Prediction of Land Cover and Land Use Changes in the Greater Gaborone Area of Botswana

Abstract

Changes in land cover and land use (LCLU) have been observed in the greater Gaborone area of Botswana since the 1980s. These changes are driven by human and livestock population pressures, rapid urbanization, and socioeconomic and industrial development. Being able to predict LCLU changes is useful for policy and land use planning. Advances in geospatial modelling techniques have become central to the monitoring and assessment of past and future LCLU changes. This study employed the Module for Land Use Change Evaluation (MOLUSCE) to simulate LCLU in 2042 based on the current LCLU. Classified Landsat images from 1988 and 2002 and Sentinel 2A images were used to simulate the likely LCLU changes between 2022 and 2042. Model validation was done by comparing the Kappa statistics for the simulated and actual 2022 LCLUs. The Kappa coefficients for the actual and simulated images were 0.93 and 0.79, respectively. The simulation results showed a projected increase of 65.81 km2 and 28.19 km2 in the shrubland and built-up classes between 2022 and 2042, respectively, while the tree cover, cropland, bare land, and water body classes are expected to decrease by 56.69 km2, 25.05 km2, 11.89 km2, 21.89 km2, and 0.36 km2, respectively. These dramatic changes in future LCLUs could negatively impact the environment, and jeopardize water quality and quantity and food security. The results from this study provide useful information to land use planners to make better decisions about improving future land use policies within the framework of sustainable land use planning.