MODELING LAND DEGRADATION FOR CONSERVATION PLANNING IN KALEHE TERRITORY, EASTERN D.R. CONGO

Abstract

The Eastern D.R. Congo is experiencing land degradation resulting from unsustainable land use which prevent the achievement of the land degradation neutrality in this region. This thesis aims to model the land degradation (LD) for conservation planning of natural resources in this region by using the Kalehe territory as a case study. Based on the system theory, a mixed approach combining field surveys, GIS, and remote sensing techniques was adopted. The geospatial data were used to assess the LULC changes, their implications on land productivity (LP), ecosystem service value (ESV), and soil erosion dynamics. Furthermore, a multi-criteria decision analysis (MCDA) based model was developed to assess the land degradation vulnerability (LDV). The results of these geospatial analyses were triangulated with community perception data to identify the DPSIR (Drivers-Pressures-State-Impacts-Responses) indicators toward the LD management. By analyzing the LULC changes from 1987 to 2020 using the Landsat images and forecasting the future LULC for 2030-2070 through Markov modelling, the study identifies trends of increasing built-up, shrub land, and cropland at the expense of forestland, grassland and wetland. These changes contributed to 34.17% of land cover degradation over the last three decades. The analysis of LP dynamics through the linear trend analysis of NDVI time series data reveals that 31.25% of the territory has experienced a decrease in LP. The assessment matrix was used to link the perceived ESV and the LULC categories. Through this approach, it was demonstrated that the potential supply of ecosystem services decreased in 28.44% of the land over the 1987- 2020 period. The assessment of soil erosion dynamics through the RUSLE model indicated that the mean annual soil loss has increased over time from 32.08 t/ha/year in 1987 to 44.35 t/ha/year in 2020. Under the current LULC trend, the annual soil loss is projected to increase to 46.42 t/ha/year by 2030, 46.79 t/ha/year by 2050, and 48.38 t/ha/year by 2070. The adoption of conservation practices would result in the reduction of the current erosion rate by 86.56%, 62.28%, 54.05%, and 11.61% for bench-based terracing, agroforestry, strip cropping, and contouring, respectively. Moreover, the LD dynamic is influenced by the landscape characteristics since the decrease of forestland, and patch’s shape complexity, the increase of patch’s isolation, landscape heterogeneity, and fragmentation positively influenced the soil erosion dynamics. Hence, the need for land consolidation, connectivity and forest conservation during the future land use planning. This study also demonstrated that the LDV model based on MCDA can be used to predict the occurrence of physical LD processes in eastern DR Congo with an accuracy of 77.82%. Thus, it can be supplemented with the outcomes of land capability (LC) analysis for restoration planning and adaptive land use planning to reduce the LDV. To address the challenges of LD, this study proposes a conceptual model of LD management and a conservation action plan including the sustainable land use according to LC, implementation of conservations practices, environmental education, and improvement of community livelihoods.