【作者】 刘刚；

【导师】 沈守云；

【作者基本信息】 中南林业科技大学 ， 生态学， 2011， 博士

【摘要】 湿地景观研究是近十余年来新兴的热点领域,湿地碳储量是研究全球气候变化和生态安全核心内容之一。中国是世界上湿地资源最丰富的国家之一,洪湖是我国第七大淡水湖,在长江中下游的湖泊湿地中具有典型性和代表性。对洪湖开展湿地景观和碳储量研究,对我国开展湿地资源生态安全、保护和合理利用等研究方面具有重要的示范意义。论文在3S技术支持下,收集了1974-2010年期间内洪湖市Landsat系列、CBERS02-CDC、ALOS遥感数据,共8个时间段的9期影像,建立了湿地资源空间数据库,结合地面相关辅助资料,运用面向对象的影像分割方式进行湿地信息提取；在此基础上并结合景观生态学原理,选取23个景观格局指数,开展洪湖湿地景观的时空变化、湿地演替分析,分析各种湿地类型的景观变化特征和转移变化。利用Markov和CA-Markov模型对洪湖今后一段时间内的景观变化进行了预测和分析,同时利用高分辨率的ALOS影像对整个研究区各种景观类型的碳密度和碳储量进行了估算。论文研究的主要结论如下：(1)参照《国际湿地公约》和国家林业局《湿地调查技术规程》,结合洪湖市湿地的实际情况,确定了洪湖湿地遥感解译分类标准。采用面向对象的信息提取方式进行湿地信息的提取,确定了合适的分割参数设置(空间参数为10、颜色参数为10、最小区域参数为120)。(2)选取了23个景观指数,开展湿地景观时空变化分析研究,研究表明过去36年间研究区湿地景观斑块数量增加明显,湿地破碎化程度增大。人为活动加剧是影响洪湖湿地景观斑块数量、景观破碎化程度的主要原因,导致景观破碎化程度上升,香农均匀度指数与香农多样性指数增大。其中斑块数量变化起伏较大的地类为旱地、农用池塘和稻田/冬水田。面积增长最快的属淡水养殖池塘,淡水养殖池塘增加的面积90%以上来自稻田/冬水田的转化。(3)通过景观类型构成的转入转出贡献率和保留率等方面进行的动态变化研究,揭示了不同景观组分转移特征在景观整体变化中的地位和作用,描述景观动态变化的总体特点和内在驱动机制；在景观类型动态度变化研究方面,分析了全市景观类型面积转移的活跃性,波动性变化,通过对景观类型质心变化研究,计算了各景观类型质心位置偏移方向及其偏移量,直观的再现了景观组分的时空演变过程。(4)利用CA-MARKOV和MARKOV模型对洪湖市景观类型格局进行预测,并将其预测结果与解译结果进行了对比,对比发现：CA-MARKOV和MARKOV模型的都得到了较好的估测效果,均可对洪湖市景观类型变迁进行科学、合理的预测,在估测精度方面CA-Markov模型的预测精度更准确。(5)从土壤有机碳和植被碳储量两个方面开展洪湖碳储量研究,计算结果表明洪湖土壤有机碳密度为4.557-67.519kg/m2,各土壤有机碳密度大小为：湖底淤泥>洪泛平原湿地>草本沼泽>林地>农田；总的土壤碳储量为3.537×107t,其中湖底淤泥的碳储量占了45%的比例,为1.597×107t,碳储量大小为：湖底淤泥>林地>农田>草本沼泽>洪泛平原湿地；洪湖植物有机碳含量碳密度及碳储量的计算结果表明洪湖植物有机碳密度为4.93-63.03kg/m2,碳密度大小为：林地>草本沼泽>农田>水生植物>洪泛平原湿地；总的植物碳储量为1.74×106t,其中农田的碳储量占了59%的比例,为1.02×106t,碳储量大小为：农田>林地>水生植物>草本沼泽>洪泛平原湿地。更多还原

