The Use of Remote Sensing to Map and Monitor Coastal Dune Vegetation Change at Southampton, Ontario, Canada

摘要

Coastal dune ecosystems in the Great Lakes Basin are fragile, rare ecosystems that are under increasing threat due to anthropogenic and natural forces. The Chantry Dune system in Southampton, Ontario is one of five major dune systems along the eastern shores of Lake Huron. The dune complex provides habitat for a diverse range of vegetation species, some of which are endemic, rare, and threatened. This research mapped and monitored dune vegetation change at the Chantry Dune system from 2005-2012 using multi-temporal normalized difference vegetation index (NDVI) images produced from QuickBird and GeoEye-1 imagery acquired in 2005 and 2012, respectively. Next, a post-classification comparison change-detection technique was applied to determine the patterns of change in vegetation cover. Finally, the maximum-likelihood classifier (MLC) was applied to the GeoEye-1 data to produce a land-use/land-cover map. Results revealed that increased vegetation growth occurred throughout the dune system while NDVI values remained unchanged or increased slightly from 2005-2012. Application of the MLC resulted in a map output with an overall classification accuracy of 97%. The results and outcomes of this research will provide much needed baseline information, which can be used by local stakeholders and authorities to improve dune management practices.