【作者】 刘铁冬；

【导师】 罗传文；

【作者基本信息】 东北林业大学 ， 森林经理学， 2011， 博士

【摘要】 岷江江上游是四川省乃至西部地区重要的生态屏障,天然林保护工程的重点实施区域和长江上游的生态屏障,且长期以来人类不合理的资源开发和利用方式造成人口-资源-环境之间的矛盾日益突出,成为区域经济社会发展的主要限制性因子,造成生态环境脆弱化且已危及到长江流域的生态安全。基于此背景下,本文以岷江上游干旱河谷—杂谷脑河流域为研究对象,借助空间信息技术GIS和RS),以1980-2006年四期遥感数据为主要数据源,结合地形分析,从时间序列上和空间尺度上对研究区域的土地利用及其景观格局进行了深入分析,运用土地利用图谱模型对土地利用类型的数量变化、时空演变过程和空间转移规律进行了剖析,在此基础上,对研究区域开展生态环境脆弱性的评价研究,探索生境脆弱性生态区的生态恢复途径和方法,以期为研究区的生态保护和可持续发展提供理论基础和科学建议,对该区域生态环境的恢复与重建、流域可持续发展等方面具有重要的战略意义。本文得出的结论如下：‘(1)研究区域内各景观类型面积、斑块数量与优势度分布不均衡,差异明显。区域自然景观(森林景观、草地、灌木林地、岩石裸露地等)占绝对主导地位,面积总和均达到90%以上,而半自然景观、人为景观(采伐迹地、农地、居民地等)占据较小的比例,说明杂谷脑河流域以自然景观主导区域景观格局特征,人为景观则明显处于支配地位。(2)1980-2006年杂谷脑河流域的土地利用类型变化表现为林地和未利用地面积减少；灌木林地、草地、耕地、居民地和水域面积增加的态势,林地面积减少的最大,未利用地面积减少的最小；灌木林地面积增加的最大,水域面积增加的最小(3)采用5x5km的格网作为统计单元,并在ArcGIS的支持下,对土地利用综合程按自然断裂法(Natural Break)对土地利用综合度进行空间聚类,获取其空间分布格局,土地利用综合程度较高的区域主要在杂谷脑河流域沿岸的区域,归因于区域气候适宜,适宜人口生存,人类的干扰程度较大,呈现出不同土地利用兼备的格局现状。(4)林地的重心坐标变化相对较为复杂,变化趋势为高纬度----低纬度----高纬度；草地重心坐标由低纬度向高纬度转变,高海拔梯度的分布频率逐渐增加；灌木林地的重心坐标由低纬度区域向高纬度、高海拔区域转变,变化较为曲折性；未利用地的重心坐标由高纬度低海拔区域向低纬度高海拔区域转变。(5)河谷区(1980-2001年)形成森林-耕地-水域等交错的区域景观格局；2006年形成森林-灌木林地-耕地等交错的区域景观格局；亚高山区主要形成森林-灌木林地-草地等交错的亚高山针叶林带景观格局；半高山半山区主要形成草地-灌木林地-林地等交错的亚高山灌丛草甸带景观格局,该地带是森林景观向草地景观的过渡地带；高山区主要形成草地-未利用地-林地等交错的高山草甸带景观格局。(6)居民地在各个坡度分区上都有分布,主要集中分布在平坡上,水域主要集中分布在平坡区域,居民地和水域都是在无坡向区域的面积达到最大；林地和灌木林地都是在陡坡上处于优势地位,林地主要分布于半阳坡和半阴坡区域,灌木林地在阳坡上的分布比阴坡偏多,草地的分布刚好相反,且草地和灌木林地在坡向上的分布差异明显。(7)1980-2006年稳定型图谱单元在各类变化模式图谱单元中占比例较大,占研究区总面积的92.5245%,面积最大的图谱变化模式为“林地--林地--林地--林地”；前期变化型次之,变化面积6758.28hm2,占研究区总面积的3.1005%,面积比例最大的变化模式是“林地--灌木--灌木--灌木”；后期变化型面积比例最大的变化模式是“林地--林地--林地--灌木”；中间过渡型面积最大的变化模式是“林地--林地--灌木--灌木”；反复变化型面积变化类型比例最大的变化模式是“林地--灌木--草地--林地”；持续变化型是面积最小的图谱单元,其面积最大变化模式是“未利用地--草地--灌木林地--林地”,形成草地-灌木林地-林地等交错的亚高山灌丛草甸带的格局现状。(8)粒度效应的研究所采用的景观指标中：斑块数量、斑块边界密度、蔓延度、分维数和景观斑块聚集度有明显的粒度效应,它们随粒度的增加表现出有规律的变化趋势；斑块边界密度、蔓延度和景观斑块的聚集度没有出现粒度突变效应；景观多样性指数和分离度在这次研究中没有出现明显的粒度效应,对粒度变化的敏感程度较低。(9)林地景观的聚集度最大,分离度最小,其次是草地,表明林地在所有景观类型中集中分布程度最高,构成了区域的控制性景观,草地在所有景观类型中集中分布程度较高,构成了区域的主要景观类型；水域景观的聚集度指数最小,分离度较大；居民地的分维数最小。(10)聚集度指数呈逐年减少的趋势,蔓延度的指数呈逐年增加的趋势,且平均斑块面积减小,区域景观破碎化程度加剧,结果造成景观多样性、景观均匀度呈现递增,景观优势度减小,景观多样性和优势度两指数呈现出一定的负相关关系,景观多样性和均匀度两指数基本上呈现出一定的正相关关系,反映了各景观类型所占比例差异减小,而且景观类型组成成分分布较均匀,景观类型的分布由集中趋向分散交错,景观斑块分布的均匀程度较大,景观的完整性减弱,连通性降低,景观的异质性增强。(11)草地、未利用地作为研究区内面积比重较大的景观组分类型,其保留率一直较高；林地和灌木林地具有较高的转入率和转出率,转入率要大于转出率,且双向变化最剧烈；水域保留率高的原因主要是受林区深切峡谷的地形庇护；居民地等更容易受到人为的影响,其保留率得以维持或快速变化。(12)从预测的面积来看,居民地面积一直呈现增加的趋势,耕地和未利用地呈现减少的趋势,草地、林地和水域呈现先减少而后增加的趋势,灌木林地呈现先增加而后减少的趋势,表明经过18年的的恢复,人工林和次生天然林的逐步演替,其生物量将会增长,其生态屏障功能,如水源涵养、水土保持、径流调节、生物多样性保护等也将逐步恢复,区域生态环境的质量得到了好转,再次表明天然林的生态恢复是一个很缓慢的过程。(13)计算土地利用变化的终极状态概率,并将其换算为各用地类型占地面积,本文首次将终极状态概率和CA-Markov相结合,创造性地对研究区域的土地利用进行模拟和预测,其最终结果表明：耕地、灌木林地和未利用地面积呈现减少的趋势,灌木林地和未利用地减少的幅度远大于耕地,草地和林地的面积呈现增加的趋势,林地增加的幅度较大,水域的面积基本上没有发生太大的变化。(14)强度侵蚀的主导植被类型是耕地和草地区域,中度侵蚀区域集中分布在灌木林地地带,轻度侵蚀和微度侵蚀的主导植被类型是落叶阔叶林和针阔混交林,其中微度侵蚀主要集中分布于灌木林地,无侵蚀和微度侵蚀的主要植被类型是针叶林,总体而言,从控制土壤侵蚀强度的能力大小看：针叶林>落叶阔叶林>针阔混交林>灌丛>草地>农用地。(15)杂谷脑河流域生态脆弱性评价,可以划分为5个等级,依次分别为：EVI低于25为微度脆弱,其所占的面积比例为2.7040%；EVI在26-40之间为低度脆弱,其面积比例为36.0134%；EVI在41-55之间为中度脆弱,面积比例为29.8752%；EVI在56-70之间为高度脆弱,面积比例为19.5512%；EVI高于70为极度脆弱,面积比例为11.8562%,其中微度和低度脆弱区所占比例之和仅为38.7174,说明区域生态系统总体运行状态一般,且有31.4074%左右的面积属于重度和极度脆弱区,表明生态环境问题依然严峻,气候干燥、土壤贫瘠、干旱缺水和地质灾害贫乏是其生态脆弱性形成的自然基础,人类活动加剧尤其是不合理放牧和大量垦荒是生态脆弱性进一步加剧的重要人为根源。(16)现有的生态脆弱性评价,几乎都是单向性的,没有对评价结果进行独立的验证,在查阅文献的基础上,本文创造性地提出了生态脆弱性评价结果的验证方法,通过网格的中心点提取平均盖度和环境脆弱度评价值,对两者进行回归分析,研究两者之间的关系,该结果表明植被指数决定了大于3/4左右的生态脆弱性区域,从某种意义上讲区域植被覆盖密度很大程度上可以决定着一个地区的生态脆弱性程度,通过研究植被盖度与生态脆弱度之间的这种关系,为我们提供了根据植被指数来验证生态脆弱度评价结果的可能性,表明评价结果能够较好地反映研究区域生态脆弱性实际情况。(17)对不同海拔、坡向生境脆弱度的三维趋势面模型的分析可知,生境脆弱度较低的区域集中在半山半高山区域,在海拔3400-3800米之间是其主要的分布区；其次是河谷区,在该区域随着海拔的升高,脆弱度呈现减少的趋势；在亚高山区和高山区,随着海拔的升高,脆弱度呈现增加的趋势。各海拔带生境脆弱度随坡向的变化不是特别明显,总体上是呈现缓慢增加的趋势。(18)生境脆弱度在坡度方向呈现“U,,变化曲线,而在坡向方面呈现缓慢递增的趋势,生境脆弱度较低的区域主要集中分布于坡度在21-420和坡向在60-1600之间的交叉区域,坡度在42-900区间,随着坡度的增加,生境脆弱度呈现增加的趋势；坡度600-900范围内,随着坡向的增大,evi呈现“U,,变化曲线,生境脆弱度在阳坡和半阳坡(180-2100)较小,而在阴坡和半阴坡(28-800和320-3600)较大。(19)生境脆弱度在高程方面呈现“U,,变化曲线,而在坡度方面呈现弧度较大的下降圆弧曲线,生境脆弱度较低的区域主要集中分布于坡度在21-420的半山半高山区。更多还原

