0639-B1
J.A. Dechka, A. Beaudoin, D.G. Goodenough, M.D. Gillis, R.J. Hall, D.G. Leckie, J.E. Luther and M.A. Wulder 1
In support of national and international reporting requirements, the Canadian Forest Service, in partnership with the Canadian Space Agency, is using space-based, earth observation technologies to support the monitoring of sustainable development of Canada's forests through an initiative called Earth Observation for Sustainable Development of Forests (EOSD). This project will assist in the implementation of Canada's international obligations in relation to climate change and sustainable forest management.
The current priority of EOSD is mapping circa-year 2000 land cover in the forested areas of Canada, in partnership with the provinces, territories, universities and industry. These maps are scheduled for completion in 2006. A longer-term goal of EOSD is to produce maps of forest change, biomass and land cover on a five- to ten-year cycle. An important component of EOSD is research directed at fulfilling these long-term goals. Research is focusing on developing, improving and automating procedures for land cover, change monitoring, and biomass measurements.
In addition to thematic maps, products generated in this programme will include data, methods and systems that will be freely available and distributed via the internet. These products will be an integral component of Canada's National Forest Information System, which is a new forest measuring and monitoring system that will help the public and interested organizations understand the composition, distribution and dynamics of Canada's forests.
Canada has a great interest in the sustainable management of its forests since they cover nearly 50% of the total landmass of the country (Natural Resources Canada 2001), and they contribute to the Canadian economy in various ways. Forests support the largest industry in the country by generating over 373 000 direct jobs and contributes over $37 billion to our balance of trade (Forestry Canada 2001). In addition, our forests support a multibillion dollar recreation and tourism industry. Thus, Canada's ability to manage its forest resources in a sustainable manner is essential for the future well being of the forest, recreation, and tourism sectors.
An ecosystem-based approach to sustainable forest management (SFM) has become an important objective due to the multitude of environmental, social, and economic values that are derived from Canada's forests. The new National Forest Strategy has been assembled as a framework to guide the implementation of policies that will influence the practice of sustainable forest management into the new millennium (Canadian Council of Forest Ministers 1998). For example, two foci within this Strategy are to broaden the scope of forest inventories and to develop objective measures of sustainability through a set of criteria and indicators. These measures will allow Canada to both assess the results of its management practices and monitor the potential effects of a changing climate on the sustainability of its forests.
Achieving sustainable forest management will require an effective measuring and monitoring program that can provide timely forest information to support policy issues related to sustainable development. Currently, Canada's ability to deliver on our national and international commitments related to sustainable forest management is currently beyond the combined capacity of federal, provincial, territorial, and industrial inventory and information systems. Therefore, Canada is developing the National Forest Information System (NFIS) to provide information to various user groups and stakeholders that will allow them to address questions about sustainable forest management (Addison et al. 1999).
As one component of the NFIS, the Canadian Forest Service (CFS), in partnership with the Canadian Space Agency (CSA), is undertaking an initiative called Earth Observation for Sustainable Development of Forests (EOSD). This project will use space-based earth observation technologies to create products and develop methods in support of the NFIS (Wood et al. 2002; Goodenough et al. 1999). EOSD will use Landsat 30 m spatial resolution satellite data to map the land cover2, change, and biomass of the forested areas of Canada. Specifically, the goals of the EOSD project are to:
A priority of this project is to produce land cover maps of the forested areas of Canada to support reporting requirements related to climate change (UNFCCC and Kyoto Protocol) and landscape management. In addition, research will be focused on developing and improving methods for mapping forest land cover, forest change, and biomass. An important component of this research and development is the work on integrating, automating, and streamlining of processes and systems. Completion of these tasks will require the participation of various federal, provincial, and territorial government agencies, and universities and industries.
Implementation of the EOSD program for the purpose of monitoring forest land cover began in the Spring 2002. Its priority was to produce land cover map products of the forested area of Canada for circa year 2000 (Wulder 2002). In addition to land cover maps, other key products that EOSD will provide in the long term include:
These products will be distributed freely over the web to the public and other interested parties.
Using Landsat-7 Enhanced Thematic Mapper (ETM+) imagery as the primary data source, and based on a classification approach developed by Wulder et al. (2002a), land cover mapping is scheduled to be complete for the forested area by 2005/2006. This schedule assumes timely delivery of a national Landsat-7 ETM+ orthoimage coverage (Wulder et al. 2002b) suitable for vegetation mapping. Modifications to the production schedule will occur based on the supply and seasonality of these data, as processes become more efficient, and if there are changes in resources or priorities.
As part of this mapping process, a classification legend for EOSD has been derived from the National Forest Inventory (NFI) class structure (Wulder and Nelson 2001) (Table 1). The NFI legend is a hierarchical land cover classification that was developed in co-operation with provincial and territorial forest inventory agencies through the Canadian Forest Inventory Committee (CFIC). The level of information detail being collected for EOSD represents a layer in the NFI hierarchy that is more generalized in order to be compatible with information typically captured from Landsat satellite imagery. This level would also be more appropriate for translation among other national and international vegetation classification systems (National Biological Service and National Park Service 1994).
