【作者】 王英安；

【导师】 张林泉；

【作者基本信息】 山东师范大学 ， 人口·资源与环境经济学， 2003， 硕士

【摘要】 人口是区域可持续发展的一个重要因素，我国又是一个人口大国，因此，对于区域人口指标的获取方法始终都是学术界研究的热点问题之一。对于人口指标的统计，目前大多采取人口普查和抽样调查的方法来计算某一区域内当前的人口数量。由于我国面积辽阔，自然条件多种多样，因此人口的抽样调查数据并不能准确的反映全国各地人口的实际分布状况。而人口普查又是一个规模巨大的系统工程，牵扯面广，需要耗费较长的时间和巨大的人力、物力和财力，同时更新周期比较长，不能获得及时、动态的人口数据。为此，本文将着重研究利用中国模拟人口密度模型(SPD)来反演估算中国的人口密度空间分布状况。中国模拟人口密度模型是由中国科学院地理科学与资源研究所首先提出的，这个模型试图利用先进的格网生成技术，将城市、交通基础设施等社会经济因子和陆地数字高程(DEM)、陆地植被净第一性生产力(NPP)等自然因子相结合，通过模型的反演，模拟某一时期的中国人口密度空间分布状况。这种模型反演的方法，突破了传统的按行政区界线统计人口密度的方法，改为按照均匀分布、规则大小的格点单元来计算人口密度，丰富了获取人口密度指标的方法，提高了人口密度指标的精确程度和应用范围，将有利于人口、经济政策的决策过程，有利于区域可持续发展。 本文的主要内容分为三个大的部分： 第一部分系统论述中国模拟人口密度模型(SPD)的理论基础和技术方法，模型的各个变量因子、计算公式和模拟结果。在这一部分中，首先简单介绍了基于格网生成技术的人口密度空间分布模拟的意义和理论基础，然后回顾了该领域目前国内外的研究现状，接着详细阐述了SPD模型中各个主要变量因子的选取、预处理和计算方法，最后给出了我们对于中国人口密度空间分布的模拟结果。 第二部分主要是采用几种典型的误差分析方法，主要包括相关性分析和回归分析，对模型的原始输入数据的时间误差、空间误差和统计数据误差进行了相对误差和绝对误差的分析，然后又对模型中各影响因子进行了相关分析和回归分析，最后综合以上误差分析的结果得出该模型的理论精度为87％。 第三部分主要论述中国模拟人口密度模型(SPD)模拟运算和误差分析过程中开发使用的软件系统。该系统全面采用了基于元数据库管理的思想和面向对象技术，并在此基础上构建一个简单的模型库管理系统，初步建成了一个集模型管理、误差分析和图表输出为一体的集成计算系统。这一部分详细论述了该软件系统的结构、功能和集成方法等内容。 本文的主要目的是探索利用格网生成技术和数据融合理论表达区域人口密度分布这一重要经济要素的新方法。主要研究方法是借助计算机工具实现人口密度指标的模拟和空间表达，主要成果是在应用国际上较为先进的数据融合理论，在引入格网生成的空间数据计算和表达方法的基础上，模拟生成了中国的人口密度空间分布图。本文最后还提出了今后基础理论和技术应用方面需要进一步改进和深入研究的问题。 总之，中国模拟人口密度模型(SPD)以数据融合理论为指导，综合应用了多尺度数据融合和多源数据融合，方法上使用目前国际上推崇的格网生成技术，完成了中国人口密度空间分布的模拟。这些新的理论和技术都是可持续发展信息系统领域研究的前沿问题，并在实际应用过程中具有一定的创新性。更多还原

【Abstract】 The population is an important factor for the sustainable development of one region and our country has so much people, therefore, the method research to obtain the region population indexes is one of the focus problems in the academic field. Statistics about population indexes, is taken by the census and the method of the sampling mostly now to compute with a current population in some district. Because our country is broad and natural environment is various, the sampling population data can’t reflect the condition of national population distribution accurately. But the census is an enormous system engineering in a scale and it need longer time with the enormous manpower, material resources and financial power. It has longer renewing period, so we can’t acquire the dynamic population data in time. For this, this article will emphasize on the research of estimating the spatial distribution of Chinese population density by the Simulated Population Density model (SPD). The SPD model is made by the Institute of Geographical Sciences and Natural Resources Research of Chinese Academy of Sciences. It combines social economic factors, such as city and transportation foundational facilities, with nature factors, for example the Digital Elevation Model (DEM) and Net Primary Productivity (NPP) of plants to simulate the spatial distribution of population density of China by the advanced grid generation technology. The method of the model retrieval breaks down the traditional population density statistics method according to the administrative area boundary and changes to evenly distributed and same size grid units to compute the population density. It enriches the methods to obtain the population density indexes and increases the accurate degree and application fields of the indexes. So, it will benefit to the decision process of the population and economic policy, benefit to the sustainable development of regions.The main contents of this article is divided into three parts:In the first part we systematically discuss the theoretical foundation and the technique of the Simulated Population Density model (SPD), explain each one of the model variables, the formulas and the simulation result. In the part, we first introduce the significance and theoretical foundation of the simulation in the spatial distribution of population density basing on the grid generation technique. Then we review the current condition of domestic and international research in that field, detailedly expatiate the selection, pretreatment and computation method of the main variables in the SPD model.In the second part, we adopt a few typical error analysis methods, primarily including the relativity analysis and the regression analysis, to analyses the relative error and absolute error of the time error, spatial data error and statistic data error of the raw input data in the model. Then, we take each influence factors in the model into the relativity analysis and the regression analysis. Finally, we synthesize the results of the above error analysis to figure out the theoretic accuracy of that model as 87%.The third part is mainly about the software system developed and used in the simulating calculation and error analysis process of SPD model. The software adopted completely the thought of metadata database management with the Object Oriental Programming (OOP)technique, set up a simple modelbase management system, got out a integrated computing system for models management, error analysis and charts output and so on. In this part we detailedly discuss the framework, functions and system integrated methods of the software.The main purpose of this article is to research the new method to express the spatial distribution of region population density that is an important economic element, by using the grid generation technique and the data fusion theory. The main research method is to achieve the simulation and spatial expression of the population density index with the help of computer tools. Finally, We got the map of the spatial di更多还原