Assessment and monitoring of soil erosion risk parameters in the trans-boundary region of Lake Kivu and the Ruzizi River basin

Kurzfassung

Anthropogenic activities like urbanization, intense land use and unsustainable land management practices together with an increase in weather and climate extremes, including longer and more severe droughts and an intensification of precipitation events, lead to a strong increase in soil erosion risk. The Lake Kivu and Ruzizi River basin, located in the trans-boundary region of the Democratic Republic of Congo, the Republic of Rwanda and the Republic of Burundi, is characterized by a mountainous topography and a high population density and is therefore particularly threatened by soil erosion. We present an analysis to assess and monitor soil erosion risk parameters in the Lake Kivu and Ruzizi region utilizing Earth observation technologies. The remote sensing analysis is performed to enrich an already existing comprehensive Revised Universal Soil Loss Equation (RUSLE) study, which was performed by our local partners for the same region. For the investigation of the complex setting in the study region, we focus on vegetation dynamics and uncovered soil in combination with climate data. Low vegetation cover, for example caused by inappropriate land management or droughts, can lead to a strong destabilization of top soil and with this increases its vulnerability to erosion. The vegetation dynamics in the region are analyzed using the Normalized Difference Vegetation Index (NDVI). For this, the more than 20-years long MODIS NDVI time series is used. This is combined with Sentinel-2 NDVI data for the past 5 years which is calculated from more than 1000 granules and aggregated to monthly medians. With the long time period covered by MODIS and the high resolution of Sentinel-2, areas and time periods are detected, for which vegetation cover is low and erosion risk is increased. The ESA-CCI S2 prototype landcover (LC) map with 20m resolution is used to relate the vegetation dynamics to land-use and land-cover classes. In a second step, the local precipitation pattern is analyzed. For this, the daily product of Rainfall Estimates from Rain Gauge and Satellite Observations (CHIRPS) data set is used. With this data, we are able to determine periods of drought but also short-term strong precipitation events. When strong precipitation and low vegetation cover affect a region in our study area, we consider this a high-risk region. As an additional indicator for soil erosion and soil erosion intensity, turbidity information of Lake Kivu and of Lake Tanganyika is analyzed. The turbidity analysis is based on the Lake Water Quality (LWQ) product of the Copernicus Land Monitoring Service (CLMS). For exemplary cases, we will show whether an increased erosion risk, determined by precipitation and vegetation pattern, leads to an increase in lake turbidity in the following days, which can be considered as proof for soil erosion. The analysis shows the capabilities and constraints of using remote sensing data to assess soil erosion risk parameters on a spatio-temporal scale without considering further relevant parameters like topography or soil types. Furthermore, the comparison to land use and land cover information can support the identification of land use practices that lead to an increase in soil erosion and, hence, to a loss of fertile topsoil. The presented approach contributes to a better understanding of soil erosion risk parameters and is conducted with freely and globally available remote sensing data which makes it scalable and transferable to other regions of the World.