【作者】 余坤勇；

【导师】 刘健；

【作者基本信息】 福建农林大学 ， 森林经理学， 2012， 博士

【摘要】 林地生产力是营林的重要基础，实现林地生产力及其演变基础数据的获取，对于指导森林资源经营管理、促进适地适树、挖掘林地资源潜力、提升林地生产力，促进林业可持续发展具有重要意义。本文于2008年8月14日至8月29日在永安市布设118个样地点，采集土壤样本，实验测定各样点土壤有机质、土壤含水量等理化性质，结合国家自然科学基金《林地立地质量遥感反演技术研究》（40971043）项目和福建省科技计划重点项目《基于3S技术工业原料林林地优化经营技术应用研究》（2008N0003）成果提供的1999年和2008年永安市的林地土壤有机质专题图和土壤侵蚀图，实现林地生产力的遥感估测和林地脆弱性的测定，采用熵突变理论分析不同时期林地生产力演变与林地脆弱性的关系，构建基于单期遥感数据的林地生产力演变遥感监测模型，实现1999年、2008年永安市林地生产力演变的时空效应分析，基于永安市林地生产力的演变、脆弱性的影响等因素，综合开展永安市林地资源开发利用的适宜性测定研究，为林地资源的优化经营、林业可持续发展提供辅助依据。主要研究结果如下：(1)利用1999年和2008年的遥感数据，采用面向对象的专题信息提取方法，实现1999年、2008年研究区马尾松林、竹林、阔叶林、桉树林、经济林、杉木林、灌木林、疏林地、其他林地共9种森林类型专题信息的提取，1999年的分类精度为90.63%，Kappa系数为0.8934；2008年为86.54%，Kappa系数为0.8419。(2)选择林地生产力的估测影响因子，以传统林地立地质量等级为参照，通过数量化理论I方法，构建了永安市林地生产力遥感估测模型，估测精度为81.15%：林地生产力=3.432370.14864×坡度1+0.63546×坡度2+0.13683×坡度3+0.50472×坡向1+0.81227×坡向2+0.2211×坡向30.20431×海拔1-0.63461×海拔2+1.32375×土层厚度1+0.31163×土层厚度20.40069×侵蚀10.54288×侵蚀20.6219×侵蚀3+0.00087×有机质0.00028×NPP0.09522×腐殖质(3)选定降雨侵蚀R因子、坡度、地形起伏度、土壤侵蚀K和人为干扰指数5个因子作为林地脆弱性评定指标因子，通过咨询相关专家和模糊层次分析法确定林地脆弱性各指标权重，采用综合指标数法实现林地脆弱性的测定。(4)将信息熵理论引入林地利用结构演变分析，永安市1999年和2008年林地利用结构信息熵分别为0.7404、0.7858；1999年、2008年林地生产力等级结构“熵”为0.5468、0.5624。通过林地脆弱性与采用2008年和1999年的林地生产力比值构建的林地质量Site08/99指数相关性分析，结果表明，10a来永安市林业经营过程中注重脆弱性高区域林地资源的保护，脆弱性高的地方林地生产力向提升方向发展。(5)构建了1999年和2008年的林地生产力“熵”，以林地脆弱性等级为步长，构建林地生产力熵的尖点突变模型，首次引入了熵突变理论分析永安市不同时期林地生产力演变与脆弱性的关系。结果表明，1999年时永安市林地生产力未发生突变，而2008年永安市林地生产力在不同脆弱性空间上发生了突变。2008年林地生产力熵均值为0.7755整体比1999年林地生产力熵显下降趋势，随着脆弱性的提升而下降。(6)首次选定林地土壤有机质（yjz）、植被净第一性生产力（NPP）和土壤侵蚀量(qs)3个因子构建了基于单期遥感数据林地生产力的遥感监测模型，模型估测结果与构建的林地质量Site08/99指数相关性达到了0.9，二者在表达林地生产力演变表现出一致的规律性。(7)依据生产力水平是否高于平均水平，将林地生产力划分为林地生产力为“一般”区和“良好”区。分析林地生产力演变遥感监测模型估测的林地生产力演变指数值与干扰指数、利用1996年、2006年永安市的二类调查固定样地资料和理查德方程构建的杉木、马尾松和二者综合的地位指数比值之间的关系，综合确定林地生产力演变遥感监测模型估测的林地生产力演变值0.9323，作为林地生产力“一般”区和“良好”区的阈值条件。并利用1996年、2008年永安市林业小班数据库中的林木生长量是否高于林木平均生长量作为林地生产力“一般”和“良好”区的判定依据，检验林地生产力演变遥感监测模型识别林地生产力“一般”和“良好”区的有效精度，精度为81.95%。(8)叠加、确定1999年-2008年永安市林地生产力“一般未变”、“良好未变”、“一般转良好”和“良好转一般”4个演变区。结合脆弱性、立地质量等级、生态公益林和商品林区、坡度条件以及林地利用方式等因素，分析1999年-2008年永安市林地生产力演变的时空效应。结果表明，桉树、经济林的林地生产力下降程度较大，而杉木下降程度也相对较大，马尾松和竹林则相对较好，生态敏感性较强区域的林地生产力得到提升。(9)利用野外获取、测定的水分、土壤密度和总孔隙度样本数据，采用反距离加权的空间确定性插值法插值得到永安市2008年林地水分、土壤密度和总孔隙度空间分布专题图，插值估测精度达95%以上。(10)选择林地生产力演变指数、林地现况生产力、林地脆弱性、林地土壤侵蚀、坡度、坡向、海拔、水分、土壤密度和总孔隙度共10个因子，结合层次分析法构建永安市林地资源开发利用适宜性评价指标体系，采用0-10等级赋分法和综合指数法实现林地适宜性的综合测定。通过模糊聚类分析和加权处理，将永安市林地资源按适宜性得分值的“7.78以上、7.78-6.02、6.02以下”3个区间划分为适宜性好、适宜性一般和适宜性差三个等级。结果表明：永安市林地资源中适宜性好、适宜性一般面积合计达到了212693.76hm2，比重为85.25%。商品林中有87.43%的面积为适宜开发利用的林地资源，10a来永安市的林地生产力发生了一定的变化，林地适宜性开发的面积较大，整体上体现了永安市林业发展水平较高，林地资源利用的集约化技术水平整体较为理想。(11)结合林地现况生产力等级、坡度等级、脆弱性等级和1999年至2008年林地生产力的演变情况等要素，分析各要素相应等级上适宜性林地资源的分布情况。结果表明，林地适宜性在坡度上的变化与生产力等级存在着显著的关系，商品林中生产力好的区域，林地适宜性开发利用的程度较高，随着脆弱性和坡度等级的提高、林地生产力水平的下降或生产力低于平均水平时，林地资源的开发利用适宜性均显降低趋势，不适宜开发利用面积显增大趋势，商品林地资源中适宜性差的林地面积主体分布在坡度3、4等级和生产力水平下降区，综合表明了本文划分的林地适宜性等级具有良好的合理性与适用性。更多还原

【Abstract】 The forestland productivity is the important foundation of the planting forest, acquisitionthe evolution basic data of the forestland productivity is of great significance for guiding forestresource management, promoting the matching tree species with site, mining the potential offorest resources， lifting the forestland productivity and promoting the forest sustainabledevelopment. The soil physical and chemical properties such as soil organic matter and soilmoisture were measured with118sampling points in Yong’an city were layouted on August14,2008to September29,2008.The forestland productivity and the forestland fragile were estimatedby using the remote sensing data on1999and2008,combining forestland soil organic matterthematic map and soil erosion map in Yong’an city on1999and2008which provided byNational Natural Science Foundation《On the Inversion Technique of the Forestland Site QualityBased on Remote Sensing》（40971043）and the key project of Fujian province science andtechnology program—《On the Industrial Raw Material Forestland Optimization ManagementTechnology Based on Geomatics Technology》(2008N0003). By using entropy catastrophetheory to analysis the development relation between the