Estimating population surfaces in areas where actual distributions are unknown: dasymetric mapping and pycnophylactic interpolation across different spatial scales

Spatially distributed estimates of population provide commonly used demand surfaces in support of spatial planning. In many countries, spatially detailed population summaries are not available. For such cases a number of interpolation methods have been proposed to redistribute summary population totals over small areas. Population allocations to small areas are commonly validated by comparing the estimates with some known values for those areas. In areas where spatially detailed estimates of the population do not exist, that is where the actual population in small areas is unknown, such as Nigeria validation is problematic. This research explores different interpolation methods applied at different scales in areas where the actual population distribution is known and where validation is possible. It then applies the parameters developed from these results to areas where the distribution is unknown. The binary dasymetric method using land cover data derived from a classified 30m spatial resolution satellite imagery as the ancillary data input and with disaggregation over 30m support grids, was found to provide the best target zones estimates of the population. The demand surfaces were then used to evaluate current health facility locations and then to suggest alternative spatial arrangements for health centres in Port-Harcourt, Nigeria. The average distance from each demand point to the nearest healthcare centre was found to be 1204m. When alternative locations for the current health centres were identified, the results suggest 13 service provision points would provide almost the same demand coverage as the 17 current PHCCs. This research develops methods that can be used to support informed decision making in spatial planning and policy development