【作者】 罗龙海；

【导师】 胡庭兴；

【作者基本信息】 四川农业大学 ， 森林经理学， 2007， 硕士

【摘要】 本研究在对沐川县土地利用／覆盖变化进行分析过程中，引入了景观格局分析方法，将遥感与地理信息系统技术有机地结合起来，综合运用计算机自动分类技术与专家知识对该区域的LUCC、景观格局变化及其迁移特征进行了定量分析。为弄清地方林业产业发展对区域生态的影响程度及范围，进而充分识别该区域生态环境的变化趋势及其内在因素，制定合理的生态管理与土地利用决策提供了参考和借鉴。现将本文主要结论总结如下：1、以1990、2001、2006年二个时期的LANDSAT-5 TM遥感影像和1999年的森林资源分布图作为数据源，采用监督分类结合专家知识分类方法获得了沐川县三个时期的土地利用专题图，并进行了精度评估，分类精度分别达到了81.42％、86.52％、86.10％。2、采用谱间关系法(TM2+TM3＞TM4+TM5)和阈值法(TM4＜60)提取了实验区不同时相TM影像的水体界限，减少了地形阴影被误分为水体的机率。对建设用地(居民地、公路)采用了基于光谱知识的居民地、公路提取模型来提取，以修正监督分类的结果。然后将重新提取的水体和建设用地导入ArcGIS，更新监督分类的结果，并将其转换为了.img格式，从而找出了一条利用ERDAS与ArcGIS的优势功能与人工目视解译相结合的方法来进行遥感图像的自动分类的途径，提高了分类精度。3、得到了沐川县1990年、1999年、2001年、2006年各时期的森林覆盖率，分别为47.55％，50.55％，53.84％，57.13％。4、将ArcGIS下更新后的矢量数据转换为GRID格式的栅格数据，并进行了面积统计、地形分异研究和相关景观指数的计算。最终获得了土地利用转移矩阵，并结合景观地形分异研究和景观格局分析找出了影响沐川县土地利用景观变化的主要因素。为认识沐川县区域景观行为，制定相应的生态管理与土地利用决策提供了依据。5、通过土地利用转移矩阵的计算，获得了以下有研究意义的研究结果：1999～2001年为耕地向林地转化的一个最重要时期，期间大量耕地向林地进行了转化。而该时期为退耕还林初期，很明显，是国家的退耕还林政策引导了这一大量耕地向林地(竹林)的转化过程。林地的转出面积比例中，大部分都是相互之间的转入和转出。水域和建设用地面积变化不大，水域面积变化主要受马边河流域新建水电站的影响。建设用地面积变化主要受政府改善城市投资环境政策影响，且主要由耕地和林地转化而来。6、对研究区土地利用类型按面积统计和动态度分析可知：研究区耕地面积随时间的变化成逐年下降趋势，到2006年仅占全县总面积的42.15％；林地中，变化面积最大的地类是竹林，且变化时期主要集中在1999～2001年。针叶林、阔叶林面积1990年～2001年11年期间比例基本保持不变，2001年开始才逐渐有一定幅度的增加。水域、建设用地面积比例变化较小。灌木林1990～1999年的动态度最大，达到14.12％；竹林1999～2001年的动态度最大，达到44.89％；针叶林2001～2006年的动态度最大，达到12.31％。将不同时期各地类的动态度做纵向比较可知，1999～2001年为竹林产业发展最重要时划。7、通过对景观指数的计算，获得了如下有意义的研究结果：1999～2001年，大量的竹林的栽植，对景观的连通性产生了消极作用，景观破碎度达到最大值，加强了生物行为或生物过程相关的渗透性。同时也说明了这一时期内景观的连通性受到了外界干扰的巨大影响。2001年后，大量竹林成林，并向周围迅速扩大面积，促进了景观连通性的好转，景观破碎度有所下降。1999～2001年间，在西南部高丘河谷平坝区，斑块的多样性增大，破碎度增加。而西南部高丘河谷平坝区主要是耕地，是沐川县的粮食主产区。这同时也说明，是耕地的破碎化导致了景观多样性的增大。在海拔较高的区域，由于耕地数量较少，退耕还林对耕地的影响变小，此时景观的连通性变好，景观中斑块被分割的可能性变小，破碎度变小。同时也说明人为因素对景观破碎度的影响主要集中在低海拔地区。8、对研究区土地利用区域差异分析可知：在西南部高丘河谷平坝区，1990～1999年耕地的逐年减少主要是受竹产业发展的影响，一部分耕地逐步转变为了竹林地。1999年退耕还林后，由于耕地紧张，农民担心粮食收成及经济收入不佳，将部分距离居民区较近，地力较高、郁闭度较低的针叶或阔叶林下栽种了经济作物，并加速了耕作，导致了耕地面积有所增加的现象。另外，1999～2001年两年内，退耕后竹林由于郁闭度很小，在影像上多数被判读为了耕地，这也是导致耕地面积增加的又一重要原因。在西北部低山区(500～1000m)，耕地面积在1999～2001年期间年均降幅为4.30％；相反，竹林与针叶林在1999年后大幅增加，尤其是竹林面积，在1999～2001年期间年均增幅达52.02％，面积大小则占到低山区总面积大小的33.42％，成为这一区域内的主导林种。东南部低山区(1000m以上)与西北部低山区一样，耕地面积变化也表现出相同的趋势，但1999～2001年期间的变化更为显著。竹林在这一区域面积增幅不大，主要原因是退耕还林结束后，研究区的竹林已经开始从造林阶段转向生产利用阶段。9、通过地形分析得知：研究区的耕地主要是坡耕地，且主要分布在缓坡和斜坡上。1999～2006年，陡坡、急坡范围内耕地面积由17012 hm~2下降到6799 hm~2。这说明，在国家对25°以上坡耕地退耕还林政策的影响下，这一坡度范围内的耕地己经有60％退出了耕地用途；针、阔叶林主要分布在斜坡和陡坡上。2001～2006年，针叶林面积在陡坡上的增大主要是由于该坡度范围的针叶林受人为影响很小，一部分退耕的耕地被栽种了针叶林。阔叶林面积在急、险坡上随着坡度的增大而下降，在斜坡以上坡度范围内，阔叶林面积随时间的变化而逐年下降，主要原因是在大力发展竹产业这一地方经济产业发展的推动下，一部分阔叶林在与竹林的生态位竞争中处于弱势，演变成了竹类占优势的混交林；灌木林主要分布在斜坡、陡坡和急坡上，其面积在各个坡度等级上所表现出的分布规律基本都是：1999年＞2001年＞1990年＞2006年。竹林主要分布在斜坡和陡坡上，沐川县的竹林1999年后在坡度≤25°范围得到了较大发展，尤其在斜坡上，平均增幅达到154％。水域和建设用地都主要分布在平、缓坡上，受坡度因素影响较小。耕地和竹林面积大小随坡向的总体分布规律相似，都是阴坡＞半阴坡＞半阳坡＞阳坡＞平地。竹林面积增加主要是依靠研究区竹林产业产业的发展。灌木林和阔叶林面积的坡向总体分布规律相似，都是：半阴坡＞阴坡＞半阳坡＞阳坡＞平地：阔叶林面积在平地上一直维持在比较稳定的水平。针叶林面积大小的坡向总体分布规律是：半阴坡＞半阳坡＞阴坡＞阳坡＞平地。研究表明，沐川县的林业产业发展是其景观变化的直接驱动因素，退耕还林和国家的天然林保护工程是引导这一林业经济产业发展的直接驱动力。生物过程在景观动态变化中居于次要地位。更多还原

【Abstract】 The analysis methods of landscape structure were introduced when we analyze the LUCC (Land Use and Cover Change) of Muchuan County, and combined with the RS and GIS technique, including the computer automatic classification and the expert knowledge, the LUCC, landscape structure and their transfer character were analyzed quantitatively. The results provided references for understanding the influence extent and rang of the development of regional forestry industry acted on the eco-system. Moreover, it also helped us to identify the trend of change of the eco-system and its internal factors, make suitable eco-system management plans and land use decisions. All conclusions can be summarized as:1. The remote sensing image of Landsat-5 TM in 1990, 2001, 2006 and the forest distribution map in 1999 were used as information source, and the method of supervised classification combined with the expert knowledge were used for the classification. At last, three classification maps of remote sensing images have been got. Their accuracy arrived at 81.42%, 86.52% and 86.10% respectively.2. Methods of spectrum relations (TM2+ TM3＞TM4+ TM5) and critical values (TM4＜60) were used to distill the