forestland productivity and theforestland fragility on different periods. The forestland productivity evolution remote sensingmonitoring model based on single-period remote sensing data was constructed and was used toanalysis on the time-space effect of the forestland productivity evolution in Yong’an city on1999and2008. The forestland suitability were determinated on the basis of the effect factors such asthe evolution and fragility of forestland productivity in Yong’an city by developing in the study,which would provide aided basis for the optimal management of forestland resource and forestrysustainable development in Yong’an city. The main results are the following：(1) The9forest types thematic information which include masson pine, bamboo forest,broadleaf forest, eucalyptus, economic forest, Chinese Fir, shrub, open forestland and otherwoodland were extracted by using object-oriented methods and the remote sensing on1999and2008, the classification accuracy is90.63%in1999and86.54%in2008,the Kappa coefficient is0.8934in1999and0.8419in2008.(2) The forestland productivity remote sensing estimation model in Yong’an city wasconstructed on the basis of the forestland productivity estimation influencing factors and theQuantification Theory I method, according the traditional grade of forestland site quality and the estimation accuracy is81.15%:forestlandproductivity=3.432370.14864×slope1+0.63546×slope2+0.13683×slope3+0.50472×aspect1+0.81227×aspect2+0.2211×aspect30.20431×altitude1-0.63461×altitude2+1.32375×soilthickness1+0.31163×soilthickness20.40069×erosion10.54288×erosion20.6219×erosion3+0.00087×organicmatter0.00028×NPP0.09522×humus(3) By selecting rainfall R factor, slope, amplitude of landforms, soil erosion K and humandisturbance index as the factors of forestland fragility evaluation indices, the forestland fragilitywas determinated with the factors weight which were work by the Fuzzy Analytic Hierarchy andthe advices from the relevant experts.(4) By analysing the forest using structure information with the help of the entropy theory,the forest using structure comentropy is0.7404in1999and0.7858in2008, the forestlandproductivity hierarchical comentropy is0.5468in1999and0.5624in2008in Yong’an city. Byanaslysing the correlation between the forestland fragility and the forestland qulity site08/09structured by the forestland productivity ratio between2008and1999, the results showed thatthe protection which had promoted the development of the highest Fragilityforestlandproductivity were put on the high Fragilityof regional in Yong’an city in the last10years.(5) The forestland productivity entropy tine catastrophe model were structured with the theforestland productivity “entropy” in1999and2008structured in the paper and the Fragilitygradefor step length, the entropy catastrophe theory was firest used on the analysis of correlationbetween the forestland productivity evolution in different periods and Fragilityin yong’an city.The results showed that the forestland productivity catastrophe according to the vulnerabilitiesdid not occur in1999, but occurred in2008.The mean value of the forestland productivityentropy is0.7755in2008which show down trend than the forestland productivity entropy in1999and decline with the Fragility of ascension.(6) By selecting soil organic matter, Net Primary Productivity and soil erosion3factors, theforestland productivity remote sensing monitoring model based on single-period remote sensingdata was structured, the correlation is0.9between the estimated result by the model and theforestland quality site08/99developed in the paper, which showed a consistent regular on theexpression of the forestland productivity evolution.