【作者】 王志伟；

【导师】 陈全功；

【作者基本信息】 兰州大学 ， 草业地理信息， 2010， 硕士

【摘要】 人,社会构成的主体。而描述人在地理空间分布的构成形式就是人口分布,这种人口的空间分布形式既有历史传承下的特性,也有当今生产力发展水平影响下的因素。它是自然环境和社会经济状况的反映,对自然环境演变和社会经济发展产生深远的影响作用。研究人口分布,将其同地理信息系统相结合,可以更好的揭示人口分布变化、特征、规律及其背景,为相关自然、人文科学研究提供研究基础。人口分布空间化方法主要有面积内插和空间建模两类。面积内插法将人口密度在研究区内取平均值,存在不能区分研究区内部差异,研究区之间差异变化过于显著等缺点。空间建模方法考虑了多种因素对人口分布的影响,但存在指标过多、难以完善说明彼此相关性的问题。利用插值方法中的点插值方法比传统方法平均人口密度得到的人口空间分布更符合实际。同时在解决大系统的单一问题时,仅仅用多因子的一次综合模拟并不准确,模拟因子过多,不宜取值,因子间关系不明确,最终不能执行插值,没有结果。研究表明,人口分布空间化时先进行插值得到一个粗略的可检验的结果,然后对此结果以多因子进行修正,会更加易行,更加准确。本文首先以2000年人口普查数据等为基础,分别比较了几种常用插值方法在研究区(巴彦淖尔市)人口分布空间化中的应用,研究表明：五种面插值方法得出的研究区人口总数与研究区统计数据均相差不到0.1%,但是存在相邻两地区在行政区划边界线内外数值非线性巨变的缺点；行政中心点控制下的点插值方法虽然能够克服前述缺点,但是得出的研究区人口总数是统计人口总数的4倍多,此结果显然不能作为人口分布结果；多点控制下的点插值方法得到的人口总数是统计人口总数的3/4,虽然控制点已经足够多,但是仍不能克服前述缺点,也不能作为人口分布的结果；研究区外围调节点控制下的点插值方法得出的人口总数占统计人口总数的98.22%，同时能够克服前述缺点。以这种方法对全国人口数据进行人口分布空间化插值,分布结果符合胡焕庸先生提出的：我国人口分布在爱辉-腾冲线以东的地区属人口密度极高区,在其以西的地区除少数地段外都属极稀区。再将全国人口密度和农牧交错带进行栅格赋值比较,得出人口密度同农牧交错带的一致性为0.6413。然后在西汉数字化地图的基础上,结合西汉人口数据,完成西汉时期人口密度分布和分级图。得出西汉时期我国人口分布空间格局是人口多集中于黄河中下游地区,南方人口较稀少的结论。并在此基础上对其成因进行探讨,西汉时期人口分布的空间格局是在自然因素、政治因素和历史因素的共同作用下形成的。更多还原

【Abstract】 Human beings, who constitute the main body of the community, its description of the geographical distribution are population distribution. The form of spatial distribution of population possesses historic properties and it also under the influence of the current productivity factors. It is a reflection of the natural environment and socio-economic status. At the same time, population distribution has a profound effect on natural environmental evolution and socio-economic development. Population distribution which is combined with geographic information systems can better reveal the changes, characteristics as well as its background. It laid a solid foundation for the relevant natural and human science research.The methods of population spatial distribution can be mainly attributed to two: area interpolation and spatial modeling. Interpolation area of population density is an average value in the area which was used to study. There are many shortcomings for this method. It not only does not distinguish the differences within the study area but also the change between the different study areas is too significant. Spatial modeling considers that a variety of factors impact on the population distribution, but there are still too many indicators and hardly prove the problems which have a closely relationship. The result of population interpolation is far more realistic than the average population density traditionally. By the way, using many factors to solve a single and large-scale system just at one time is not accurate enough. Because if there are too many factors, which is not easy to get value and figure out the relationship between each other. Population interpolation can get a rough result which can be tested, then modifying the result, which is easier and more accurate than other methods.Several widely used methods were compared in spatial distribution of population for BayanNur city by us, which were based on Census Data of 2000. The results showed that:The gap between the research area gross population that obtained by five surface interpolation methods and the research area statistical data does not over 0.1%. But it has the shortcoming that values of neighboring two areas have the non-linear great change in theirs administrative regionalization boundary inside and outside. Although the administrative central point as control point spot interpolation method can overcome the fore cited shortcoming, however, the research area gross population obtains from it is more than 4 times in construct to its statistics gross population. Obviously, it cannot be taken as the result of population distribution. The method of many spot as control point spot interpolation obtains the research area gross population account for 3/4 of its statistics gross population.Even though control points have already enough for this method, whereas it still could not overcome the shortcoming which was cited in the former content. Simultaneously, it also cannot be taken as the result of population distribution; The research area periphery adjustment points as control points spot interpolation method obtains the research area gross population take 98.22% of its statistics gross population, and it can overcome the fore cited shortcoming. The result about the distribution of population is consistent with the conclusion which was put forward by Mr. Hu Huanyong. He pointed out that the distribution of population in the east of Aihui-Tengchong line is an extremely high population density area; the area which located at the east part of it is all extremely rare to see human beings, that is to say, its density is a small number. Hence we use the method carries on the population distribution specialization interpolation to the national population data. Finally, our grid nation population density chart and the farming and animal husbandry interlock the belt chart and then carry on 22 comparisons for each grid evaluation. The result showed that the population density and the farming and animal husbandry interlock the belt the uniformity is 0.6413.Based on the digital map and the population data of Western Han Dynasty, we used GIS completed the map of distribution in population density and population density classification chart. From the two maps and related information, we can easily infer the final result that the spatial patterns of China’s population in Western Han Dynasty is larger in the middle and lower reaches of the Yellow River region and smaller in south of China. We researched the reasons on the basis of the result. It is evident to be seen that it formed under the huge role, which including natural factors,political factors and historical factors and so on.更多还原