water boundary, which could obviously reduce the error that water been classified to shadow. As to the building land, the model of "Distilling on the basis of the building and road" were used to correct the results of the supervised classification. Then, the water boundary and the building land were imported to ArcGIS to update the results of supervised classification. These data were converted to format of *.img which meant that the predominance of ArcGIS and ERDAS were combined with the artificial interpretation, a new classification method for automatic classification of remote sensing image was found to enhance the classification accuracy.3. Four period’s forest coverage rate in Muchuan county (years of 1990, 1999, 2001 and 2006) were got, arrived at 47.55%, 50.55%, 53.84% and 57.13% respectively.4. The updated vector data were converted to format of *. GRID. Four period’s data were analyzed by their area, terrain distribution and landscape structure index. At last, the transfer matrix of land use was got and the main factor, which influenced the land use and landscape change, was found. These results provided rules for recognizing regional landscape structures, planning corresponding eco-system management and land use decisions.5. By calculating the transform matrix of land use, some significant research results were got as below.Years between 1999 and 2001 were the most important period that the farmland converted to forestry land. In this course, large numbers of bamboos were planted, and obviously was the policy of "farm land conversion to forestry land "led the transform course. Among the forestry land transform rate, most were their own transform course. Water and building land have little change. Change of water area influenced mainly by the newly built water-electric station. Change of building land area influenced mainly by the policy of the government’s improving city’s investment environment, and mainly transformed from farmland and forestry land.6. Statistics of land use type in study region by area and dynamic analysis showed that:Area of farmland decreased gradually with the change of time. By 2006, the area of farmland reached only 42.15% of the total study region. Bamboo land had the great change in the forestry land among years of 1999 and 2001. Area of conifer land and broadleaf land kept steadiness on the whole in 11 years from 1990 to 2001, and after year of 2001, it began to increase. Area of water land and structure land changed small.Dynamic rate of shrubbery land had its biggest values of 14.12% among years of 1990 and 1999; Bamboos land had its biggest values of 44.89% among years of 1999 and 2001; Conifer land had its biggest values of 12.31% among years of 2001 and 2006; Comparing of different land type in different period showed that years among 1999 and 2001 was the most important development period of bamboos economy.7. By calculating the landscape structure index, some significant research results were got as below:Years between 1999 and 2001, large numbers of newly planted bamboos led negative effect to landscape connectivity, and the fragmentized rate of landscape got its maximum which strengthened the infiltration correlated with biological behavior or biological course. It also interpreted that the landscape connectivity been greatly disturbed in this course. After years 2001, large numbers of bamboos converted forest, and diffuse toward every direction, improved the landscape connectivity, the fragmentized had a certain extent decrease.Years between 1999 and 2001, in those area altitudes below 500 meters, the diversity rate and fragmentized rate of patches increased. Considering were mainly farmland in those area, which also meant that the fragmentation of farmland led to the increase of diversity of landscape.In those areas with high altitude, because of their low amount of farm and, the engineering of farmland conversion to forestry land had less influence to farmland. The landscape connectivity got better, the possibility of landscape been divided got small, and the fragmentation of landscape got small. At the same time, it also meant that the influence of human’s activities to fragmentation of landscape mainly in those areas with low altitude. 