Denudation is a key parameter controlling the evolution of the Earth’s surface, the production of soils, the stability of relief or the long-term evolution of climate. Climate fluctuations conversely have a strong impact on denudation, but these complex feedback mechanisms are still under-constrained. To better predict future changes that will affect our habitat, and understand links between climate and denudation, precise quantification of paleo-denudation rates is required. In this work, we measure cosmogenic radionuclides (10Be) in turbidites of a well-dated marine sedimentary core recovered in the Mozambique Channel to provide a 900 ka long near-continuous record of paleo-denudation rates over the 100 ka climatic cycles. Neodymium isotopes and heavy mineral analysis were used to provide constraints on the provenance of terrigenous sediments exported from Madagascar to the studied site and show that temporal variations in sediment provenance are limited and decoupled from climatic cyclicity. Our 10Be-based paleo-denudation rates are in the same order as modern rates, ranging from 17.4 ± 5.8 mm/ka to 73.9 ± 29.4 mm/ka, and do not show major variations through the Middle and Late Pleistocene. Importantly, we did not identify a systematic significant impact of glacial/interglacial cyclicity on denudation rates. Denudation of this subtropical island may instead have been controlled by variability of monsoon intensity associated with shifts in the Inter Tropical Convergence Zone, but this interpretation remains speculative at this stage as it cannot be recorded within the resolution of cosmogenic-derived denudation rates.

The Glaciations impacted erosion during the Late Cenozoic but no consensus has emerged whether they led to increased erosion rates globally. In the Himalayas, recent work used past sediment concentrations of the terrestrial cosmogenic nuclide (TCN) 10Be and demonstrated that erosion rates have not permanently increased in the Himalayas. However, for the Quaternary, the published sedimentary records suffer from provenance uncertainties which prevent to elaborate on the causes of steady erosion rates. Here, we document the new, 4,000-m thick Valmiki Section (VS) to address this question. In the remote Valmiki Tiger Reserve, the VS consists of Siwalik sediment deposited in the Himalayan foreland basin by the Narayani River, a major river of Central Himalayas. To quantify past Himalayan erosion rates from TCN 10Be measurements, we determine: (1) the magnetostratigraphic deposition age model, (2) provenance using major elements and Sr-Nd isotopes, and (3) the recent cosmic exposure related to Siwalik exhumation using TCN 36Cl measurement in feldspar. The VS records Himalayan erosion from 7.5 to 1.25 Ma. Our 10Be results confirm steady erosion rates, close to modern values, 1.4-2.3 mm/y, with a brief increase by 35% at 2 Ma, possibly due to sustained glacial erosion of the high peaks as suggested by the geochemical signature. The Narayani Catchment may be more sensitive to the onset of the Glaciations because of larger glacial cover (presently ~10%) than elsewhere in the Himalayas. Despite this sensitivity, our results support that over long timescales, rather than climate, tectonics control Himalayan erosion.