Simulation of Chinese Population Density Based on Land Use
暂无分享,去创建一个
Land use data integrates lots of information of factors affecting population distribution. It is the nature of land, the household responsibility system, the household registration system, and the production mode of agriculture in China that establish a close spatial relation between land use and population distribution. According to the idea of modeling separately by town and country, by ecological zones along with the population of counties as restrictive conditions, we build the following model based on land use to simulate the population density in 1 km square grid-cells of China: POP■ = P■×V■■V■ + P■×V■■V■. Linear weighted model is used to calculate rural population indices. It firstly picks out the indicators for the model which are correlated positively and remarkably with the population of counties in each of the 12 ecological regions, then stepwise regression is introduced to calculate the indices; and finally the weighted coefficients of various indices are determined in combination with the productivity of land and the relations between habitation and variant land. For the urban population indices, a power exponential model based on the scale of town and the distance from the center of town, rooted in distance decay function, is built as V■ = A■×lnA■×exp(-1.9874r■■A■■). The outcome of the simulation shows that 94.58% of the population inhabit in the southeast of Hu Huanyong population line, and its density is 21 times as that of the northwest. In the east, population is centralized in four regions and one zone. They are Huang-Huai-Hai plain, Sichuan basin, plains of the middle and lower Yangtze River, Northeast China plain and the zone of 100 km from coast. Moreover, a "points-axis" pattern of population distribution is discovered on the southeast coast. Because this paper has fully recognized the relation between land use and the population distribution, the outcome of the simulation proves to be highly credible. Compared with other models, this method is more efficient.