Search

The movement of large, slow-moving, deep-seated landslides is regulated principally by changes in pore-water pressure in the slope. In urban areas, drastic reorganization of the surface and subsurface hydrology—for example, associated with roads, housings or storm drainage—may alter the subsurface hydrology and ultimately the slope stability. Yet our understanding of the influence of slope urbanization on the dynamics of landslides remains elusive. Here we combined satellite and (historical) aerial images to quantify how 70 years of hillslope urbanization changed the seasonal, annual and multi-decadal dynamics of a large, slow-moving landslide located in the tropical environment of the city of Bukavu, Democratic Republic of the Congo. Analysis of week-to-week landslide motion over the past 4.5 years reveals that it is closely tied to pore-water pressure changes, pointing to interacting influences from climate, weathering, tectonics and urban development on the landslide dynamics. Over decadal timescales, we find that the sprawl of urbanized areas led to the acceleration of a large section of the landslide, which was probably driven by self-reinforcing feedbacks involving slope movement, rerouting of surface water flows and pipe ruptures. As hillslopes in many tropical cities are being urbanized at an accelerating pace, better understanding how anthropogenic activity influences surface processes will be vital to effective risk planning and mitigation.

The rapid and typically uncontrolled growth of many African cities leads to a plethora of problems and challenges. One of these is the formation and expansion of large urban gullies (UGs) in many (sub)tropical cities. UGs typically lead to the destruction of houses and other infrastructures, displace large numbers of people and often claim casualties. As the formation of such gullies is strongly linked to land use and rainfall intensity, the problems associated with UGs are likely to aggravate in the near future as a result of continued urban expansion and climate change. However, this newly emerging geo-hydrological hazard hitherto received very little research attention. Several studies report on the occurrence and impacts of UGs. Yet, they remain limited to specific local case studies. A clear understanding of the patterns, impacts and driving factors of UGs at larger scales is currently lacking. To address this gap, we aim to better understand the spatial patterns and UG occurrence at the scale of Africa.In order to achieve this, we are documenting cases of UG occurrence across Africa through the visual analysis of very high spatial resolution satellite imagery. This mapping already allowed us to identify more than 3,500 UGs in 11 countries (mainly across D.R. Congo, Angola, Republic of the Congo, Nigeria and Mozambique). Using on this database, we develop a logistic regression model that accurately simulates the likelihood that UGs occur within (peri-)urban areas across Africa. Our preliminary results show that a combination of rainfall characteristics, topography, soil type and variables describing the land use/urban context can already robustly explain why certain cities are extremely susceptible to the problem and others not. Overall, our dataset and model are first crucial steps to better understand the current and future risks of UGs across Africa.

Large urban gullies (UGs) cause major infrastructural damages and often claim casualties in many tropical cities of the Global South. Nonetheless, our insight into this new type of geo-hydrological hazard remains limited to some case studies and the overall impacts remain poorly quantified. Here, we aim to bridge this gap by making a first assessment of the number of persons affected by urban gullies at the scale of the Democratic Republic of Congo (DRC). We used Google Earth imagery in combination with local news sources and earlier research to identify 25 cities in DRC where UG occur at a significant scale (at least ten UGs). This list is likely exhaustive. Next, for each of these cities, we used Google Earth imagery and other high resolution satellite images to map all visible UG, evaluate their expansion rate and inventorize detectable damages to houses and roads. In total, >2,000 UGs were mapped across the 25 affected cities. Overall, the problem of UGs in DRC is especially acute in the cities of Kinshasa, Mbujimayi, Kikwit, Tshikapa and Kananga. Over 90% of these gullies were active during the observation period (typically from 2002 to 2020). Next, we assessed the total number of persons that are directly affected, as well as the number of persons currently at risk. Using available high resolution population density data and taking into account the current position of urban gullies, we estimate that around 68,700 people were directly displaced due the formation and expansion of UGs over the last 15 years. This corresponds to an average of ca. 4,300 persons per year. By considering the population that lives in the direct vicinity (Overall, this research shows that urban gullying is a very serious problem in the Democratic Republic of Congo, but likely also in many other countries of the Global South. More research is needed to better understand this processes and, ultimately, to prevent and mitigate its impacts. The results and the database of this study provide an important step towards this.

