Tracking the Past and Predicting the Future Land Use/Land Cover Dynamics in Semi-Arid Regions: The Case of Giba Basin, Northern Ethiopia

Analysis of historical and future land use/land cover (LULC) dynamics using spatiotemporal data is crucial for better management of natural resources and environmental monitoring. This study investigated LULC transformations for the period of 60 years (1984-2044) in the Giba basin, northern Ethiopia. An ArcGIS and the Cellular Automata and Artificial Neural Network (CA-ANN) model were used to develop the past (1984, 2004, 2014, and 2024) and future (2034 and 2044) LULC maps of the basin, respectively. The results show that LULC categories experienced shifts from one class to other classes by 35%, 33%, and 40%, in 2004-2014, 2014-2024, and 2004-2024, respectively. During the 1984-2024 period, the smallest and largest percentage of positive changes were observed in settlement (7700%) and shrubs and bushes (25%), which increased from negligible to 78 km2 and from 1668 km2 to 2082 km2, respectively. Whereas barren land and forestland showed the largest (-80%) and smallest (-37%) decreasing rate during the same period. Results of the CA-ANN model showed that majority of the LULC categories (including water body, forest, bushes and shrubs, grassland, and barren land) will decrease except a slight increase of cropland (+6%) and settlements (+16%) in the future two decades during 2024-2044. In general, the Giba basin has undergone significant changes during the study period. High population increase, changes in government policies, and armed conflicts were found to be the most influential driving factors of changes in the basin.