【作者】 李欢；

【导师】 刘沁萍；

【作者基本信息】 天津工业大学 ， 公共管理， 2024， 硕士

【摘要】 海岸带是海、陆、气交互作用的特殊地带,具有自然资源丰富、生态环境脆弱、人类活动频繁的特点。中国海岸带作为人类生产、生活的重要场所,面临着社会发展、环境保护矛盾和承载力不足等问题。探究中国海岸带土地利用变化特征及规律,并预测海岸带未来土地利用情况,能够有效促进生态保护、经济发展和社会效益的有机统一,为沿海地区多角度规划和发展提出科学性的对策和建议。本文以中国海岸带为研究区,揭示其1985—2021年长期的土地利用变化规律和特征,依据可持续发展目标15.1.1、15.2和15.3.1,评估中国海岸带关于可持续发展方案的实施情况。结合研究区实际情况,采用Logistic回归方法对多种驱动因子进行回归分析,分析驱动因子与各土地利用类型的影响关系及程度。利用Markov-FLUS模型预测2030年五种情景下中国海岸带土地利用情况,最后综合分析预测结果,对研究区域存在的问题和未来发展方向提出相应的建议。主要研究结论如下:(1)耕地和林地在中国海岸带土地利用类型中占主导,1985—2021年耕地、草地、灌丛和裸地呈减少趋势,不透水面、水域和林地呈增长趋势。从动态变化来看,2005—2010年研究区综合动态度最大,各地类变化程度剧烈。从地类转换和变化图谱来看,耕地向不透水面转换的变化比率最高,灌丛向其他地类转换的比率最低。(2)1985—2021年中国海岸带部分林地得到可持续管理,森林覆盖率上升,满足可持续发展目标15.1.1和15.2要求。研究区土地退化面积高于土地改善和恢复面积,退化率呈波动变化趋势。当前研究区与可持续发展15.3.1还存在一些差距,这是由于在中国海岸带城市建设过程中过度侵占耕地和林地,加剧了土地退化。(3)通过Logistic模型对自然因素(高程、坡度)、区位因素(到河流水系、铁路、公路、县级中心的距离)和社会经济因素(GDP、人口密度)八种驱动因子进行回归分析。结果显示各地类与这些驱动因子的ROC值大于0.7,表明本文建立的回归模型是有效的,所有驱动因子均对研究区各土地利用类型的空间分布产生影响,可适用于模拟预测。(4)Markov-FLUS耦合模型,在2018年土地分类数据基础上模拟2022年土地利用数据,精度结果Kappa系数为0.87,总体分类精度为92.15%。说明该模型的马尔可夫转移概率和各项参数达到满足预测要求,可行性较高。(5)本文设定“自然发展”“耕地保护”“生态优先”“可持续发展目标”和“政策引导”五种情景,模拟2030年中国海岸带土地利用变化情况。结果表明五种情景下的各地类扩张和缩减程度均得到了一定的调控,但存在土地类型比例失衡、空间布局不合理等问题。因此在未来土地利用过程中,需要综合考虑研究区实际情况和土地利用方式,提高土地利用率,从而促进人与自然和谐共生。更多还原

【Abstract】 The coastal area is a special zone where sea,land and air interact.It is characterized by abundant natural resources,fragile ecological environment and frequent human activities.Coastal zone of China plays a vital role in human production,life and ecosystems.While mankind promotes social and economic development,it has also caused land to be destroyed to varying degrees.Explore The characteristics and laws of the coastal zone of mainland urban land use spatial pattern change can effectively obtain the future development direction and distribution pattern of land use,so that scientific countermeasures and suggestions can be put forward.This paper takes Coastal zone of China as the research area,based on the land use classification data from 1985 to 2021,and uses GIS to analyze the spatial and temporal characteristics and processes of land use in the region over the past 36years.Land use change rate,transfer matrix of land use,geo-spectrum of land use change,indicators of United Nations’Sustainable Development Goals(SDGs)15.1.1、15.2 and 15.3.1 were used to analyze the characteristics of land use change and the progress in achieving the SDGs in the coastal zone of China.Combining the study area own conditions,select multiple driving factors that affect land use changes,and use Logistic regression to analyze the selected driving factors to explore the mechanism of land use change and the relationship between driving factors and it;according to different future development trends of the area,set five development scenarios,comprehensively use Markov-FLUS coupling models to predict the land use distribution pattern of the study area under the five scenarios in 2030,and finally put forward corresponding countermeasures for the study area based on the results.The main research conclusions are as follows:(1)Cropland land and forest land dominate the land use types in Coastal zone of China.During 1985-2021,cropland,grassland,scrub and bare land showed a decreasing trend,while impervious,water and forest land showed an increasing trend.The comprehensive dynamic attitude was the largest in 2005-2010.According to Geo-spectrum of land use change,the conversion ratio of cultivated land to impervious surface was the highest,and the conversion ratio of scrub to other land types was the lowest.(2)During 1985-2021,forest land has been sustainably managed,and the forest coverage rate has increased,meeting the requirements of 15.1.1 and 15.2 of sustainable development.The land degradation area was higher than the land improvement and restoration area in the study area,and the degradation rate showed a fluctuating trend.This is due to the rapid economic development of Coastal zone of China during the research phase and the urbanization process has led to an increase in the demand for construction land.(3)The Logistic model is used to analyze the driving factors.The results show that the ROC values of land use classification and the 8 types of driving factors are all close to 0.7,indicating that the regression model established in this paper is effective,and the selected driving factors can explain land use well,which can be further explain that the factors affecting land change do not exist independently.(4)Combining the Markov and FLUS model to simulate the land use distribution pattern in 2030 based on the current land use data in 2022.The simulation result is verified by the Kappa coefficient to be 0.87 indicating that the model meet the accuracy requirements,the feasibility is higher.(5)This paper sets five scenarios of"natural development","cropland land protection","ecological protection",“Sustainable development goals”and“political guidance”to simulate land use changes in Coastal zone of China in 2030.The results show that the land use under the five scenarios can be regulated to a certain extent,but the types of land use there are still problems such as unreasonable structure and layout.This study found that under the policy guidance scenario,all land types are basically stable,cultivated land and ecological environment are protected,impervious water surface growth rate is controlled,human-land relationship is more harmonious and symbiotic,and land use structure is more reasonable.更多还原