A methodological approach to estimate population density in a land-use cover change model - a case study in Bogota - Colombia

Abstract

In rapidly growing cities on developing countries, urban sprawl has been a trend. The city of Bogota and its adjacent municipalities have undergone densification processes during the last decades, especially on peripheral locations where low-income settlements have established. This study uses ordinary least squares (OLS) and geographically weighted regression (GWR) models to include density analysis in a Land-use cover change (LUCC) model based on cellular automata. Variables related to residential land-use and transport and urban oriented variables on housing built area were used to predict future population densities. The global regression model (OLS), is used to estimate statistically significant explanatory variables and overall model significance, while the local regression model (GWR), estimates parameters differently depending on spatial weighting among neighboring geographical zones. An indirect methodological approach was established to calculated population density on a Iow-resolution scale, using a well-specified GWR model that determine housing built area. The methodology proposed was tested in a case study in the city of Bogota. Results for Bogota suggest the methodology proposed improved the capability of LUCC models to derive densities. Further research is needed to better understand densification processes in the context of sustainable development for the city of Bogota and its region