Spatiotemporal distribution of regional landslides and erosion rates in Southeastern Tibet

The difficulty in estimating ages of regional landslides hampers to assess frequency of landslides and hence to quantitatively assess regional landslide hazard. In this study, we used radiocarbon dating of organic sediment on boulder rock varnish to estimate landslide ages in Langxian (LX) arid region in southeastern Tibet. Samples of rock varnish with organic sediment were collected on site for radiocarbon dating, leading to landslide ages from 1880 ± 30 to 18,430 ± 30 yr B.P. To measure surface roughness characteristics of 109 remotely-mapped large bedrock landslide deposits, we estimated average standard deviation of slope (SDS) over an area of ∼640 km 2 by calculating the slope gradient of each raster cell and using a rectangular moving window method in ArcMap from a 5 m-resolution Digital Elevation Model generated from helicopter-obtained photographs. Combing estimated landslide ages ( t ) with average surface roughness of mapped landslide deposits ( R ) quantified by SDS, we fit an exponential landslide deposits surface roughness-age function ( t =1.47×10 6 × e −1.46 R , r 2 =0.63) that was used to estimate regional landslide ages in LX. We conclude that three periods with clusters of regional landslides in LX were revealed by different surface roughness of landslide deposits combing roughness-age function, with the values of 5563–7455 yr B.P., 1724–4151 yr B.P., and 960–1287 yr B.P.. Furthermore, we used our estimates of landslide ages to quantify landslide erosion rates of three corresponding hillslopes in LX ranging from 0.50 to 2.42 mm yr −1 . Although rock varnish radiocarbon dating provides us a feasible option for timing regional landslides of arid regions, the epistemic uncertainty in the dating method should arouse our attention, which could be reduced by increasing the number of samples.