【Abstract】 Minjiang upstream in Sichuan province and western regions is an important ecological barrier. Also it is the significant experimental subject of Natural Forest Protection Programme and the vital ecological barrier of the headwater of Yangtse River. For a long time, that the unreasonable resources development and utilization ways of human cause the contradiction among population, resources and environment increasingly. The contradiction limits the regional economic social development and gives rise to the week ecological environment which has endanger the ecological security of the Yangtse River basin. Based on this background, in this paper I choose arid valley of Minjiang River-Zagunao River as the subject, use space information technology(RS, GIS and RS), treat remote sensing data from 1980 to 2006 as the main data source and combine with terrain analysis to deeply analyse the subject of land use and landscape pattern in time series and space scale. Moreover, I dissect the quantity change of land use type, spatio:temporal evolution and space transfer law by using land use mapping model. Then by assessment of ecological environmental vulnerability in study area and exploration of ecological restoration approach in habitat vulnerability biota, in ecology protection and sustainable development of subject I expect to give the rationale and scientific advice which will effect an optimistic strategic influence in restoration and reconstruction of the ecological environment and sustainable development of river basin. The conclusions will be given in the following passages.(1) In the study area, proportion of landscape type, quantity of plaque and dominance distribute lopsidedly. Regional natural landscape (such as forest, grassland, shrub land and rocks exposed areas) account for the absolute dominant position, taking up over 90%. However, half natural landscape and anthropogenic landscape (such as cut over land, agricultural land and settlement place) occupy smaller proportion. It illustrates that in Zagunao River the dominated regional landscape pattern is natural scenery, and man-made landscape is obviously in ascendancy.(2) From 1980 to 2006, the transformation of land use pattern in Zagunao River area displayed in two ways. Firstly, the proportion of forest and unused land had reduced. The area of forest declined maximally, and that of unused land decreased minimally. Secondly, the acreage of shrub land, grass land, agricultural land and water area remained escalation. The increasing proportion of shrub land made up the biggest part of all,in the meanwhile, that of water area took up the smallest one.(3) Under the support of ArcGIS, I use 5* 5km grid as the statistical unit to get the spatial clustering of synthetic degree of land use by method of Natural Break. It could give us the space distribution pattern. From that, we can see the areas of high land use comprehensive level mainly locate in the place along Zagunao River. The main reasons why coast wise of Zagunao River display the high level of land use is the climate which is suit for living and the big degree of human interference. Also that the area which expresses the pattern status of reasonable use of various lands is very important.