Areal interpolation of population counts using pre-classified land cover data

The need to combine spatial data representing sociodemographic information across incompatible spatial units is a common problem for demographers. A particular concern is computing small area trends when aggregation zone boundaries change during the trend interval. To that end, this study provides an example of dasymetric areal interpolation using the pre-classified land cover data available through the US Geological Survey's National Land Cover Dataset (NLCD) program. Areal interpolation of population estimates is preferable to traditional reaggregation techniques, and the use of land cover data as a weighting factor in interpolated estimation has been shown in earlier studies to be highly accurate. In this study, the NLCD data set performs well and, because it requires no classification, it compares favorably with other land cover data sets for areal interpolation when considered on the basis of accuracy, precision and ease of use.