Abstract

Badri Mussa Omary, Edmund Mutayoba, Lusajo Mfwango, Nyemo Chilagane

Land Use and Land Cover (LULC) change is a major driver of landscape transformation, influencing ecosystem integrity, hydrological processes, and watershed sustainability. This study assessed three decades of LULC change (1994–2025) and predicted future landscape dynamics (2035–2055) in the Morogoro river subcatchment, Tanzania. Multi-temporal Landsat imagery (1994, 2005, 2016, and 2025) was classified using the Random Forest algorithm in a GIS environment, while future LULC scenarios were simulated using a Cellular Automata–Markov (CA–Markov) model. Classification accuracy ranged from 81.15% to 86.87%, with Kappa coefficients between 0.76 and 0.83, and the CA–Markov model achieved a validation Kappa of 0.83, indicating reliable predictive performance. Between 1994 and 2025, built-up land expanded by 322.65% (499–2,111 ha) and cultivated land by 171.95% (626–1,701 ha), whereas forest, woodland, shrubland, wetlands, and open water declined by 41.51%, 77.86%, 66.03%, 74.09%, and 89.07%, respectively. Transition analysis showed that agricultural expansion and urban growth were the principal pathways driving the conversion of forests, woodlands, and riparian vegetation into cultivated and built-up land. Under a business-as-usual scenario, the model projects continued declines in forest (53.63%) and woodland (36.66%) by 2055, alongside further agricultural expansion and near-complete loss of wetlands. These findings highlight increasing anthropogenic pressure on the Morogoro river subcatchment and underscore the need for integrated land-use planning, ecosystem restoration, and strengthened conservation measures to safeguard watershed functions and long-term environmental sustainability.