Abstract
Wildfires disrupt geomorphic and environmental equilibria, amplifying the risk of subsequent rainfall-triggered landslides. This study introduces a scalable, data-driven framework for evaluating the spatiotemporal dynamics of landslide susceptibility across pre- to several post-wildfire phases. The framework integrates multi-epoch landslide inventorying, landslide density-anomaly analysis, phase-specific susceptibility modelling, and susceptibility-transition analysis. The approach is demonstrated in Lake County, California, a region where more than half of the land area has burned over the past decade. Using open-source Earth observation data and machine learning based classification, refined through clustering and facet-wise simplification, a multi-epoch inventory of 5045 landslides across ten epochs was produced. Although intense storms produced substantial interannual variability in landslide occurrence, density anomalies indicated that wildfire further amplified rainfall-triggered landsliding. Annual landslide occurrence varied markedly, with 2727 and 791 landslides recorded in 2005 and 2020, respectively, compared with fewer than 100 in both 2016 and 2018, highlighting substantial interannual variability associated with rainfall forcing Although intense storms influenced landslide frequency, the density-anomaly pattern indicated that wildfire further amplified rainfall-triggered landsliding. Landslide density peaked during the first post-wildfire year and generally returned towards pre-wildfire conditions within 2–3 years. Phase-specific susceptibility modelling similarly revealed a sharp post-wildfire increase followed by gradual recovery, although susceptibility-transition trajectories varied spatially across the landscape. Areas exhibited contrasting susceptibility trajectories, including persistence near pre-wildfire levels, recovery following an initial increase, and sustained elevation beyond the broader recovery period, reflecting differences in topographic, geologic, and vegetation controls. By linking temporal changes in landslide activity with spatial transitions in susceptibility, the proposed framework moves beyond static post-wildfire susceptibility assessment and provides a basis for time-dependent screening and prioritization of wildfire-affected terrain for engineering-geologic investigation, monitoring, and mitigation.
