Comparison between different methods of zonal interpolation for population estimate: Case study of the Federal District urban areas

Abstract

The data about population characteristics, normally released by the census, are of paramount importance because subsidize numerous social, economic and environmental studies. These data are aggregated arbitrarily considering the homogeny information in a given space. An alternative to work around this problem and generate information with more accurate spatial quality is the use of dasymetric mapping. The dasymetric mapping uses auxiliary data to refine the representation of the spatial distribution of the variable analyzed. During the last decades it has become a widely used technique, and many studies have been conducted with different methodological approaches. Thus, this study aims to evaluate the performance of three different approaches of dasymetric mapping to estimate the urban population of the Federal District. For this purpose, we used the dasymetric model per areal weighting (MD1);using subjective relative density (MD2) and; intelligent dasymetric method (IDM) with relative density sampled by the centroid method (MD3). For the generation of maps census data were used with the population value in the sub-district level. The land cover map used has a resolution of 1m and was modified of the mapping of urban structures types (UST) of the Federal District, which ranks the urban area in relation to its homogeneity considering aspects of functionality, covering material and physical characteristics. For the evaluation of the generated methods used statistical methods and graphs, using as comparison information of the population at the level of census tracts. According to the analysis applied the model demonstrated the best performance was the MD3, with this technique applied to the census tracts information level as final statement of estimated of the spatial distribution of the urban population of the Federal District.