(7) The forestland productivity is divided into "General" and "Good" according to whetherthe productivity level is higher than average. By analysing the correlation among the humandisturbance index, the forestland productivity evolution index estimated by the forestlandproductivity evolution remote sensing monitoring model, the ratio of the forestland site qualityindex which were came from the Chinese Fir, masson pine forestry resource survey samples on1996and2006in Yong’an city,0.9323was determined as the forestland productivity evalution value for different the "General" and "Good" forestland productivity estimated by the forestlandproductivity evolution remote sensing model. The forestland productivity was classified as"General" and "Good" according to the value that the volume of forest tree from forestryinformation database growth is higher than the volume mean value of forest tree growth or not totesting the effective accuracy that estimated by the forestland productivity evolution remotesensing model, the result showed the accuracy is81.95%.(8) The time-space effect of the forestland productivity evolution were analysed by theevolution of the forestland productivity in Yong’an city from1999to2008which were classifiedas4level such as “general keeping”,“good keeping”,“general to good” and “good to general”and combining with the factors such as fragility,the grade of forest site quality, ecological publicwelfare forest and commodity forest, slope and the useful forestland. The result shows that theworst decline in the forestland productivity are eucalyptus and economical forest, the worsedecline in the forestland productivity is Chinese Fir forest, the forestland productivity of massonpine and bamboo forest are better, the forestland productivity has been promoted in the seriousFragilityregional.(9) By utilizing samples data of water, soil density and soil total porosity which obtainedand measured from soil plots set in the paper, the thematic information map of forestlandwater,soil density and total porosity were extracted with the help of the inverse distanceweighted interpolation method based on GIS, the interpolation estimating precision of more than95%.(10) The forestland suitability were evaluated by the forestland suitability evaluation indexsystem which were structured by the10factors such as the forestland productivity evolutionindex, the forestland productivity, the forestland fragility, the forestland soil erosion, slope,aspect, altitude, water, soil density and soil total porosity and the AHP(analytic hierarchyprocess) method, the0-10class assignment method, the Comprehensive index. The suitability offorestland were classified into three grade with “better suitability”,“general suitability” and “badsuitability” according to the score of forestland suitability measured in the paper which has3intervals:more than7.78,6.02to7.78, under6.02, measured based on the Fuzzy ClusterAnalysis and weighted processing. The accumulative area of the “better suitability”,“generalsuitability” forestland is212693.76hm2, accounting for85.25%,87.43%of commercialforestland is suitable for exploitation and utilization. There are some favorable changes in theforestland of Yong’an city in the last10years, the area of forestland suitable for development islager, all which show that the forestland management is more Advanced in Yong’an city(11) Combining with the grade of forestland productivity, the grade of slope, the grade offragility and the forestland productivity evolution from1999to2008, the area and thedistribution of the suitable forestland were analysed and measured. The results show that there are significant relationship between the slopes and the level of forestland productivity in thechanges of forestland suitability, the more forestland area for development the more high level inthe commercial forestland productivity, the forestland area for development is decline and thearea of the “bad suitability” forestland is rising along with the ascending of the fragility andslope, the forestland productivity is decline or the forestland productivity level is low than theaverage productivity. The main area of the bad suitability forestland in the commodityforestland are on the slope in grade3,4and the declining productivity forestland, all whichdemonstrated the forestland suitability in the paper are more reasonal and applicability.更多还原