8. By analyzing the district difference of land use, some results were also got below:Years between 1999 and 2001, in those district altitudes below 500 meters, the decrease of field land mainly influenced by development of bamboos industry and some field lands converted to bamboos land. After the engineering of "farmland conversion to forestry land" began to put in practice in year 1999, farmers were afraid of lacking foodstuff and economy income and planted some economic crops under the fertile and thin conifer forest and broadleaf forest which near their residential area. Intense cultivation led the increase of field land. Moreover, years between 1999 and 2001, bamboos land come from farmland were distinguished to farmland for their low coverage rate, and became another important reason of increase of farmland.In those district altitudes between 500 and 1000 meters, the area of field land decreased by 4.30% every year from year 1999 to 2001. On the contrary, area of conifer land and bamboos land increased greatly after year 1999, especially area of bamboos land, increased by 52.02% every year from year 1999 to 2001 and account for 33.42% of low mountain region’s total area.District altitudes above 1000 meters had similar change rules in field land’s change with district of altitudes between 500 and 1000 meters, but had more significant change from year 1999 to 2001. Bamboos land changed little in this district mainly for the bamboos had turned to utilization stage from forestation stage.9. By analyzing the terrain, some results were also got below:Farmland in study region was mainly sloping field and distributed on slow slope and slant slope. From year 1999 to 2006, area of farmland in this slope range decreased from 10172 hectares to 6799 hectares. This meant sloping field above 25°had 60% out of its primary use after the engineering of "farmland conversion to forestry land" began to put in practice. Conifer land and broadleaf land mainly distributed on slant slope and abrupt land. Increase of conifer land area among year 2001 and 2006 mainly for small influence of human’s activity or plantation of conifer trees. On rapid slope and perilous slope, area of broadleaf land decreased with the increase of slope and above slant slope range, area of broadleaf land decreased with the change of time. The reason can be drawn as: under the impulse of forestry economy industry’s development, some broadleaf land got inferior position when competed with bamboos land and converted to mixed forest which bamboos got predominance. Shrubbery mainly distributed on slant slope, abrupt slope and rapid slope, and their areas in every slope classes followed the rules: Area of 1999＞Area of 2001＞Area of 1990＞Area of 2006. Bamboos land mainly distributed on slant slope and abrupt slope. Area of bamboos got great development after year 1999 under slope of 25°, especially on slant slope, and the increase extent reached 154%. Water land and structure land were mainly distributed on even slope and slow slope, and had little influence of slope.Area of farmland and bamboos land had the similar slope aspect change rules: Shady Slopes＞half Shady Slopes＞half sunny slopes＞sunny slopes＞flat. Increases of bamboos land area mainly depend on the development of study region’s bamboos economy. Shrubbery land and broadleaf land had the same slope aspect change rules: half Shady Slopes＞Shady Slopes＞half sunny slopes＞sunny slopes＞flat. Area of broadleaf land maintained a steady level on the flat land. Conifer land’s slope aspect rules was: half Shady Slopes＞half sunny slopes s＞Shady Slope＞sunny slopes＞flat.In a word, years between 1999 and 2001 were Muchuan County’s most important period of LUCC and landscape structure change. Development of forestry industry was the direct factor, and the engineering of farmland conversion to forestry land combined with the engineering of natural forest protection were the direct drive force. Biology course was the secondary factor of the landscape dynamic change.更多还原