(4) The change of woodland barycenter coordinate is relatively complex. And it often changes from circumpolar latitude to low latitude, then back to circumpolar latitude. Grassland barycenter coordinate transforms low latitude into circumpolar latitude. Distribution frequency of high altitude gradient increases gradually. Shrubland barycenter coordinate turns low latitude into circumpolar latitude and high elevation curvilinearly. Unused land barycenter coordinate usually converts the areas of circumpolar latitude and low elevation into the places of low latitude and high elevation.(5) Valley area (from 1980 to 2006) had formed the regional landscape pattern which is forest-cultivated land-waters interlaced. In 2006, it had shaped the regional landscape pattern of stagger forest, shrub land and agricultural land. Half-high-mountain ares mainly formed sub-alpine acerola forest landscape pattern mixing with forest, shrub land and agricultural land. However, the main pattern in half-high-mountain and semi-mountain area was grassland-shrub land-forest interlaced sub-alpine bush meadow belt landscape pattern. This place was transitional zone from forest landscape to meadow landscape. High-mountain area primarily formed the high-mountain meadow landscape pattern with grassland, unused land and forest mixed.(6) Settlement place scatters in every slope partition, but mainly distributed in the flat slopes. The distribution of waters mainly concentrated in flat slope area. And the acreages of settlement place and waters arrived the peak in the area without slope. Both the forest and the shrub land stayed in the positive status in abrupt slope. Most forest was always in the area of partial adret slope and partial shady slope. And the distribution of shrub land was usually in the adret slope, not the shady slope. However, that of the grassland was in the opposite side. Also at the slope, the difference of distribution in grassland and shrub land is obvious.(7) The period from 1980 to 2006 witnessed a fact that the stable mapping unit occupied large proportion of units of every variation pattern map, making up 92.5245% in whole study area. And the variation pattern map, the area of which remained the top of the table, was’forest-forest-forest-forest’. The changing area of prophase type was 6758.28hm", ranked second with taking up 3.1005% of all. In this type, the variation pattern map which had the largest acreage was’forest-shrub land-shrub land-shrub land’. However, that in later period was’forest-forest-forest-shrub land’. The model with the biggest change of area in transitional type was’forest-forest-shrub land-shrub land’. In reiterative change model the variation pattern, the acreage of which is the biggest one, was’forest-shrub land-grassland-forest’. The sustainable change model was the mapping unit with the smallest area in whole. And in that type, the model with the largest change of acreage was’ unused place-grassland-shrub land-forest’. Then the pattern status of sub-alpine shrub-meadow area with grassland-shrub land-forest interlaced had formed.(8) The study of grain-size effect chose several landscape indexes.And in these indexes, the grain-size effect was obviously found in plaques number, plaques boundary density, spread degrees,fractal dimension and landscape plaques gathered degrees which expressed the regular trend of variation.However, there was no granularity mutation effect in plaques boundary density, plaques shape index and landscape plaques gathered degrees. Landscape diversity index did not result in any obvious granularity mutation effect. So it can be said that landscape diversity index had low sensitivity of that.(9) Woodland landscape had the high gathered degree and the low separation degree, and the second was grassland. It expressed that in whole type of landscape, forest which constituted the dominant scenery had the highest degree of concentration distribution. And grassland had the higher degree of concentration distribution, constituting the main scenery. The gathered degree of waters landscape was the smallest of all, but the separation degree of that was the biggest. The settlement place had the smallest HFD in whole.