Urban growth assessment using machine learning algorithms, GIS techniques, and its impact on biodiversity: The case of Sululta sub-city, Central Oromia, Ethiopia

Abstract

Ecological services in metropolitan areas are degrading more quickly due to changes in LULC brought about by urban expansion. To make a sustainable choice about the ideal location, however, merging the existing simulation approach with LULC optimization approaches involves several intricate procedures. Therefore, the goal of this study is to develop a unique technique that can forecast urban expansion over an extended period and to link with optimization of LULC techniques so as to make meaningful decisions on the impacts of urbanization on loss of biodiversity. In this study, three primary procedures were used: (1) an SVM-based supervised classification technique for LULC classification; (2) a Markov chain and cross-tabulation method for the examination of LULC trends in space and time, (3) utilizing the CA-Markov approach to forecast urban growth. By using Landsat imagery of 2008, 2015, and 2023, the study determined how urban cover changed over time, and what kind of LULC-to-urban transition occurred. The study revealed that uncontrolled and haphazard urban expansion was observed in the Sululta sub-city, which could have disastrous repercussions on physical, biological and urban ecosystem. The %age of urban area increased from 9.04% in 2008 to 13.07% in 2015. However, because of the internally displaced people from the Ethio-Somali Region, who have been resettled there since 2017, the ratio of urban areas grew from 13.7% in 2015 to 24.65% in 2023. Furthermore, by 2040, the sub-city will have grown by 27.69 %. The kappa coefficient statistics of the three classified images of the years 2008, 2015, and 2023 were 93.3 %, 97.5%, and 97.5 %, respectively. To identify the areas that would be covered by future city growth, it is advised that this innovative technique be integrated with optimizing land use strategies.