critical gradient,” which is possibly the angle of repose of dry mound-forming material, they become vulnerable to slumps. However, slopes of material containing water and clay can grow steeper than this gradient in the absence of triggering events. The frequency distribution of cell slopes becomes bimodal, indicating the dominance of two types of slopes divided by the critical gradient, one below and another above. The former represents surfaces formed mainly by fluvial processes, and the latter surfaces formed by and/or waiting for slumps. Slumps reduce relief, but average height (zmean) does not fall unless sediments produced by slumps are carried away by fluvial processes. The combination of slope failure, fluvial processes, and uplift eventually works to keep zmean stable around a certain height depending on uplift rates, while relief repeats decrease and increase. The experimental landforms in runs 2, 3, and 4 seem to have achieved a certain steady state with uplift and erosion. The landform development in these runs is considered to be in the “steady-state phase,” with uplift rates between the lower and upper thresholds. These observations and interpretations are surprisingly consistent with studies on real landforms and can be of value in interpreting their development.