Landslides in a Changing Tropical Environment: North Tanganyika-Kivu Rift Zones

Abstract

The North Tanganyika and Kivu Rift zones (Burundi, DR Congo, Rwanda) encompass a region where environmental
factors such as heavy rainfall, tectonic activity, and steep topography favor the occurrence of landslides.
These landslides have a negative impact on the livelihoods of the local population that suffers yearly from damage
to infrastructure and losses of agricultural land. Moreover, the area is characterized by a high and continuously
increasing population density resulting in (1) the expansion of urban areas and their associated road infrastructure
and (2) deforestation caused by the creation of farmland. At present, however, it is not known whether landslides
can be considered as a natural phenomenon or whether they are the consequence of human impact through e.g. the
modification of the land cover.
Therefore the main objective of this research is to assess the long-term impact of LULC changes on the
prevalence of landslides. In a first stage, a database of circa 4000 landslides was compiled on the basis of a visual
inspection of Google Earth Imagery. This allowed making a distinction between recent landslides (<10y) and
older landslides (>10y). For each of these landslides topographic setting, land use, and peak ground acceleration
(PGA) were assessed. The first results of the analysis show that recent landslides are in many cases triggered by
extreme rainfall events, often in combination with disturbances in land cover related to road construction and/or
mining activities. The spatial pattern of the older landslides can be correlated with earthquake-triggered activities
along major fault systems.
In order to better understand the role of human activities on the prevalence of landslides, historical photographs
and archives will be used to characterize LULC changes over the past 60 years. The results of this
historical analysis will enable the identification of interactions between environmental changes and landslides.

Full text

Faculty of Sciences
Department of Earth and Environmental Sciences
Landslides in a Changing Tropical Environment: North Tanganyika - Kivu Rift Zones
Arthur Depicker1, Gerard Govers1, Anton Van Rompaey1, Hans-Balder Havenith2, Jean-Claude Maki Mateso3,4 and Olivier Dewitte5
1Katholieke Universiteit Leuven, Department of Earth and Environmental Sciences, Belgium (arthur.depicker@kuleuven.be), 2Université de Liège, Départment of Geology, Belgium, 3Centre de Recherche en Sciences Naturelles,
Department of Geophysics, Lwiro, DR Congo, 4Université catholique de Louvain, Earth and Life Institute –Environmental Sciences, Louvain-La-Neuve, Belgium, 5Royal Museum for Central Africa, Department of Earth Sciences,
Tervuren, Belgium
Figure 1: An overview of the North Tanganyika –Kivu Rift zone.
In total, 6300 landslides were identified.
Landslide Database
A database of 6300 landslides was
compiled on the basis of visual
inspection of Google Earth Imagery
(Figure 2).
Figure 3: Composition of the landslide database
(percentages). Based on the classification by
Hungr et al. (2014)[1].Landslides linked to mining
activities were not considered.
Figure 6: Aerial photograph of Bukavu, 1959
(credits to the Royal Museum for Central Africa,
Belgium).
References
[1] Hungr O, Lerouel S, Picarelli L (2014) The Varnes classication
of landslide types, an update. Landslides, 11(2):167-194
[2] Larsen IJ, Montgomery DR (2012) Landslide erosion coupled
to tectonics and river incision. Nature Geoscience, 5:468-473
[3] Malamud BD, Turcotte DL, Guzetti F, Reichenbach P (2004)
Landslide inventories and their statistical properties. Earth Surface
Processes and
Landforms, 29:687-711
[4] Rossi M, Malamud BD (2014) D. 5.3. Prototype SW for
determination of landslide statistics from inventory maps
(http://www.lampre-project.eu/index.php?option=com_phocadownload
&view=category&id=6:wp5-triggered-event-landslides&Itemid=203)
  
 





(1)
Preliminary Results
•Recent landslides seem to cluster,
which could indicate an extreme
rainfall event trigger.
•Recent landslides often co-occur
with land cover disturbances, road
construction, or mining.
•Old landslides seem to correlate
with seismicity along major fault
systems.
•Through comparison of the slope
distributions of the landslides and
the study area, a threshold angle
(TA)[2] for the slope stability of
24.8°is derived (Figure 4a).
Figure 4b shows that the TA differs
for the different landslide types.
•The frequency density curve for
the landslide area, fitted to the
inverse distribution (Eq. (1)),
seems to indicate that the
database is representative for the
study area[3].
Figure 4: (a) Slope distribution of the
landslides, the buffer area around the
landslides (buffer of 0.01°), and the total study
area. The black vertical line indicates the TA
(b) Visualization of the slope distribution for
slides, flows, and their surrounding area (buffer
of 0.01°)
Figure 5: Inverse frequency density function fit
to frequency of the landslide areas in the study
area[4]
Research Questions
This study has two main objectives:
1. Characterize landslides in the study
area and link their occurrence to
environmental factors.
2. Explore the relationship between the
occurrence of landslides and LULC
changes since the 1960’s.
In order to provide an answer to these
questions, a landslide database is
compiled. In this poster, the first research
question is addressed.
Figure 2: Landslides triggered by rainfall, Eastern
shores of Lake Tanganyika (Google Earth)
DRC
Rwanda
Burundi
Tanzania
Uganda
Study Area
The study area (Figure 1)is situated in
the western branch of the East African
Rift and has a total surface of 86,744
km2.It ranges from the North Tanganyika
Rift Zone in the south to the Virunga
Volcanic Province in the north. On the
one hand, the area is prone to landslides
due to
•a steep topography,
•a tropical climate, and
•volcanic and seismic activity.
On the other hand, dramatic land use
and land cover changes have been
observed as a result of
•migration induced by military
conflicts and
•high and rapidly increasing
population density (415 inh/km).
For each landslide, this database
contains information on parameters
such as the topography, year of origin,
landslide area, and type (Figure 3).
(a)
(b)
Future Work
•Analyzing the occurrence of
landslides with regards to the
geodynamic of the region
(tectonics, lithology) and the
current environmental variables.
•Characterizing LULC changes
since the 1960’s using historical
database of ca. 300,000 aerial
photographs. The results enable
the identification of the interactions
between environmental changes
and landslides.
http://pasteca.africamuseum.be/home