Modelling channel geomorphic change from landslide sediment delivery during Typhoon Mangkhut in the Philippines

The Antamok River in the Philippines experienced a complex geomorphic response to Typhoon Mangkhut in September 2018, which triggered >500 landslides in the Ambalanga catchment. Landslides are known to influence channel geometry by delivering large amounts of sediment. However, the interaction between landslide sediment delivery and channel geomorphic change during extreme events is poorly understood and rarely examined in tropical settings. The study catchment also has a legacy of anthropogenic modifications, such as the presence of extensive small-scale mining and tailings storage facilities (TSFs) from large-scale mining activities. The aim of this study was to use a mapping and modelling approach to test the hypothesis that landslide sediment delivery is a major control on channel geomorphic change. To accomplish this, we have applied the multi-phase model r.avaflow to a reach along the Antamok River encompassing the highest density of landslides and where it displayed a highly dynamic channel morphology. Landslide sediment delivery and TSFs are represented within r.avaflow by digitising areas of mapped landslides, allowing sediment to be delivered into the channel as well as being transported by the flow. To account for modelling uncertainty, we also tested the influence of several key-parameters by carrying out a sensitivity analysis. Using the approach described, r.avaflow was used to simulate the effects of landslide delivery and TSFs on channel erosion and deposition during the typhoon event. Results showed a good agreement between observed and simulated channel width change. When compared with traditional methods (e.g., unit stream power), the model results were considerably more accurate and consistent with observations. Furthermore, sensitivity analysis suggested that simulations are dependent on the type of sediment and physical processes considered, whilst other parameters only had negligible effects. Overall, the model simulations suggested that the impact of landslide and TSF sequences is highly dependent on the amount of sediment delivered by landslides, and a multi-phase model such as r.avaflow is possibly one of the most appropriate tools for simulating active channel width changes. Further research using these mapping and modelling tools is needed to better understand the contribution of sediment supply on channel geomorphic change during extreme events, that are otherwise difficult to observe and model.