Application of Geospatial Techniques and Logistic Regression Model for Urban Growth Analysis in Limbe, Cameroon

Abstract

Little is known about the nature of ecosystem loss, rampant changes in land use and land cover (LULC) and urban growth taking place in Limbe. The aim of this study is to analyze urban growth in Limbe, Cameroon from 1986-2019 using geospatial techniques and Logistic Regression Model (LRM). Landsat Thematic Mapper (1986), Enhanced Thematic Mapper+ (2002) and Operational Land Imagery/Thermal Infrared Sensor (2019) were utilized in this study. The images were classified into land cover classes using supervised image classification algorithm in ENVI software. The classification output was subjected to LRM application to evaluate urban growth. Image difference of urban growth between 1986 and 2019 was calculated as dependent variable and the independent variables were produced by calculating the Euclidean distance and Buffer of built-up, waterbody, road and farmland as driving factor for urban growth. Future urban growth was determined for 2035 using the Land Change Modeler in IDRISI Selva. Classification overall accuracy for the three date were not less than 99%. LRM results show a good fit with relative operation characteristic of 0.8344 and Pseudo R2 of 0.21. Analysis of LULC shows that built-up increased from 3.5% (1986) to 17.6% (2019). An urban land expansion rate of about 23% was observed for 2035. Transition probability matrix revealed high probability (0.6345) of build-up to remaining build-up by 2035, while the probability for it changing to waterbody, bare land, farm land and vegetation are 0.1099, 0.0459, 0.1939 and 0.1221, respectively. This study successfully demonstrates the application of geo-spatial techniques and LRM for land use/land cover change detection and in understanding the urban growth dynamics. It also identifies the potential areas of future urban growth, which can help land use policy planners for making optimum decisions of land use planning and investment.