Use of lake deposits to reconstruct the effects of multiple environmental stressors on landslide-derived sediment in the Daliang mountain region, SW China

Abstract

Assessment of historical landslide-derived sediment dynamics can improve our understanding of the temporal pattern of landslide erosion under environmental changes to help inform better management strategies. The Daliang Mountain region, featured by frequent seismic activities and concentrated precipitation, is highly prone to landslide erosion in Southwest China and has experienced intense human disturbance over the past several decades. Here, we reconstructed the proportion of sediment and associated sediment deposition rates derived from landslide (DRs) recorded in the dated lake deposits in response to seismic activity, climate change, and human activities in a representative tectonically active catchment from the Daliang Mountain region, using a machine learning algorithm and Bayesian model. Three geochemical tracer sets (i.e., element (E) set, specific element ratios (S) set, and their combination set (E + S), were respectively examined to discriminate between landslide and other sediment sources. The results indicate that landslide erosion was the dominant source, contributing over 70 % of the core sediment in the representative core. We identified three general stages of “increasing DRs→highest DRs→decreasing DRs” reflecting environmental changes in the region over the past 70 years. Before 1966, land reclamation promoted by the Yi serfdom reform and the Cheng-Kun Railway's construction, exacerbated massive deforestation, causing a rapid increase in DRs during the initial period. DRs were the highest during 1967–2004 due to the combined effects of frequent seismic events and intensified precipitation. The decreasing DRs during the latest stage (after 2005) correspond to the implementation of check dam initiatives. These findings improve our understanding of the temporal pattern of landslide erosion in response to environmental changes impacted by multiple factors, providing evidence for supporting landslide risk management in the region.