Tracking landslide sediment: 10Be dilution of detrital quartz in multiple grain sizes following the 2016 Mw 7.8 Kaikōura earthquake, New Zealand

Cosmogenic radionuclides (CRNs) have been proposed as a tool for tracking landslide sediment, yet interpreting the in-situ ¹⁰Be signal through time with respect to the post-seismic sediment cascade has proved challenging in the few examples where it has been applied. To determine how useful the ¹⁰Be concentration is in providing insights into the fluvial response of co-seismic sediment delivery, we measured the ¹⁰Be concentration in multiple grain sizes in two catchments in New Zealand affected by the 2016 Kaikōura earthquake. A 6-year post earthquake time series of ¹⁰Be measurements was compared to a sediment budget produced by differencing lidar surveys over the same study period. Results showed that the ¹⁰Be concentrations in the sand fraction are controlled by the post-seismic hillslope delivery, whereas the ¹⁰Be concentrations in the coarser grains respond to the total volume of landslide material in the active channel. Faster ¹⁰Be dilution and recovery of the sand fraction indicates that the transport rate of the sediment pulse is highly dependent on grain size. The residence of the fine-grained sediment is estimated to be within a decade of the earthquake, although it may take several centuries to evacuate the entirety of the landslide pulse. We observe different magnitudes of ¹⁰Be dilution between two study catchments, which is attributable to a difference in connected landslide volume from hillslopes. We also demonstrate that the dilution factor has little relationship to conventional landslide metrics on a global scale. Our analysis suggests that the magnitude of dilution is controlled through a complex relationship between the connected landslide volume, the degree of sediment mixing and the delivery and evacuation time of landslide sediment. Additionally, if coarse sediment is primarily sourced through mass-wasting events, measuring just the sand ¹⁰Be concentration will bias the catchment-averaged erosion rate.