植被净初级生产力估算及与城市化水平的关系研究

【作者】 张婷媛；

【导师】 陈家治； 林文鹏；

【作者基本信息】 上海师范大学 ， 自然地理学， 2010， 硕士

【摘要】 植被净初级生产力(Net Primary Productivity,NPP)的估算是碳循环研究的基础和重要内容。本文在对SPOT5遥感影像进行几何校正、辐射校正以及土地覆盖分类等预处理的基础上,利用植被、气象等地面空间数据,应用CASA模型对2005年6月份上海市东南区域NPP进行了估算;在此基础上,对研究区域NPP空间分布特征、NPP与城市化水平之间的关系以及不同土地覆盖类型NPP之间的差异性进行了分析;最后依据分析结果对上海市的绿化布局发展方向提出了一些可行性建议。本文主要的研究工作和结论如下:(1)采用基于大气辐射理论的FLAASH模型和6S模型分别对SPOT5遥感影像进行大气校正,通过比较校正前后地物光谱曲线和利用NDVI分析两种模型的实际校正效果,证明FLAASH模型能较好地消除大气的影响,对SPOT5影像具有更好的校正作用。(2)通过支持向量机分类方法和最大似然法分类方法分别提取了研究区域的土地覆被信息。分析分类结果图和Kappa系数,对两种方法的分类结果进行评价,结果表明:支持向量机分类准确度高,地物错分现象相对较少。(3)通过对研究区NPP空间分布分析发现: NPP在各区内均出现了高值,总体上市区高值出现的概率相对较低,浦东、南汇、奉贤、金山等郊区高值出现的概率相对较高。这与市区存在林地及农田较少,郊区存在林地和农田较多的现状还是比较吻合的。(4)NPP与城市化水平之间的关系研究表明: NPP与城市化水平是密切相关的,总体上呈现出随城市化水平提高,NPP逐渐降低的趋势。对各区域内不同土地覆盖类型NPP比较分析发现:林地NPP最大,固碳能力最高,草地次之,农田植被固碳能力相对较低。更多还原

【Abstract】 Estimate of NPP is the basis and main content of carbon cycle research. On the basis of the Vegetation Classification of SPOT5 remote sensing image,Remote sensing estimation of vegetation NPP of parts of Shanghai in June 2005 was realized.The estimate made use of CASA model,weather and other ground data. The thesis has analysised the relationship between NPP and the level of urbanization and NPP of various land cover types in various regions.Finally, the study put forward advice for the sustainable development plan of Shanghai .In this paper, the research work and conclusions are as follows:(1)The FLAASH and 6S are used as the atmospheric correction models.In order to evaluate and validate the correction results, The object spectral reponse curve and NDVI profile curve were used to analyse the images after atmospheric correction.As a result, the FLAASH atmospheric correction model is better to SPOT5 remote sensing images than 6S model.(2) The land cover information was extracted from the study area making use of the support vector machine classification and maximum likelihood classification. Evaluating the classification results of two methods evaluate, the results show that the accuracy of Support vector machine classification is higher and the phenomenon of surface features misclassification is less.(3)According to NPP density distribution map, High value of NPP is in the various regions of the study area. Generally, the probability of high value appears relatively low in urban area, and the probability of high-value appeared relatively high in Pudong, Nanhui, Fengxian, Jinshan and other suburbs. The phenomenon consistent with the facts that woodland and farmland are less in urban, and woodland and agricultural land are relatively more.(4) NPP is closely related to the level of urbanization.With the level of urbanization increased, NPP is decreased. NPP of different land cover types are comparatively analyzed in each region.The results show that the forest NPP is the largest;carbon sequestration capacity is the highest, followed by grassland, and cropland carbon sequestration capacity is relatively low.更多还原