Risk Assessment of Sediment Deposition in Built-up Environments Caused by Extreme Meteorological Events
Sobhan Emtehani is a PhD student in the Department of Applied Earth Sciences. (Co)Promotors are prof.dr. V.G. Jetten, prof.dr. C.J. van Westen and dr. D.B.P. Shrestha from the Faculty ITC.
Floods, debris flows, and landslides are among the most damaging natural hazards and often transport large volumes of sediment into built-up areas. Sediment deposition can cause structural damage to buildings, destruction of household and business contents, disruption of roads and drainage networks, and substantial clean-up costs. Despite these impacts, sediment is often omitted in conventional flood risk assessments, which can lead to major underestimation of total losses.
This thesis develops an integrated framework to assess the risk of sediment deposition in urban environments by combining field observations, remote sensing, physically based process modelling, and economic loss estimation. The research focuses on the Caribbean island of Dominica, which has experienced severe hurricane-induced flooding and mass movements, notably during Tropical Storm Erika (2015) and Hurricane Maria (2017).
First, sediment deposition extent, depth, and volumes are quantified using pre- and post-event UAV photogrammetry and LiDAR-derived elevation data.
Next, the LISEMHazard model is applied to simulate landslides, debris flows, and sediment deposition, and its sensitivity and uncertainty are evaluated to understand the robustness of predicted sediment patterns and volumes.
The framework is then implemented at catchment scale for two heavily affected watersheds (Coulibistrie and Grand Bay). Modelled hazard intensities are combined with exposure data and vulnerability functions to develop loss and vulnerability curves for buildings, contents, roads, and channels, explicitly accounting for both direct damages and indirect clean-up costs.
Results show that including sediment can increase estimated annualised losses by more than 99% compared to flood-only scenarios, demonstrating that sediment processes are critical for credible risk estimates.
Finally, the approach is upscaled to the national level to produce island-wide sedimentation risk estimations for historical events and synthetic rainfall scenarios (5-, 10-, 20-, and 50-year return periods), and the web-based tools FastFlood and FastSlide are tested for rapid assessment applications.
