本文为英国格拉斯哥大学（作者：Fathi Goma Al Sghair）的博士论文，共281页。

遥感（RS）和地理信息系统（GIS）方法与地面实况相结合，为高级生态系统管理提供了新的工具，提供了在本地、区域和全球范围内监测随时间变化的能力。在这项研究中，利用遥感（Landsat TM和航空照片）和地理信息系统，并结合地面实况的工作，评估英国两个对比湿地区域的植被随时间变化情况，分别是：1984年至2009年的威肯汾淡水湿地，1988年至2009年的卡雷洛克保护区盐沼。

2010年6月14日至18日，在威肯汾进行了地面实况研究：共采集40个1 m²的样方，沿不同植被类型的6个横断面随机放置。2011年7月5日至9日在第二个研究区域卡雷洛克保护区进行了调查，共48个1 m²样方沿研究区域内不同植被类型的七个横断面随机分布。采用双向指示物种（TWINSPAN）分析法对地面真值样本进行分类，对高值特征值（>0.500）进行分离，确定样本的端组。根据威肯汾40个样方的数据，将样本分为4个样方组，而卡雷洛克保护区的48个样方分为5个样方组。利用地理信息系统（GIS）结合地面实况资料，对航空照片中的植被覆盖进行了初步分析，采用与航空摄影解释相同的无监督和监督分类技术解释Landsat TM图像中的植被覆盖。

在威肯汾，1984年8月18日和2009年8月23日使用了Landsat TM图像；对于卡雷洛克保护区，1988年5月14日和2009年7月11日使用了Landsat TM图像。威肯汾的航空摄影图像来自1985年和2009年；而卡雷洛克保护区的航空摄影图像来自1988年和2009年。航空照片和Landsat TM图像分析的结果显示，在威肯汾研究期间，植被发生了重大的时间变化，最有可能是由管理方案产生的，而不是由于自然变化。在卡雷洛克保护区，航空摄影解释的结果表明，在1988年至2009年期间，灌木覆盖层略有变化，但在研究期间其他变化不大。结果表明，航空摄影分类精度高于Landsat TM数据。

航空摄影和Landsat TM分类精度的差异，特别表现在卡雷洛克保护区，这是由于Landsat TM图像的分辨率较低，一些植被所占面积小于TM图像像素大小。根据测绘实践，航空照片比Landsat TM图像能够产生更好的植被类别（与地面真实数据相比），因为航空照片比Landsat  TM图像具有更高的空间分辨率。也许本研究最重要的结论是，本文提供了证据来证明RS/GIS方法可以提供有关湿地植被随时间变化的有用基线数据，并跨越相当广阔的区域，从而为湿地生态环境的管理和保护提供有价值的信息。

Remote Sensing (RS) and GeographicInformation System (GIS) approaches, combined with ground truthing, areproviding new tools for advanced ecosystem management, by providing the abilityto monitor change over time at local, regional, and global scales. In thisstudy, remote sensing (Landsat TM and aerial photographs) and GIS, combinedwith ground truthing work, were used to assess wetland vegetation change overtime at two contrasting wetland sites in theUK: freshwater wetland at WickenFen between 1984 and 2009, and saltmarsh between 1988 and 2009 in Caerlaverock Reserve. Ground truthing studieswere carried out in Wicken Fen (UK National Grid Reference TL 5570) during 14th- 18th June 2010: forty 1 m2quadrats were taken in total, placed randomly along six transects in differentvegetation types. The survey in the second Study Area Caerlaverock Reserve (UKNational Grid Reference NY0464) was conducted on 5th - 9 th July 2011, with atotal of forty-eight 1 m2quadrats placed randomly along seven transects in different vegetation typeswithin the study area. Two-way indicator species (TWINSPAN) was used forclassification the ground truth samples, taking separation on eigenvalues withhigh value (>0.500), to define end-groups of samples. The samples wereclassified into four sample-groups based on data from 40 quadrats in WickenFen, while the data were from 48 quadrats divided into five sample-groups inCaerlaverock Reserve. The primary analysis was conducted by interpretingvegetation cover from aerial photographs, using GIS combined with ground truthdata. Unsupervised and supervised classifications with the same technique foraerial photography interpretation were used to interpret the vegetation coverin the Landsat TM images. In Wicken Fen, Landsat TM images were used from 18thAugust 1984 and 23rd August 2009; for Caerlaverock Reserve Landsat TM imageryused was taken from 14th May 1988 and 11th July 2009. Aerial photograph imageryfor Wicken Fen was from 1985 and 2009; and for Caerlaverock Reserve, from 1988and 2009. Both the results from analysis of aerial photographs and Landsat TMimagery showed a substantial temporal change in vegetation during the period ofstudy at Wicken Fen, most likely primarily produced by the managementprogramme, rather than being due to natural change. In Cearlaverock Reserve,results from aerial photography interpretation indicated a slight change in thecover of shrubs during the period 1988 to 2009, but little other change overthe study period. The results show that the classification accuracy usingaerial photography was higher than that of Landsat TM data. The difference ofclassification accuracy between aerial photography and Landsat TM, especiallyin Caerlaverock Reserve, was due to the low resolution of Landsat TM images,and the fact that some vegetation classes occupied an area less than that ofthe pixel size of the TM image. Based on the mapping exercise, the aerialphotographs produced better vegetation classes (when compared with ground truthingdata) than Landsat TM images, because aerial photos have a higher spatialresolution than the Landsat TM images. Perhaps the most important conclusion ofthis study is that it provides evidence that the RS/GIS approach can provideuseful baseline data about wetland vegetation change over time, and acrossquite expansive areas, which can therefore provide valuable information to aidthe management and conservation of wetland habitats.

1 引言及文献回顾

2 研究方法

3 威肯汾研究区域1

4 卡雷洛克保护区研究区域2

5 调查方法的一般讨论与比较

6 结论与研究建议

更多精彩文章请关注公众号：