【作者】 陶健；

【导师】 徐跃通；

【作者基本信息】 山东师范大学 ， 地图学与地理信息系统， 2010， 硕士

【摘要】 土地系统作为人-地作用最紧密的自然资源系统,其发展演化是人-地相互作用、相互影响的时空动态变化过程。矿山作为人类工程活动对地质影响最为强烈的环境之一,其土地利用变化以及扩张异于其他区域,不仅仅表现为人类直接作用于土地系统,影响、改变整个土地生态系统及景观,更交织了土地系统对人类生存环境的反作用影响。从地理模拟系统的发展及其应用入手,介绍了现代社会发展对地理学发展的新要求推动了传统描述性地理学发展以地球信息科学为代表的现代地理学定量研究手段,然后介绍了作为新一代地理模拟系统的元胞自动机的基本概念,并介绍了元胞自动机与地理网格系统的相似性,说明了其在地理学中的适用性。根据国家顶级驱动力决定根本驱动力推动直接驱动力(人口、经济等)驱动土地利用变化理论,研究时段选择从我国社会主义市场经济体制改革时期附近开始,据此从已有数据中选取连续等间距数据,并根据研究实际情况在前人研究的基础上对其进行重分类;运用数学统计方法定量化分析研究区土地利用年际变化的规模、速度以及幅度,并综合运用基于地学信息图谱理论的地球信息科学技术手段分析研究时段内南屯矿区土地利用时空格局演化及其空间圈层结构变化,总结研究区内土地利用扩张的过程、动态、结构及机理;采用统计和地统计方法提取规则,同时结合空间分析认识确定限制条件,采用元胞自动机方法对南屯矿区土地利用变化进行模拟,首先根据已有数据对模型进行精度检验,表明模拟结果的数量精度以及空间相似度都能较好的反映实际情况,然后模拟了研究区2017年土地利用状况并对其变化趋势进行分析。通过上述静态数学模型定量描述评价与地球信息科学技术时空动态分析相结合的步骤分析了1987~2007年南屯矿区土地利用变化过程,总结了南屯矿区在这20年间土地利用时间、空间变化总体趋势以及空间圈层结构变化,并预测了未来变化趋势。研究区基础设施建设用地增长迅速,尤其是在矿区1997～2007年扩大再生产阶段,其中生产建设用地在各圈层中都在增长,矿区采空塌陷“迫使”生产建设用地逐步向矿区边缘推移,形成新的居民聚集区域,这就使得生产建设用地的距离谱曲线由单峰态逐步转变为双峰态;采煤活动导致的采空塌陷现象严重,20年内塌陷地面积扩大了3.35倍,并“逼迫”其他用地类型向矿区边缘方向发展,使得矿区中心3000米范围内土地利用变化极不稳定;当地生态重建工作初见成效,塌陷地面积扩张呈减小趋势,矿区中心3000~4000米处分界线差距变小,用地类型结构开始趋于合理化与均衡化。在系统阐述地理模拟系统、元胞自动机的基础上,详细论述了本次研究所用模型方法。同时采用基于数学统计的马尔可夫预测方法以及元胞自动机预测方法对研究区2007、2017年土地利用情景进行预测,并利用2007年模拟结果与实际情况对比分别进行精度分析。由预测结果与实际值对比(表4-12)可知,马尔可夫定量预测总体精度为84.9292%,元胞自动机模拟的总体精为81.5357%(Overall Accuracy=(10629/13036)),Kappa系数为0.6609。两者基本都能够较好的模拟矿区土地利用变化,但在数量上后者的精度小于前者。本研究为南屯煤矿区人类经济社会发展提出预警,并对煤矿区可持续发展提供科学的空间决策依据。最后运用元胞自动机对研究区2017年土地利用进行空间定量模拟可以为该区域未来经济发展及生态环境重建提供参考。?更多还原

【Abstract】 As one of natural resource systems, land system has the closest man-land relationship, which is changed dynamically by interleaving interactions between man and land. As one of the most strongly influenced environments by human activity, mining area environment differ from all others, especially in land use change, which not only has uniqueness in characteristic, process and mechanism of dynamic change, but also interleaves interactions between man-land relationships.Beginning with the developments and applications of Geographical Simulation Systems (GSS), theories of combining descriptive and quantitative (represented by Geo-Information science in modern geography) methods were introduced, then basic concepts of Cellular Automata (CA), which is regarded as the new generation of Geographical Simulation Systems (GSS), and highlighted the similar aspects between Cellular Automata (CA) and Geographical Grid Systems to outstand its applicability in geography.Based on the theories, fundamental driving forces depend on national climax driving forces, and effect according to direct forces, which are put forward by Liu Jiyuan on the basis of studies on spatial - temporal dynamic changes of land use and driving forces analyses of China in the 1990s’, data of equal-time interval was picked up from existing data, and reclassified according to actual situation.Changing scale, speed, extent quantified by mathematical statistical methods as well as characteristics of spatial-temporal pattern and spatial ring structure of land use were analyzed by geo-information science technology based on Geo-information Photographical Methodology from 1987 to 2007 in Nantun coal mining area separately to summarize process, structure, dynamic and mechanism of land use in study area.Using geo-statistical/statistical model and determined rules by spatial analysis, land use change was simulated by Cellular Automata (CA) in Nantun coal mining area. The result of that in 2007 was verified with existing data, and then analyzed the trend of change from 2007 to 2017 according to the prediction of 2017.Combing quantitative description from static mathematical statistical methods and spatial-temporal dynamic change from geo-information science technology, process of land use change was analyzed to summarize general trend of spatial-temporal expansion and spatial structure change. Last but no means the least, predicting was made to analyze the trend of future.During the study period, infrastructural land increased rapidly, especially in stage of expanded reproduction from 1997 to 2007, in which production and construction land rose in all rings and compelled to move to edge becoming new residential area, whose curve converted from single-peak to double-peak; Trend of extremely unstable in land use change was caused by mining collapse aggravating, which has enlarged 3.35 times during that 20 years, and tressed all others far away from center; Gratifying achievements in ecological reconstruction has success initially, manifested morphologically as: trend of decreasing in mining collapse, stabilizing in land use of all rings and rationalizing and equalizing in structure of land use.Based on the description of Geographical Simulation Systems (GSS) and Cellular Automata (CA), model in this paper was described detailedly. Situation of land use in 2007, 2017 was simulated by Markov and Cellular Automata (CA) simultaneously, and the result in 2007 was used to validate the simulations of the two. The validation showed that the accuracy of Markov was 84.9292% while that of Cellular Automata (CA) was 81.5357%, and the Kappa coefficient was 0.6609. Both of these two methods simulated land use change preferably, however, the accuracy of the latter was better than that of the former.This paper provided early-warning for production and life in study area and basis of spatial determination. Moreover, the prediction supplied the programming of economic development and ecological reconstruction.更多还原