Uvira is a rapidly growing city of about 600,000 inhabitants in DR Congo. Squeezed between the shore of Lake Tanganyika and steep mountains hillslopes, and under the influence of a tropical climate, the city is familiar with flash floods. Nevertheless, the impacts of the flash flood event of April 2020 have been unprecedented in the last decades. Debris-rich flash floods led to at least 43 deaths, nearly 200 injuries, more than 5,500 houses destroyed and at least 70,000 people made homeless. Dozens of socio-economic infrastructures were damaged and nearly 280,000 people were left without hydroelectric power as a result of damage to water and electricity distribution networks in the city. In this work we explore the natural and anthropogenic causes of the exceptional impacts of these flash floods. To do so, we use satellite images, historical aerial photographs, social media reports, field observations, and details provided by local stakeholders and citizen observers. We show that a high-magnitude rainfall event, that occurred at the end of a wetter-than-usual rainy season, triggered, over an exceptionally large area for the region, hundreds of landslides in the upper parts of the watersheds. These landsides transported extra material to the flooded rivers, increasing their sediment content and lateral mobility. We also show that the landscapes of the watersheds where this compound event occurred remained rural in the last six decades and were not impacted by significant forest cover changes and road construction, hence eluding the role of key potential human activities on the magnitude of the flash floods. However, downstream, the city expanded with little consideration of the geomorphological context of the environment where the alluvial fans and flood plains of the rivers have been significantly urbanized in the last decades, often in an informal manner. The impacts of the April 2020 flash flood event are still present almost after two years and are anticipated to remain at least a few more years; most remarkably, the disturbance of the river dynamics and the severe bank erosion caused by the large supply of sediment. Furthermore, often-uncontrolled sediment mining, seen here by the local people as an opportunity, exacerbates the unpredictable dynamics of the system. In the meantime, flooded areas where houses were destroyed are being resettled. Although authorities and inhabitants are aware of the danger, they face limited means to improve the management and planning of the city.

Kinshasa, the capital of the D.R. Congo, is strongly affected by urban mega gullies. There are currently hundreds of such gullies, having a total length of >100 km. Many of these gullies (typically tens of meters wide and deep) continue to expand, causing major damage to houses and other infrastructure and often claiming human casualties. To mitigate these impacts numerous efforts are being implemented. The type and scale of these measures varies widely: from large structural measures like retention ponds to local initiatives of stabilizing gully heads with waste material. Nonetheless, earlier work indicates that an estimated 50% of the existing urban gullies continue to expand, despite the implementation of such measures. As such, we currently have very limited insight into the effectiveness of these measures and the overall best strategies to prevent and mitigate urban gullies. One reason for this is that gully erosion is typically very episodic with long periods of stability, followed by sudden expansion events. As a result, understanding the dynamics of gully expansion in urban environments requires observations over sufficiently long time periods. However, most current initiatives to stabilize urban gullies happen on a rather isolated basis and are rarely evaluated afterwards.This work aims to improve our understanding of this issue by constructing a large inventory of measures implemented to stabilize urban gullies in Kinshasa and statistically confronting these measures with observed vegetation recovery and long-term gully expansion rates (derived from high-resolution imagery over a period of >10 years). Our preliminary results (based on a dataset of > 140 urban gullies) shows that the most commonly applied measures are revegetation and reinforcement of gully heads with sandbags or household waste material (implemented in around 50% of the cases). Also retention ponds and water storage tanks are frequently implemented (around 30% of the cases). Surprisingly, our results indicate that urban gullies with higher expansion rates tend to have more measures implemented in their upstream catchment. While this seems counterintuitive, it may point to the fact that more actively retreating gullies create a larger sense of urgency and therefore instigates a higher number of (often ineffective) initiatives. More research is needed to confirm this. Furthermore, the stability of gullies seems to be strongly linked to vegetation cover in the gully. Nonetheless, it is not always clear if vegetation is the cause or the result of this stability. Overall, this study provides one of the first large scale assessments of the effectiveness of gully control measures in urban tropical environments. With this study, we hope to contribute to a better prevention and mitigation of this problem that affects many cities of the tropical Global South.

