| Abstrakt: |
Land use and land cover change is driven by a complex interaction involving multiple environmental, climatic, and social factors. This study employs time-series satellite imagery from Landsat-5 Thematic Mapper (1995 and 2009) and Landsat 8 Operational Land Imager (2022) to examine LULC changes in Tanzania's Dar es Salaam Metropolitan City. Image pre-processing was performed using the Google Earth engine code editor, followed by LULC classification using Random Forest machine learning in R software. Herein, seven LULC classes were defined: Built-up areas; agriculture; forest; bushland; water; grassland; and bare soil. The overall accuracy and Kappa coefficients for the years 1995 were 81.40% and 0.77, for 2009 were 88.42% and 0.86, and for 2022 were 81.51% and 0.77, respectively.. Results showed that between 1995 and 2022, the built-up and agricultural land areas increased by 14.87% and 4.47%, respectively. In contrast, there was a significant decrease in bushland, grassland, forest, and bare land areas (14.57%, 1.75%, 2.9%, and 35%), respectively. The study identifies factors influencing these changes and employs a Cellular-Automata-Markov-Chain (CA-Markov-Chain) model to evaluate and predict LULC changes. Distance from the city centre, population density, and precipitation are major drivers for LULC change. The CA-Markov Chain analysis predicts that the built-up area and agricultural area will grow from 319km2 (19.30%) to 530 km2 (32.06%) and 471 km2 (28.49%) to 720 km2 (43.56%) between 2022 and 2050 respectively, based on the trends observed from 1995 to 2022. Therefore, policymakers should prioritise sustainable urban planning in Dar es Salaam, considering drivers of land-use change and predicted growth in built-up and agricultural areas. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink