Abstract : Fluvial erosion is a major agent of landscape evolution, noteworthy in bedrock rivers of orogenic belts. The goal of this memoir is the demonstration that an integrated study of abrasion processes (small scale and short term) is possible directly in field conditions if the “erosive background” is well constrained, in other words: (1) the characteristics of water fluxes (flood frequency, hydraulic properties and the flow structure analysis of the natural channel) and of sediment fluxes (volume of transported sediments, grain size); (2) the mechanical resistance of the bedrock. The other main objective is of course the in situ measurement of erosion, which requires the conception of new tools.
In our « natural laboratory » of the Himalayan foreland (frontal fold belt, Siwaliks of Central Nepal), yearly fluvial abrasion reached in 2006 several millimeters, even tens of millimeters. We propose a straightforward erosion model integrating fluxes characteristics and rocks erodabilities (i.e. the “erosive background”) at the scale of the channel section. Our model fits quite faithfully with measured erosion in the river bed. This manuscript is also the occasion of methodological outbreaks, mainly on the indirect measurement of the river bedload (conglomerate grain size analysis from pictures, sandstone disintegration) and on the measurement of the rocks erodability directly on the field with a Schmidt Hammer. At least we present new autonomous measurement devices of which a very promising « erosion sensor ».