Urban gullies cause major infrastructural damages and often claim casualties in many tropical cities of the Global South. Nonetheless, our understanding of this hazard is currently limited to some case studies while the overall impacts remain poorly quantified. Here, we aim to bridge this gap by making a first quantification of the number of persons and buildings affected by urban gullies at the scale of the Democratic Republic of Congo (DRC). We used Google Earth imagery to identify and map urban gullies for cities in the DRC and evaluate their expansion rate and the resulting damages where possible. In total, more than one thousand urban gullies were mapped across 22 affected cities. Over 80% of these gullies were active and, by analyzing their expansion, we identified 1463 houses and 386 roads destroyed. Nonetheless, the actual impacts are likely much larger since the limited amount of imagery available does not allow to quantify all impacts.We therefore also made an estimate of the total number of persons directly affected by urban gullies (i.e. displaced due to the destruction of their house). For this, we calculated the areal fraction of urban gullies in affected cities (which ranged from 0.12% to 4.66%) and combined these fractions with the urban population density. From this, we estimate that a total of 212 000 people have been affected. The problem is especially acute in the cities of Kinshasa, Mbujimayi, Tshikapa, Kananga, Kabinda, and Kikwit. Given that these gullies are linked to recent urban growth and typically less than 30 years old, we estimate that at least 7000 people/year lose their house as a result of urban gullies in DRC. This is likely an underestimation since (i) not all urban gullies are detectable; (ii) urban gullies may disappear and reappear over time; and (iii) many of these gullies are likely more recent than 30 years. Furthermore, this assessment does not take into account numerous other indirect impacts of urban gullies (e.g. impacts on traffic and sanitation, increased flood risks, real estate value loss and intangible impacts like fear or stress). Overall, this research shows that urban gullying is a serious problem in DRC, but likely also in many other tropical countries. More research is needed to better understand this processes and, ultimately, to prevent and mitigate its impacts. The results and the database of this study provide an important first step in this direction.

Accurate precipitation data are fundamental for understanding and mitigating the disastrous effects of many natural hazards in mountainous areas. Floods and landslides, in particular, are potentially deadly events that can be mitigated with advanced warning, but accurate forecasts require timely estimation of precipitation, which is problematic in regions such as tropical Africa with limited gauge measurements. Satellite rainfall estimates (SREs) are of great value in such areas, but rigorous validation is required to identify the uncertainties linked to SREs for hazard applications. This paper presents results of an unprecedented record of gauge data in the western branch of the East African Rift, with temporal resolutions ranging from 30 min to 24 h and records from 1998 to 2018. These data were used to validate the Tropical Rainfall Measuring Mission (TRMM) Multisatellite Precipitation Analysis (TMPA) research version and near-real-time products for 3-hourly, daily, and monthly rainfall accumulations, over multiple spatial scales. Results indicate that there are at least two factors that led to the underestimation of TMPA at the regional level: complex topography and high rainfall intensities. The TMPA near-real-time product shows overall stronger rainfall underestimations but lower absolute errors and a better performance at higher rainfall intensities compared to the research version. We found area-averaged TMPA rainfall estimates relatively more suitable in order to move toward regional hazard assessment, compared to data from scarcely distributed gauges with limited representativeness in the context of high rainfall variability.

Studying the causes and triggers of landslides is essential to understand the key process of hillslope evolution and the hazards they generate. Such understanding is crucial in tropical areas where landslide impacts are high and on the rise, and the dearth of accurate processes characterisation is large. Here we investigate the timing and the mechanisms of relatively slow-moving deep-seated landslides in weathered tropical environments through the analysis of a landslide located in the Kivu Rift (DR Congo). This landslide, developed in weathered basalt, shows obvious deformation features at its surface indicating large deformations during recent years, making it a unique natural laboratory in an underexplored area. High-resolution topographic data, historical aerial photographs, satellite imagery and careful field investigations are used to detail the landslide mechanisms and investigate failure development over a 60-year record. By confronting rainfall time series and earthquake sequences with the different deformation episodes, we show that the relation between instability triggers and slope failure is not straightforward; e.g., the largest instability occurred at the end of a dry season during a period of relatively low seismicity. Instead of direct influence of external triggers, we show that some phases of instability may be caused by the intrinsic evolution of the hillslope associated with weathered-related weakening of the slope strength through time. Our results question the relative weight of the commonly recognized causes and triggers of slope instability in this area. Analysis of landslide causes and triggers provided here should help improve the understanding of how surface processes influence the pace of hillslope evolution. It also contributes to a more accurate evaluation of the landslide hazard in the area and across other regions where similar environmental conditions are met.