Please use this identifier to cite or link to this item:
|Title:||Wetland Classification and Environmental Monitoring in the Peace-Athabasca Delta using LANDSAT Digital Data|
|Authors:||Wickware, McIntosh Gregory|
|Abstract:||<p>Through the use of LANDSAT digital date and automated image classification techniques, habitats of the Peace-Athabasca Delta region in northeastern Alberta were classified, and mapped, a number of change detection techniques evaluated, and the potential of using LANDSAT digital data in a long-term monitoring program assessed.</p> <p>The General Electric Image 100 Multispectral Image Analysis system was used to conduct a supervised classification for a 1200 Km² test site centered on Mamawi Lake at the east end of the Delta complex. To provide optimum opportunity for detecting change and for testing the classification of habitats under differing conditions, two images, one with normal water levels (August 1, 1976) and one with above normal water levels (August 26, 1973), were selected for analysis. A total of eight habitats was identified for the 1976 data and seven for the 1973 date.</p> <p>Classification accuracy of the 1976 habitats was considered excellent when compared with available ground reference data. Although ground reference data were not available for 1973, classification results are considered good. Spatial filtering of the classified scenes was found to improve the visual quality of the images, although only marginal improvements in classification accuracy resulted.</p> <p>The classified scenes were found to be useful in delineating major topographic subdivisions within the Delta. These subdivisions can be used to identify areas most susceptible to change. Using an image registration technique, the 1973 and 1976 scenes were overlayed and two change detection analyses carried out. These were ratioing and post-classification change detection. Ratioing was found to be relatively fast and simple method for determining the areal extent of flooding and areas of major vegetation change. Post classification change detection was carried out for the entire study ares as well as 1/2 and 1/8 areas. The technique was found to be useful in identifying the nature of change, although further research is required before an evaluation of the absolute values generated in the analyses can be fully assessed.</p> <p>It is anticipated that when extrapolated over the entire 37,500 Km² area of the Delta, the use of digital LANDSAT data will provide a relatively low cost and yet effective means of maintaining a current habitat map of the region, as well as providing an effective method for detecting and monitoring environmental change.</p>|
|Appears in Collections:||Open Access Dissertations and Theses|
Items in MacSphere are protected by copyright, with all rights reserved, unless otherwise indicated.