Modelling debris-flow source-area connectivity and impacted stream channels in the semi-arid central Andes of Chile using random walks and network analysis

In the last decade, major debris flows events in remote areas of the semi-arid central Andes of Chile have led to critical water supply shortages for large populated areas such as Santiago de Chile. There is therefore a crucial need for modelling debris-flow sediment connectivity to stream channels to identify both vulnerable stream channel sections and sediment source locations to focus mitigation efforts to ensure the reliability of drinking water supplies. In this research, we couple a statistical learning model of debris flow source areas with a process-based random-walk runout simulation to estimate the probability of source areas connecting to stream channel networks in a large catchment area of the upper Maipo river basin using a 12.5 m resolution digital elevation model. The runout model parameters are regionally optimised and validated using a spatial cross-validation approach.   Additionally, we perform network analysis to model the cumulative impact of potential debris flow sediment delivery to the stream channel network. The proposed methods are also designed for flexibility to adapt for assessing potential debris flow impacts and source areas corresponding to other critical features such as roads and buildings. Overall, the resulting predictive models of  runout sources and impacted areas provide not only valuable insights for characterising the potential impacts of debris-flows on stream channel networks, but also provides a model framework that can be potentially linked to weather forecast data for establishing early-warning systems of debris-flow related water supply shortages and quality issues in remote areas.