The feasibility of using landsat TM imagery for longterm vegetation monitoring in the fynbos biome

[摘要] ENGLISH SUMMARY: Those who would successfully manage the fynbos biome require managementoriented information. Managers, strategic planners and researchers often lackinformation about fynbos areas and changes that occurs in these areas upon whichthey can, with reasonable confidence, base decisions. Landsat Thematic Mapper (TM)data provides a potentially strong base to work from. The objective of this project wasto determine the feasibility of using TM imagery to quantify fynbos communitieswithin a specific area that can be used for the long term vegetation monitoring in thefynbos biome. TM imagery have advantages over other data sources such as aerialphotography, for large area land cover classifications, in part because of their frequentrepeat cycles, large-area sample, wide spectral range, cost effectiveness andamenability to automated classification. Considering the very detailed scale the studywas done at, another objective of the project was to determine the feasibility of usingTM imagery to quantify fynbos communities at a far more detailed scale than therecommended operating scale for TM imagery. At this detailed scale fynbosvegetation can be managed effectively in the long term. A small scale vegetation andvegetation community study of the study area was done to obtain a detailed groundmap. The vegetation communities as defined by image processing of the TM data wascompared with the accurate ground map. By means of this comparison it was shownthat TM imagery could be used effectively to obtain a detailed fynbos communitydescription of a fynbos area at the desired scale. Another objective of the study was todetermine the feasibility of Principal Component Analysis (PCA) to reduce theamount of natural variation and noise within TM imagery. It has been shown thatPCA can be used effectively to achieve the latter and by doing so, baseline imagerycan be obtained for a specific area and time period in terms of its vegetationcommunities, while portraying the maximum environmental variation within theimagery. The final part of the study entailed the determination of the feasibility ofusing TM imagery to determine vegetation change over time. It has been shown thatTM imagery can be used effectively for this purpose. This resulted in some veryinteresting findings, especially with the recovery of burnt fynbos areas. The objectivesof the study have been met and this scale of monitoring holds much promise. Furtherapplication is warranted in different areas as well as with additional thematic data.

[发布机构] Stellenbosch University