Using remote sensing and census tract data to improve representation of population spatial distribution: case studies in the Brazilian Amazon

Abstract

This work proposes a methodological approach to redistribute population data obtained from polygonal census tracts into population density surfaces (grids) based on a cell space database. The methodology was first developed for the municipality of Maraba, Para state, in the Brazilian Amazon. We used a dasymetric method to eliminate areas of environmental restriction to human presence; then integrated environmental data indicative of human presence to generate a potential surface of population occurrence; and finally, census population count data were redistributed into cells. The methodology was subsequently adapted for 13 municipalities of the Sustainable Forests District (SFD) of BR-163, generating population distribution surfaces for 2000 and 2007. The evolution of the resident population over the SFD-BR163 showed spatial patterns compatible with the occupation process described in the literature and verified by fieldwork. To be applied over other areas, the proposed methodology must be adapted with local parameters but in this way, population density surfaces can be useful as an additional data source to study population and environment relationships.