Publications

Dille, A.,  Nobile, A.,  Monsieurs, E.,  Kervyn, F.,  Kervyn, M. & Dewitte, O. 2017. ‘Characterizing landslide dynamics in a tropical urban environment: focus on Bukavu (DR Congo)’. Journées Aléas Gravitaires 2017. Book of abstracts.

Résumé de colloque

Loss of lives associated with landslides has been shown to be concentrated in less developed countries, particularly in tropical areas where a combination of active tectonics, steep topography, intense rainfall and high population density can be found. In most of these regions, however, landslide process characterisation relevant for hazard assessment and urban planning remains rare. The fundamental challenge is to gain insight in what might be called the baseline rates/frequencies of these processes, i.e. the rates/patterns that would occur under natural conditions whereby geophysical, biophysical, and climatic conditions control temporal and spatial variations as well as the way in which humans have affected these processes, so that more effective remediation strategies can be laid out. The objective of this research is to investigate the spatio-temporal pattern and the mechanisms of landslides in an urban-sprawling context in the tropics. The region of interest lies in the urban and peri-urban environments of the rapidly expanding city of Bukavu (DR Congo), situated on the southern shore of Lake Kivu in the western branch of the East African Rift (Figure 1). The first comprehensive landslide inventory reveals that landslides cover about one third of the city area (Dewitte et al., 2017). While it is clear that some of them are rather old and exclusively related to natural factors, i.e. they predate first colonial settlements (Maki Mateso and Dewitte 2014, Dewitte et al., 2017), at the same time, new landslides occur every year and some of them have dramatic impacts (fatalities, damages to the infrastructures, etc.). To characterise the landslides affecting Bukavu, we processed 80 COSMO-SkyMed (CSK) X-band SAR images acquired between March 2015 and April 2016 with a mean revisiting time of 8 days in both ascending and descending orbits. We produced surface displacement rate maps and ground deformation time series using the Small Baseline Subset (SBAS) technique (Figure 2 a. & b). Figure 2b shows a zoom of the deformation map over the largest (~1.5 km²) active landslide affecting the city. These results support field observations, with for example indication of the presence of multiple blocks moving at different velocity over this landslide. The displacement rates are also consistent with the DGPS measurements taken at regular intervals (~4 per years) over the area. However, despite the short revisiting time offered by CSK constellation, a loss of coherence is observed within the fastest moving regions of the landslide. These non-coherent areas have displacement rates up to 0.5 m/yr according to the DGPS measures. Two very high-resolution DEMs were constructed based on stereo-pairs of Pléiades images, calibrated with the available DGPS points. These models are used to better characterize the slope deformations and will also serve for pixel-tracking on both Pléaides and CSK coherence images. Such techniques will be used in combination with the InSAR processing of longer wavelength (C-band) and shorter revisit time Sentinel 1 images and should enable to constrain the displacement rates of the fast moving zones. To complete deformation tracking Ground-based time lapse cameras will be installed and firsts UAV acquisitions performed. These measurements should enable to complement and validate the optical and SAR derived ground displacements measurements.