【Abstract】 Wetland landscape is a hot research field in the past decade, and wetland carbon storage is one of the essential elements in the study of global climate change and ecological security. As the seventh largest freshwater lake in China, which owns the most abundant wetland resources in the world, the Honghu Lake wetland is typical and representative among the Yangtze River. Therefore, research into Honghu wetland landscape and carbon storage has great significance for China’s wetland ecology security, protection and rational utilization.With3S technology support, this paper collected the Landsat series, CBERS02-CCD and ALOS remote sensing data of Honghu City during the period1974-2010, a total9images for eight different time slots, had set up a wetland resource space database, combined ground auxiliary information, and extracted wetland information using object-oriented image segmentation approach; On this basis, by employing principles of landscape ecology,23landscape pattern index were selected to carry out analysis of temporal and spatial variation of Honghu wetland landscape and its succession, and to reveal the wetland type change of landscape change and transfer characteristics. Using Markov and CA-Markov model, the future landscape changes of the wetland over time were predicted and analyzed, while utilizing high-resolution ALOS images of the entire study area, carbon density and carbon storage were estimated for target landscape types. The main conclusions are as follows:(1) Referring to "International Convention on Wetlands" and the State Forestry Administration published "Technical Specification for wetland inventory", and considering the actual situation of Honghu wetlands, the interpretation standard of remote sensing classification was proposed. Through wetland information extraction using object-oriented approach, the appropriate segmentation parameters were determined as:the spatial parameter, the color parameter and the minimum area parameter are10,10and120respectively.(2) Spatial and temporal analysis of wetland variation using23selected landscape indices suggests that, over the past36years, the number of wetland landscape patches in the study area increased significantly, and the degree of fragmentation increases. Increasing human activities main factor affecting the landscape patch number and the landscape fragmentation, leads to the rising of landscape fragmentation and the increases of Shannon evenness index and the Shannon diversity index. Land types undulating patch number changes are cropland, farm ponds and paddy fields/winter paddy field. The fastest growing of area appears for freshwater ponds, and more than90%of the area increase is from rice/winter paddy field conversion.(3) Dynamic analysis of the contribution rate and retention rate of landscape composition moving-in or-out reveals the position and role of different landscape components’characteristics in the overall change of the landscape, describes the general features of landscape dynamics and the internal driving mechanism; for changes in landscape type dynamic, the landscape transfer activity and fluctuant of the city was analyzed, through investigating into the landscape center changes, the centroid position of each direction and the partial offset Shift for landscape types were calculated, visually representing the temporal evolution of landscape components.(4) CA-Markov and Markov models were used to predict the pattern of Honghu landscape, and the predicted results were compared with the interpretation results. Comparison shows that:CA-Markov and Markov models have both achieved satisfactory estimation, can provide scientific and reasonable projections for the Honghu landscape pattern change, and CA-Markov has higher prediction accuracy as to model estimation precision.(5) Carbon storage of Honghu wetland were carried out on organic carbon from soil and vegetation carbon storage, results show that the soil organic carbon density of Honghu is4.557-67.519kg/m2, the soil organic carbon density of each soil type rank as:Bottom mud> Floodplain wetland> Marshes> Woodland> Cropland; the total soil carbon reserves is3.537x107t, of which the lake bottom sludge carbon storage takes45%, accounts to1.597x107t, carbon storages rank as:Bottom mud> Woodland> Cropland> Marshes> Floodplain wetland; Honghu plant organic carbon content of carbon density and storage of calculation results show that the organic carbon density of plants is4.93-63.03kg/m2, carbon density rank of the vegetation types is: Woodland> Marshes> Farmland> Aquatic> Floodplain wetland; the total plant carbon reserves is1.74x106t, carbon storage of farmland takes59%, accounts to1.02x106t, carbon storage rank is:Farmland> Woodland> Aquatic> Marshes> Floodplain wetland.更多还原