Using Support Vector Machines to Analyze and Predict Land Use Land Cover Changes in Malemba Nkulu in the Democratic Republic of Congo

Kikombo Ngoy

College: Hennings College of Science Mathematics and Technology

Department: Environ & Sustainability Sci

Abstract:

Sentinel‑2 remote sensing data from 2017 and 2021 were used to analyze land use/land cover (LULC) changes and to predict future spatial distributions. A Support Vector Machine (SVM) classifier implemented in TerrSet 2020, a powerful machine‑learning modeling platform, was employed for this purpose. Seven land use/land cover classes were identified and mapped.The results revealed that Miombo tropical forest cover has been declining at a relatively high rate. This loss is primarily attributed to anthropogenic activities, including tree cutting for fuelwood and charcoal production, house construction, dugout canoe carving, and the widespread practice of shifting cultivation. In contrast, farmland and savanna woodland areas showed a noticeable increase over the study period. The expansion of farmland is mainly driven by rapid population growth and the resulting increase in food demand.If effective conservation and land management measures are not implemented, the predicted future land cover changes are expected to have significant negative impacts on the ecosystem. These impacts may include the disappearance of indigenous tree species, a decline in wildlife populations, alterations in soil texture and fertility, and changes in local and regional climatic conditions.