Examining erosion in New Zealand over millennial timescales using in-situ 10Be and 14C

Landslides are the major erosional process in many orogens, and one of the most sensitive erosional process to tectonic and climatic perturbations. However, it remains extremely difficult to constrain long-term or past rates of landslide activity, and hence their contribution to long-term landscape evolution and catchment sediment fluxes, because the physical records of landsliding are often removed in <10² yrs. Here, we use the in-situ ¹⁰Be and in-situ ¹⁴C concentrations of recent landslide deposits and catchments from the Fiordland and the Southern Alps of New Zealand to: (a) estimate landslide frequencies over 10³-10⁴ yr timescales, which we compare against landslide inventories mapped from air photos (<10² yrs) to estimate changes in landslide activity, (b) quantify catchment-averaged erosion rates, and landslide’s contribution to those erosional fluxes, and (c) test whether paired ¹⁴C-¹⁰Be measurements can be used to trace erosional depth-provenance and identify transient erosion rate changes. We show that ¹⁰Be concentrations on landslide deposits can be used to estimate landslide recurrence intervals and frequency over 10³ yr timescales, and that ¹⁴C/¹⁰Be ratios reflect the depth-provenance of sediment, and possibly transient changes in erosion rates. The comparison of our ¹⁰Be-based long-term landslide frequencies with short-term published inventories suggests that landslide frequencies have increased towards the present by up to an order of magnitude. We compare sediment fluxes inferred from these long- and short-term landslide inventories with sediment flux estimates derived from ¹⁰Be catchment-averaged erosion rates, which allows us to examine fluctuations in erosion rate estimates from 10¹ to 10³ yrs timescales.