Land Cover Change Detection and Prediction in the Fafan Catchment of Ethiopia

Land cover change is a pressing environmental issue that the Fafan catchment in Ethiopia is currently grappling with. In spite of several studies on the historical period, predictions on future land cover in Ethiopia in general and the Fafan catchment in particular are limited. In line with this, studies that integrate the local knowledge with remote sensing data are scarce. Thus, the purpose of the study is to examine the land cover dynamics and its drivers for 1990 to 2050, using a combination of satellite imagery and a socio-economic survey. To investigate the dynamics of land cover change, the study employed support vector machine, post-classification, multi-layer perception-artificial neural network, and cellular automata-Markov approaches. Thematic information extraction from satellite imagery was triangulated using local knowledge derived from key informant interviews (KIIs) and focus group discussions (FGDs). The results of the study revealed that the catchment includes six land cover categories, including cropland, settlement, barren land, forest, grassland, and shrubland. For the years 1990 to 2021, forest, grassland, and shrubland decreased by 13.2%, 4.6%, and 18%, respectively, while cropland, settlement, and barren land rose by 19.2%, 11.7%, and 4.9%, respectively. During this period, the net gain for cropland, settlement, and barren land was 30,705 ha; 18,541 ha; and 7,776 ha, respectively, while the net loss of shrubland, grassland, and forest was 28,675 ha; 7,294 ha; and 21,052 ha, respectively. Similarly, for the years 2022 to 2050, cropland, settlement, and barren land are predicted to increase by 9.1%, 3.5%, and 2.2%, respectively, while shrubland, forest, and grassland are expected to decrease by −1.3, −3.65%, and −10.1%, respectively. Furthermore, the findings of the study indicated that several factors have contributed to changes in land cover, including overgrazing, population growth, resettlement, wood collection, and cropland expansion. To this end, by combining socio-economic surveys and remote sensing data, this study has developed a reasonably accurate map of land cover changes. However, the use of very high-resolution satellite imagery, combined with local knowledge, could yield even better results than those obtained from Landsat imagery.