(10) There were some facts. The gathered degree index turned to decline year by year. Spread degree index turned to increase, and average plaques area decreased. Regional landscape fragmentation degree intensified. And these result in the accumulation of landscape diversity and evenness and the decline of landscape dominance index. There was a negative correlation between landscape diversity and landscape dominance index. Also there was a positive correlation between landscape diversity and landscape evenness. It reflected that different landscape types made up the similar proportion. And the compositions of landscape types were uniformly distributed. The distribution of landscape types turned concentrated into scattered and crisscross. Most landscape patched distributed evenly. There was a decrease in landscape integrity and connectivity. Nevertheless, landscape heterogeneity increased.(11) The retention rate of grassland and unused places which were treated as the landscape types with large proportion of area in subject always stayed in the high level. Forest and shrub land had the higher switching-in rate and switching-out rate which varied most drastically. And the switching-in rate was more bigger the the switching-out rate. The result of high retention rate of waters was protected by the deep canyon in forest. The settlement place was effected by human beings easily. So its retention rate can remain stabilization or change sharply.(12) From the expected acreage, the area of settlement place always remain increase. The acreage of plowland and unused places turned to decline. That of the grassland, forest and waters decreased at first and then rose. However the area of shrub land stood in the opposite side. That expressed the gradual succession of plantation and secondary forest after the restoration of 18 years:the biomass live weight would increase, the ecological barrier function (such as water conservation, soil and water conservation, runoff regulation and biodiversity conservation) would recover stepwise, the quality of ecotope began to improve. Also that expressed ecological restoration of natural forest should experience a slow process.(13) To calculate the posibility of the ultimate state of the land use which is changed, and to converse the posibility to the area of the various land types, this paper will combine the posibility of the ultimate state with the CA-Markov firstly, and creatively simulate and predict the land use of studied area, its final result shows that:the area of plow land, shrub land and used places turned to decrease. That of shrub land and used places dropt sharply than that of plow land. The area of grass land and forest had the trend of increase. And the rise of the area of forest was more rapid. The area of waters remain stabilization nearly.(14) The dominant vegetation types in deep erosion region ware plow land and grassland. However, that in moderate erosion area mainly distributed in shrub land. The dominant vegetation type in mild erosion and micro degree erosion area was the broad leaved deciduous forest and the leaved and coniferous mixed forest. And the micro degree erosion area mainly distributed in the shrub land. The dominant vegetation type in both no erosion region and micro degree erosion region was coniferous forest. All in all, from the level of ability to control soil erosion intensity:coniferous forest>broad leaved deciduous forest>leaved and coniferous mixed forest>shrub land>grassland>agricultural land.