The computational model of multi-scale population density

A computational model of multi-scale population densities was developed based on the native population database of Shijiazhuang city. The model comprises four parts: data processing, down-scaling, data format conversion and up-scaling. Down-scaling was completed by apportioning village population to each street block. The vector dataset of block population was converted into 100m×100m grid dataset, which is the minimum grain population density. By changing the radius of circular filter factor, the up-scaling of population density was realized and the multi-scale population density was obtained. The results show that the long-term balance of human and nature environment can be clearly revealed on the maps of population densities at R = 0, R =3, R =12, R =29 and R =99. The model is helpful to observe the interaction of man-land relationship at different scales systematically, and to build a mathematical foundation for studying the MAUP (The Modifiable Area Unit Problem) of population density.