N ° d ’ ordre : 4379 ANNÉE 2011

Erosion processes in large mountain belts are controlled by tectonic uplift and the spatial distribution of topography and precipitation. The aim of this work is to understand more in detail the influence of precipitation on erosion in the Nepal Himalayas. Here, the Indian Summer Monsoon defines a very clear cyclic seasonality, with two well separated and long-lasting seasons, one very wet and one very dry. I will investigate here: 1) the evaluation of precipitation datasets, 2) the transfer of precipitation to river discharge, 3) the mobilization and transport of material from the analysis of a unique suspended sediment dataset, and 4) the determination of millennial time-scale erosion rates using comsogenic isotopes. 1) High-quality precipitation observations with good spatio-temporal coverage are needed to investigate the possible impact of precipitation on erosion. I evaluate five gridded precipitation data sets and show that the APHRODITE dataset, derived from gauge data interpolation performs best in the Himalayas. 2) The transfer of precipitation into rivers involves temporary water storage in different reservoirs, where different residence times influence the hydrological cycle. I observe an annual anticlockwise precipitation-discharge hysteresis loop, revealing more river discharge than precipitation during the falling limp of the monsoon season. This implies the temporal storage of water in a fractured basement aquifers, whose storage capacity is ∼ 28 km3 at the scale of Nepal. 3) I present a new suspended sediment data compilation, with daily resolution for the major rivers in the Nepal Himalayas. I show that suspended sediment concentrations vary through the seasons describing an annual clockwise hysteresis effect. This hysteresis effect disappears when comparing suspended sediment fluxes directly with direct storm discharge, revealing a linear relationship. All river basins show the same erosion behavior when data are normalized by drainage area and mean sediment flux,. Erosion rates calculated from a new rating model, based on direct discharge, range between 0.1 and 5.9 mm/yr. The rivers in the Nepal Himalayas are supply limited and the hillsopes as contributing source are transport limited. This results provide some new insights on erosion mechanisms in the Himalayas. 4) I present mean catchment erosion rates, calculated from in-situ produced 10Be cosmogenic isotope concentration. The calculated erosion rates range between 0.2 and 4.4 mm/yr within five large streams in Nepal and several tributary catchments. I show an unreported systematic difference beIII te l-0 06 74 91 9, v er si on 1 28 F eb 2 01 2 tween tributary and main streams erosion rates, suggesting that the system is not in an equilibrium state. However, this effect could be also due to bias in the calculation of the erosion rates. IV te l-0 06 74 91 9, v er si on 1 28 F eb 2 01 2