Deriving Spatially Refined Consistent Small Area Estimates over Time Using Cadastral Data

Areal interpolation is the process of transferring data collected over source zones to target zones. One of the applications of areal interpolation is to construct temporally consistent areal units for comparing socioeconomic data over time. Dasymetric modeling is the process of employing ancillary data to spatially refine the distribution of socioeconomic variables. In this paper, three areal interpolation methods, namely areal weighting (AW), target density weighting (TDW) and pycnophylactic (PM), with and without spatial refinement, are used to interpolate census tract populations in 2000 (source zones) to census tract boundaries in 2010 (target zones). The spatial refinement is performed using residential parcels. The study area is Hennepin County, Minnesota. Accuracy assessment is based on the interpolated population of each target zone and its benchmark population resulting from aggregating the population values of census blocks within it. According to accuracy comparisons, spatial refinement has potentials to improve areal interpolation results. However, the improvement level depends on the employed areal interpolation method.