[1] We present the results of an experimental study of topography dynamics under conditions of constant precipitation and uplift rate. The experiment is designed to develop a complete drainage network by the growth and propagation of erosion instabilities in response to tectonic perturbations. The quantitative analysis of topographic evolution is made possible by using telemetric lasers that perform elevation measurements at an excellent level of precision. We focus our study on the effect of initial surface organization and of uplift rate on both the transient dynamics and the steady state forms of topography. We show that the transient phase is strongly dependent on the initial internally drained area, which is found to decrease exponentially with time. The topography always reaches a steady state whose mean elevation depends linearly on uplift rate with a strictly positive value when uplift is zero. Steady state surfaces are characterized by a well-defined slope-area power law with a constant exponent of À0.12 and an amplitude that depends linearly on uplift rate with a strictly positive value when uplift is zero. These results are consistent with a stream power law erosion model that includes a nonnegligible threshold for particle detachment. Uncertainty regarding the sediment transport length is resolved by calibrating the transient dynamics with a surface process model. Reappraising published results on the linear dependency between mean elevation, or relief, and denudation rate, we suggest that an erosion threshold is worth considering for large-scale systems.

34The discharge of the central Himalayan rivers is governed by a strong precipitation seasonality 35 3,6,9,10 ( Fig. 1) with up to 80% of the annual rainfall occurring during the Indian Summer Monsoon 36(ISM) season 3 . The ISM precipitation is the main source for glacier mass accumulation 9 and its spatial 37 distribution is strongly influenced by orographic effects 3 48We investigate the transfer of water within the main catchments of the Nepal Himalayas (Fig. 49 1a) using a daily meteorological and hydrological dataset spanning ~30 years (Table 1). We consider 50 the three main catchments of Nepal (Sapta Koshi, Narayani and Karnali basins), some of their 51 tributaries, and three unglaciated small catchments at the front of the Himalayan range ( Fig. 1a and 52 Table 1). The main catchments drain the entire Himalayan range of Nepal, from the Tibetan Plateau to 53 the Lesser Himalayas. Most of their headwaters are located on the arid Plateau (Fig. 1a) (Fig. 1c). Most of the data considered here come from outlet stations located to the north of 58 the Siwalik foreland. The annual specific discharge of the studied basins is typically on the order of 59 ~10 3 mm yr -1 (Table 1) and their annual hydrograph clearly shows the seasonal impact of the ISM on 60 river discharge, generally peaking in July/August 3,14 (Fig. 1b). Mean annual basin precipitation is 920, 611396 and 920 mm yr -1 in the Sapta Koshi, Narayani and Karnali catchments, respectively. However, 62 precipitation is spatially heterogeneous (Fig. 1a) and is strongly controlled by orography, reaching a 63 maximum between elevations of 2 to 3 km 15,16 . The upper parts of the catchments are glaciated (Fig. 64 1a), covering between 4 and 15 % of the catchment area (Table 1). 66We calculated mean basin-wide daily precipitation rate and use daily discharge measurements 67 to compute specific water discharge for all the studied drainage basins (see Methods). Plots of daily 68 precipitation vs. specific discharge highlight a considerable scatter within the ~30 year dataset ( (Fig. 2a). A 30-day moving average highlights the temporal 74 consistency of the loop from year to year (Fig. 2a, inset). Data scattering results from inter-annual 75 variability, particularly during post-ISM, as illustrated by comparing the data during a strong or a weak 76 ISM year (see Supplementary Fig. S1). The annual anticlockwise hysteresis loop is observed in all 77 studied basins (Fig. 2b), regardless of the geological units, the presence of glaciers or snow cover 78 (Tab. 1). 80Anticlockwise hysteresis loops imply that precipitation is temporarily stored within the 81 catchments and not transferred directly to the river during pre-ISM and ISM seasons, whereas the 82 storage compartment is drained during post-ISM. Glaciers can be directly ruled out as the main 83 contributor to the observed hysteresis effect because the release of water by glacier or snow melt 84 occurs principally during pre-ISM to ISM season 3,13 ( Fig. 3b and S2), which is not consistent with the 85 ant...