Discordance between cosmogenic nuclide concentrations in amalgamated sands and individual fluvial pebbles in an arid zone catchment

Based on cosmogenic 10Be and 26Al analyses in 15 individual detrital quartz pebbles (16–21 mm) and cosmogenic 10Be in amalgamated medium sand (0.25–0.50 mm), all collected from the outlet of the upper Gaub River catchment in Namibia, quartz pebbles yield a substantially lower average denudation rate than those yielded by the amalgamated sand sample. 10Be and 26Al concentrations in the 15 individual pebbles span nearly two orders of magnitude (0.22 ± 0.01 to 20.74 ± 0.52 × 10610Be atoms g−1 and 1.35 ± 0.09 to 72.76 ± 2.04 × 10626Al atoms g−1, respectively) and yield average denudation rates of ∼0.7 m Myr−1 (10Be) and ∼0.9 m Myr−1 (26Al). In contrast, the amalgamated sand yields an average 10Be concentration of 0.77 ± 0.03 × 106 atoms g−1, and an associated mean denudation rate of 9.6 ± 1.1 m Myr−1, an order of magnitude greater than the rates obtained for the amalgamated pebbles. The inconsistency between the 10Be and 26Al in the pebbles and the 10Be in the amalgamated sand is likely due to the combined effect of differential sediment sourcing and longer sediment transport times for the pebbles compared to the sand-sized grains. The amalgamated sands leaving the catchment are an aggregate of grains originating from all quartz-bearing rocks in all parts of the catchment. Thus, the cosmogenic nuclide inventories of these sands record the overall average lowering rate of the landscape. The pebbles originate from quartz vein outcrops throughout the catchment, and the episodic erosion of the latter means that the pebbles will have higher nuclide inventories than the surrounding bedrock and soil, and therefore also higher than the amalgamated sand grains. The order-of-magnitude grain size bias observed in the Gaub has important implications for using cosmogenic nuclide abundances in depositional surfaces because in arid environments, akin to our study catchment, pebble-sized clasts yield substantially underestimated palaeo-denudation rates. Our results highlight the importance of carefully considering geomorphology and grain size when interpreting cosmogenic nuclide data in depositional surfaces.