The goals of this thesis are to assess the use of remote sensmg and
Geographic Information Systems (GIS) to map and classify coastal wetland
habitat along the entire coast of eastern Georgian Bay, Lake Huron. Little
mapping has been completed in this region where there is potentially the largest
concentration of coastal wetlands in the Great Lakes. In chapter 1, we developed a method that uses high-resolution IKONOS imagery (1-m resolution) with an object-based approach to classify wet meadow vegetation in these coastal wetlands, and assessed the transferability of classification rulesets developed independently for 3 different satellite scenes. We showed that 4 different classes (meadow/shrub, emergent, senescent vegetation, and rock) can be mapped with an overall accuracy of 76%. When classification rulesets developed for individual scenes were transferred to other scenes without gathering additional field information for those scenes, we found a difference in accuracy of about 5%. This difference in accuracy is acceptable considering the trade-off in costs associated with field surveys. We recommend that managers use IKONOS in fine-scale habitat mapping and that rulesets only be developed for geographically distinct areas. In Chapter 2, we conducted a study to test the feasibility of using this mapping approach to complete the field surveys required in Ontario Wetland Evaluation System (OWES). In addition, we determined empirically how inclusion of vegetated deep-water habitat below 2 m can affect relevant OWES component scores, because the current system does not consider any vegetated habitat below 2 m, even though this portion of coastal wetlands is known to provide critical habitat for many Great Lakes fishes. We sampled 16 wetlands that varied in size and inundation characteristics and grouped them into 4 categories: small aquatic, small terrestrial, large aquatic, and large terrestrial. When the vegetated deep-water habitat was included, total wetland area and the overall score for all assessed criteria assessed increased significantly; however, this increase was not sufficiently large to make any practical difference in the overall score using existing the point-scale. This is largely because submerged aquatic habitat is not adequately represented in current evaluation protocols and is severely undervalued. In chapter 3 we developed a method to quantify and monitor change in coastal marsh habitat in southeastern Georgian Bay using multi-temporal IKONOS imagery. We detected a significant increase in the proportion of terrestrial habitat (high marsh) at the expense of the aquatic habitat (low marsh) over six years from 2002 to 2008. There did not appear to be any effect of human activities (indicated by the number of buildings within 500 m of wetlands) on habitat changes. We conclude that water levels may currently exert greater pressure on these systems than does cottage density in the region. We recommend that the approaches developed in this study be applied as quickly as possible to comprehensively map existing wetland habitat in eastern Georgian Bay to monitor responses to further water-level and human-induced disturbance. / Thesis / Master of Science (MSc)
Identifer | oai:union.ndltd.org:mcmaster.ca/oai:macsphere.mcmaster.ca:11375/21861 |
Date | 04 1900 |
Creators | Rokitnicki-Wojcik, Daniel |
Contributors | Chow-Fraser, Patricia, Biology |
Source Sets | McMaster University |
Language | English |
Detected Language | English |
Page generated in 0.0014 seconds