Abstract
This study examines the spatial relationship between land cover change and its drivers at varying scale in Mt. Guna. The recent land cover map was generated from Google Earth Image (2018), and the historical land cover map generated from the 1957 and 1980 aerial photos. Multi-scale geographically weighted regression (GWR) and ordinary least square (OLS) were used to model the spatial relationship between land cover change and its drivers at varying scale. The change analysis revealed that Mt. Guna showed a dynamic land cover change between 1957 and 2018 dominated by conversions into cropland. Overall, 69% of the land cover change shows gains and losses, while 31% of the land cover of Mt. Guna showed persistence over 61 years. Furthermore, 77% of land cover was swapped and changed from its initial state. The result of modelling the spatial relationship between land cover change and its drivers showed that population density, slope gradient and clustered homestead density increases the likelihood of land cover change, while higher elevation, water source density and precipitation reduces the likelihood of land cover change. However, the spatial scale comparison indicated that the influence is stronger in large spatial scales than in small scales.