Grid transformation method of population data combining geographic factors and simulated township boundary adjustment

High-quality spatially referenced population information plays an important role in many social-demographic fields. This paper focuses on grid transformation method for population data by combining geographic factors and simulated township boundary adjustment. Given the location, area and census data of each town-level administrative unit and national base map (1: 25000) in China, 1km*1km gridded population data can be acquired after interpolation and several adjustments. Besides the adjustment based on geographical factors like topography, transport (roads), rivers and settlements, a new adjustment method based on township boundary simulation and total population of the town is proposed in this paper. The Voronoi polygon of town point is generated and rasterized into 1km*1km grids. Considering the area of each township and boundary line of the corresponding country, the collapse or expansion process of the rasterized Voronoi polygon is conducted pixel by pixel to minimize the differences between total count of grids with same township ID and the announced area of the corresponding township iteratively. After boundary simulation, gridded population data is adjusted based on town-level census data. The study indicated that the proposed method can acquire fine-grained grid population surface. It is demonstrated as an effective method to transform census population data into regular grids.