Cooperation with federal, provincial, and territorial agencies has benefited EOSD in developing a land cover legend and has also assisted in moving towards completing the land cover map products of Canada (Figure 1). Land cover mapping activities are currently in progress in partnership with provincial and territorial government agencies in Ontario, Alberta, Saskatchewan, Newfoundland and Labrador, Northwest Territories, and Yukon.
Second year implementation activities are being planned in Quebec, and Manitoba, while British Columbia, New Brunswick, Nova Scotia, Nunavut, and Prince Edward Island are scheduled for implementation in the third year.
Forest change and biomass are also EOSD priorities. Current activities in these areas are concentrating on method development and testing. These methods need to be streamlined and automated because of the large amount of data associated with a single coverage of the country's forested area (Wulder and Seemann 2001). Therefore, researchers are investigating a distributed data storage system based on grid architecture that will link four regional CFS production centres to provide partners and the public with access to data products.
While land cover implementation has begun, improvements and additions to existing methods and processing issues are being pursued. Some issues under investigation include looking at the most appropriate way to validate forest land cover products and a recommended procedure for combining classified images.
Table 1: NFI land cover legend for EOSD.
Class |
Description |
No Data |
|
Cloud |
|
Shadow |
|
Snow/Ice |
Glacier/snow |
Rock/Rubble |
Bedrock, rubble, talus, blockfield, rubbley mine spoils, or lava beds. |
Exposed Land |
River sediments, exposed soils, pond or lake sediments, reservoir margins, beaches, landings, burned areas, road surfaces, mudflat sediments, cutbanks, moraines, gravel pits, tailings, railway surfaces, buildings and parking, or other non-vegetated surfaces. |
Water |
Lakes, reservoirs, rivers, streams, or salt water. |
Shrub - Tall |
At least 20% ground cover which is at least one-third shrub; average shrub height greater than or equal to 2 m. |
Shrub - Low |
At least 20% ground cover which is at least one-third shrub; average shrub height less than 2 m. |
Herb |
Vascular plant without woody stem (grasses, crops, forbs, gramminoids); minimum of 20% ground cover or one-third of total vegetation must be herb. |
Bryoids |
Bryophytes (mosses, liverworts, and hornworts) and lichen (foliose or fruticose; not crustose); minimum of 20% ground cover or one-third of total vegetation must be a bryophyte or lichen. |
Wetland - Treed |
Land with a water table near/at/above soil surface for enough time to promote wetland or aquatic processes; the majority of vegetation is coniferous, broadleaf, or mixed wood. |
Wetland - Shrub |
Land with a water table near/at/above soil surface for enough time to promote wetland or aquatic processes; the majority of vegetation is tall, low, or a mixture of tall and low shrub. |
Wetland - Herb |
Land with a water table near/at/above soil surface for enough time to promote wetland or aquatic processes; the majority of vegetation is herb. |
Coniferous - Dense |
Greater than 60% crown closure; coniferous trees are 75% or more of total basal area. |
Coniferous - Open |
26-60% crown closure; coniferous trees are 75% or more of total basal area. |
Coniferous - Sparse |
10-25% crown closure; coniferous trees are 75% or more of total basal area. |
Broadleaf - Dense |
Greater than 60% crown closure; broadleaf trees are 75% or more of total basal area. |
Broadleaf - Open |
26-60% crown closure; broadleaf trees are 75% or more of total basal area. |
Broadleaf - Sparse |
10-25% crown closure; broadleaf trees are 75% or more of total basal area. |
Mixed Wood - Dense |
Greater than 60% crown closure; neither coniferous nor broadleaf trees account for 75% or more of total basal area. |
Mixed Wood - Open |
26-60% crown closure; neither coniferous nor broadleaf trees account for 75% or more of total basal area. |
Mixed Wood - Sparse |
10-25% crown closure; neither coniferous nor broadleaf trees account for 75% or more of total basal area. |
Figure 1: Implementation of EOSD for forest land cover, beginning in 2002.
Once a land cover map of Canada has been completed, future land cover maps may be created by reclassifying new imagery or by using forest change procedures to update existing products. Forest change events can be used to identify and update land cover over a 5- to 10-year period, and this will help attribute a cause to the change event (Leckie et al. 2002).
Measuring biomass is complex and must be modeled at both tree and stand levels in order to acquire sufficient source data for mapping. With a combination of remotely sensed data and forest inventory techniques, procedures are being refined to map the forest biomass of Canada and to fill gaps in the national biomass inventory. Progress in biomass mapping was recently reported (Guindon et al. 2002; Hall et al. 2002; Luther et al. 2002), and future efforts are focusing on building biomass regressions for non-inventoried areas and assessing reflectance models to improve cover type and structure estimation for biomass mapping in non-inventoried areas.
Advancements in monitoring and mapping technologies are increasing rapidly. As a result, efforts are being made to track new initiatives, and sensor, processing, and analysis technologies that may assist in the mapping of Canada's forest lands. These technologies will be implemented if they are suitable to the program. One such technology under consideration is Radarsat-2; research is being conducted to investigate its potential to estimate biomass in non-inventoried areas and to detect forest change.
