人口数据空间化研究进展

人口数据空间化旨在揭示人口在地理空间上的分布位置及数量信息,展现人口统计数据的地理学含义,其研究已经成为人口学、地理学、GIS领域的研究热点。人口空间数据库在各级政府部门的规划和决策、灾害评估、资源配置等方面,具有重要的应用价值和科学意义。经过近30年的发展,人口数据空间化研究水平逐渐成熟,模型丰富多样,已获得众多成果。为把握人口空间化研究的研究现状,本文首先依据研究目的、建模思想及模型原理的异同,从3个方面对人口空间化研究进行梳理：（1）格网大小（尺度）的确定;（2）3种常用建模思想及6类主要模型的对比分析;（3）提高人口空间化精度的措施及其应用背景、优点。在此基础上,依据现阶段人口数据空间化的研究内容,从格网尺度适宜性研究、高时空分辨率人口空间分布模拟、引入新型数据源及多思想多模型综合应用等方面探讨了人口数据空间化的研究方向。

Research Progress and Perspective on the Spatialization of Population Data

The research purpose of the spatialization of population data is to capture the size and the distribution location of population in the geographical space. It plays an important role in presenting the geographical meaning of demographic data. Spatializing the statistical population data has increasingly become a research hotspot in the fields of demography, geography and GIS. Population distribution dataset is a key achievement in the spatialization study. At present, there are a few widely-used population distribution datasets and influential population spatialization projects, including GPW/GRUMP, LandScan and UNEP/GRID & China km grid population datasets. Population distribution dataset has practical application values and the scientific significance for relevant researches, such as government planning and decision making at all levels, disaster assessment and resource allocation. After nearly 30 years' development, the spatialization researches are evolving into the maturity stage. They have obtained many achievements and produce a rich variety of spatialization models of population data. Based on the purpose of spatializing census data and the differences between modeling concept and model principle, this paper reviews the spatialization methodologies in three major aspects: (1) the method and characteristic of the selection of grid size (scale); (2) 3 types of common adopted modeling ideas and a comparative analysis between 6 types of basic models; and (3) the proper strategies used for improving the simulation accuracy and their application background and advantages. Finally, according to the research contents of population data spatialization at present stage, this article discusses the further study direction through four perspectives: (1) the suitability of grid size; (2) the simulation of spatial distribution of population at high spatial and temporal resolution; (3) the adoption of new type of data source; and (4) the comprehensive application of multi-thought and multi-model. It is significant to grasp the current status of spatialization research and promote the further development of spatialization methodologies.