Canadian Forest Service Publications
EOSD Land Cover Classification Legend Report. Version 2. 2003. Wulder, M.A.; Nelson, T.A. Natural Resources Canada, Canadian Forest Service, Pacific Forestry Centre, Victoria, BC. 81 p.
Issued by: Pacific Forestry Centre
Catalog ID: 33752
CFS Availability: PDF (download)
Concern over the state of the earth’s environment has resulted in an increased need for accurate land cover information. Ecologically, land cover mapping is necessary to provide input to scientific models, to ensure forest management is sustainable, and to monitor environmental change. However, land cover mapping is also required to meet political commitments made to international agreements such as the Kyoto Protocol and the Convention on Biological Diversity.
In large countries, such as Canada, satellite imagery is a useful tool for land cover mapping as it efficiently provides information over relatively large areas. The Canadian Forest Service’s project, Earth Observation for Sustainable Development of Forests (EOSD), is mapping Canada’s forests with satellite imagery.
The goal of this report is to review the current EOSD Land Cover Classification Scheme to determine if it is suitable for national level mapping with Landsat imagery. By reviewing several key land cover legends as well as the development of the proposed EOSD legend we are able to make recommendations on how to improve the EOSD land cover legend and to aid the efficiency of the mapping project. In this report we recommend a legend for the land cover mapping of the forested area of Canada with Landsat data. As of this version update, the recommended legend is in use by federal, provincial, and territorial mapping agencies undertaking the EOSD Land Cover mapping activities.
Our most significant conclusion is that with minor modifications the proposed EOSD legend will be appropriate for mapping Canada’s land cover with Landsat imagery. EOSD wetland classes should be regrouped and the Wetland – Bryoid class removed. The EOSD legend should continue to be based on the hierarchical National Forest Inventory (NFI) Land Cover Classification Scheme as support from the Canadian forest Inventory Committee will aid in collaboration efforts with the NFI and provinces. As well, many of the provincial large area land cover classifications legends are similar to the EOSD legend allowing for partnerships and increased efficiency of land cover mapping. Land use and common classes, such as burn and harvest, should not be added to the EOSD legend. At present, techniques to classify land use and transitional areas with spectral information from Landsat spatial resolution satellites are not well developed.