The effect of fragmentation on the distribution of hillslope rock size and abundance: Insights from contrasting field and model data

Abstract

The disintegration of rocks into soil plays an important role in geomorphological processes, such as the evolution of hillslopes and river valleys. Nevertheless, the factors and processes controlling the physical weathering of rock particles are poorly understood. In this study, the surface and subsurface distribution of rock fragments abundance and size is measured in ten soil profiles along three hillslopes transects. For the surface horizons, we observed a significant linear relation between the ninth decile of the diameter, d90, and hillslope gradient and for both surface and subsurface horizons, a significant logarithmic relation between d90 and rock fragment abundance. For the first time, the performance of various fragmentation models is compared against field data to evaluate if the surface rock distribution can be explained from fragmentation of the subsurface material only. In six profiles these fragmentation models adequately reproduced the observed particle size distribution. In the other four profiles, all located on eroding hillslopes, armouring dominates over fragmentation and the surface rock distribution is coarser compared to the subsurface. Generally, the profiles have a good fit to almost all models with the exception of two. However, we can differentiate two zones in terms of fragmentation mechanism according to the modelling results. The profiles in the transect located along a gently sloping hillslope, and the two other profiles, along a steep river valley.In the models, we consider physical weathering as the single factor of particle size evolution, except or two models where conceptual chemical weathering of the fine fraction is considered, although this chemical weathering was not specifically designed to explore soil water chemistry or mineralogy in detail. The observed discrepancies between field data and the results of this study, especially in the fine fraction, reveal the possible relevance of chemical or biological weathering, redistribution processes and the lower accuracy of the number of fine fraction classes established and measured than in the coarse fraction for the field data.