dc.description.abstract | Debris flows are mixtures of fine to coarse sediment, water and air that are ubiquitous in mountainous areas. Due to their unpredictability and high velocity they impose great risk for people living in mountainous areas. Therefore, knowledge is required to assess how debris flows evolve through space and time. The necessity for such research is also evident by the relative absence of qualitative description of debris flow systems in scientific literature and limitations in the temporal resolution of remote sensing-derived datasets in the existing literature. Furthermore, a large body of knowledge on debris flows comes from experimental modelling, whereas the experimental observations are sometimes not sufficiently linked to natural debris flow systems. The current study area east of Bayasse, France, offers naturally occurring debris flow systems with limited human intervention, ideal for studying their spatiotemporal evolution. Using a combination of Structure-from-Motion and Multi-view Stereo (SfM-MVS), historical imagery from 1956, 1973 and 2004 and drone-derived imagery from 2019 and 2023 were converted into orthomosaics and digital elevation models (DEMs). The DEMs were used to produce DEMs of Differences (DoDs) showing net elevation gains and losses over time. Together with volume calculations, cross-sectional and longitudinal profiles, the spatiotemporal evolution of the two debris flow systems were studied. The results show that the observed morphology of the debris flow systems is in great accordance with existing scientific literature. It was concluded that two source areas were dominated by two different sediment-contributing processes: rockfall from scree slopes in one source area, and rock slope deformation in the other. In the long-term (multi-decadal scale), the fan-forming channels tend to balance the cross-sectional topography of the fan (topographic compensation). Short-term channel avulsion (multiple years) was found to be the cause of within-channel sedimentation processes, among which channel plugging, gradual shifts of the distal deposition locus and possibly backstepping. While this thesis shows that further research is needed in terms of the connection between sediment supply and debris flow occurrence, error corrections and further increasing the temporal resolution of the dataset, it was shown that the SfM-MVS approach is adequate for observing the spatiotemporal evolution of debris flow systems, and can thereby contribute to hazard assessment to prevent loss of life and damage to infrastructure. | |