(15) The assessment of ecological vulnerability of the Zagunao river can be divided into five levels, as follows:EVI belows 25 as the slight vulnerability, the proportion of the area is 2.7040%;EVI between 26-40 as the light vulnerability, the proportion of the area is 36.0134%;EVI between 41-55 as the medium vulnerability, the proportion of the area is 29.8752%;EVI between 56-70 as the high vulnerability, the proportion of the area is 19.5512%;EVI higher than 70 as the extreme vulnerability, the proportion of the area is 11.8562%. Thesum of the proportion of the area of slightly vulnerable and light vulnerable is only 38.7174. It indicates that the overall operation state of the ecological system is general, and there are the high and extreme areas about 31.4074%. It shows that the environmental problems are still severe. The dry climate, the Poor soil, the drought and the water shortage and the fregeuntly geological disaster are the natrual basis of the formation of the ecological vulnerability. Human origin of the severe ecological vulnerability. More and more human activities,especially unreasonable herd and abundant assart is the essentially.(16) There is no independent validation for the existing assessment of ecological vulnerability which is almost one-way. Based on the literature review, this paper creatively puts forward the method to verify the results of the ecological vulnerability assessment. Then analyze the relationship of the rate of the average coverage and the value of the environmental vulnerability evaluation which are extracted by the center point of the grid. The result of it shows that the vegetation index decides almost 3/4 of the ecologically vulnerable zone. In a sense, the density of ragional vegetation can largly decise the extent of the ecological vulnerability in a region. The research of the relation between the rate of vegetation covering and the level of ecological vulnerability enables us verify the result of the assessment of ecological vulnerability according to vegetation index. Moreover, the research dedicates that assessment result can well reflect the actual situation of the regionally ecological vulnerability which is studied.(17) From the 3D trend surface model analysis of habitat fragile degree in different elevation and aspect, the area with low habitat fragile degree mainly distributed in the half-high-mountain and semi-mountain area, especially in the area with the elevation between 3400m to 3800m. The secondary is valley area. In this area, habitat fragile degree had the negative correlation with elevation. In sub-alpine area and alpine area, with the increase of elevation, the habitat fragile degree turned to become more higher. With the change of aspect, the habitat fragile degree in different elevation did not have any obvious transformation, whereas it expressed a trend of low increase in whole.(18) In gradient, habitat fragile degree expressed’∪’curve. Nevertheless, it expressed the trend of slow increase in aspect. The area with low habitat fragile degree mainly distribute in the cross area between 21-420 gradient and 60-1600 aspect. If gradient between 42-900, habitat fragile degree was proportional to the aspect. If gradient between 600-900, when aspect rose, EVI expressed’∪’curve. In the sunny slope and the semi-sunny slope, habitat fragile degree was more lower. However, that in the other site conditions was more higher.(19) In elevation, habitat fragile degree expressed’∪’curve. However, in gradient it expressed decline arc curve with large radian. The area with poor habitat fragile degree mainly distributed in the half-high-mountain and semi-mountain area with gradient between21-420.更多还原