Estimating long-term chemical denudation rates on soil-mantled hillslopes using cosmogenic nuclides: Effects of time-varying erosion

Many biogeochemical and Earth surface processes depend critically on chemical weathering. The immediate products of chemical weathering are present as solutes and secondary minerals in groundwater, soils, and streams, and form the nutritional foundation for terrestrial biogeochemistry. Chemical weathering also contributes to physical erosion by weakening bedrock and producing easily erodible regolith, and as the primary long-term sink for atmospheric CO2 it modulates Earth’s long-term climate via the greenhouse effect. Long-term chemical denudation rates on soil-mantled hillslopes can be estimated from cosmogenic radionuclide (CRN) concentrations in soil-borne quartz and the enrichment of a chemically inert tracer in soil relative to its parent bedrock. This technique inherently assumes steady physical erosion over the timescale of CRN accumulation. Here we assess how time-varying physical erosion rates affect the accuracy of this technique, using a numerical model for chemical denudation rates in soils. In this model, bedrock supplies fresh minerals to the soil at a rate that decreases exponentially with increasing soil thickness, and Preprint submitted to Elsevier 29 November 2007 mineral dissolution rates are proportional to soil mineral concentrations. Our modeling results suggest that CRN-based estimates of chemical denudation rates closely resemble actual chemical denudation rates averaged over the timescale of CRN accumulation, even during large-amplitude and long-period oscillations in physical erosion rates. For example, this model predicts that when physical erosion rates fluctuate sinusoidally by 50% of their mean over any period in time, CRNbased estimates of chemical denudation rates should differ from actual chemical denudation rates by less than 15%. Our model also implies that chemical denudation rates should approach zero both when physical erosion rates approach zero (because soluble minerals become depleted in the soil) and when physical erosion rates approach the maximum soil production rate (because soil thickness approaches zero). Modeled chemical denudation rates thus reach a maximum at intermediate physical erosion rates. If this relationship holds in nature, it implies that in rapidly eroding regions, further increases in physical erosion rates (e.g., due to increases in tectonic uplift rates) may not necessarily lead to faster chemical denudation on soil-mantled hillslopes.