A mechanism to provide access to distributed EOSD products is the National Forest Information System (Addison et al. 1999) (Figure 2). NFIS is the highway for Canada's new measuring system that links distributed data sources and products with users.
Two major components of this system are the EOSD (Wood et al. 2002) and the National Forest Inventory (Gillis 2001). These components are closely linked and complementary in that data from one program will be used by the other and vice versa. For example, NFI will be generating photo plot and ground plot information in inventoried areas that can be used to calibrate or validate EOSD land cover products. Conversely, EOSD will produce land cover maps from which photo plots will be created in non-inventoried areas. It is hoped that the EOSD products will extend beyond the 1% NFI sample to provide more information about the land base that is compatible with the NFI class structure. In addition, in some northern and other non-managed forest areas, EOSD land cover classifications of the NFI plots will be used instead of acquiring and interpreting aerial photographs.
Information from EOSD and NFI will provide data on our forested areas and contribute towards Canada's efforts to develop a National Forest Carbon Accounting Framework. This framework is required to meet Canada's commitments, under the UNFCCC and under the Kyoto Protocol, to report on carbon stocks.
Figure 2: An illustration of Canada's new forest measuring and monitoring system.
EOSD will contribute to meeting Canada's national and international reporting requirements related to climate change and sustainable forest management by mapping the forested areas of Canada. This mapping task will require the support and concerted effort of many partners to complete a circa year 2000 land cover map by 2006. In addition to land cover mapping, research efforts are working towards automation procedures as well as methods of change and biomass mapping. Products generated from this project will be an integral component of Canada's new forest measuring and monitoring system and will assist the public and interested organizations in understanding the composition, distribution, and dynamics of Canada's forests.
Addison, P., R. Morrison, and R. Quenet, 1999. A national forest information system for Canada. Proceedings of the 13th Annual Conference on Geographic Information Systems (GIS 99), March 1-4, 1999, Vancouver, BC.
Bhogal, A.S., D.G. Goodenough, H. Chen, G. Hobart, B. Rancourt, M. Murdoch, J. Love, and A. Dyk, 2002. "Automated methods for atmospheric correction," pp. 1316-1319 in Proceedings of the International Geoscience and Remote Sensing Symposium 2002/24th Canadian Symposium on Remote Sensing, Toronto, ON. IEEE, Piscataway, NJ.
Bonnor, G.M., 1985. Inventory of forest biomass in Canada. Petawawa National Forestry Institute, Canadian Forest Service, ON.
Canadian Council of Forest Ministers, 1997. Criteria and indicators of sustainable forest management in Canada. Natural Resources Canada, Canadian Forest Service, Ottawa, ON. Technical Report.
------, 1998. National forest strategy, 1998-2003: sustainable forests, a Canadian commitment. Natural Resources Canada, Canadian Forest Service, Ottawa, ON. 47 p. http://nfsc.forest.ca/strategy.html
Cihlar, J., 1999. Land cover mapping of large areas from satellites: status and research priorities. International Journal of Remote Sensing 21(6&7):1093-1114.
Forestry Canada, 2001. The state of Canada's forests 2000-2001, sustainable forestry: a reality in Canada. Natural Resources Canada, Canadian Forest Service, Ottawa, ON.
Gillis, M.D., 2001. Canada's national forest inventory (responding to current information needs). Environment Monitoring and Assessment 67:121-129.
Goodenough, D.G., A.S. Bhogal, A. Dyk, R. Fournier, R.J. Hall, J. Iisaka, D. Leckie, J.E. Luther, S. Magnussen, O. Niemann, and W. M. Strome, 1999. "Earth observation for sustainable development of forests-a national project," pp. 2101-2104 in Proceedings of the International Geoscience and Remote Sensing Symposium 1999, Hamburg, Germany. IEEE, Piscataway, NJ.
Guindon, L., R.A. Fournier, J.E. Luther, R.J. Hall, E. Arsenault, D.E. Piercey and M.-C. Lambert, 2002. "Mapping forest biomass on several pilot regions in Canada with Landsat TM and forest inventory data," pp. 886-888 in Proceedings of the International Geoscience and Remote Sensing Symposium/24th Canadian Symposium on Remote Sensing, Toronto, ON. IEEE, Piscataway, NJ.
Hall, R.J., B.S. Case, E. Arsenault, D.T. Price, J.E. Luther, D.E. Piercey, L. Guindon, and R.A. Fournier, 2002. "Modeling and mapping forest biomass using forest inventory and Landsat TM data: results from Foothills Model Forest, Alberta," pp. 1320-1323 in Proceedings of the International Geoscience and Remote Sensing Symposium 2002/24th Canadian Symposium on Remote Sensing, Toronto, ON. IEEE, Piscataway, NJ.
Leckie, D.G., N. Walsworth, J. Dechka, M.A. Wulder, 2002. "An investigation of two date unsupervised classification in the context of a national program for Landsat based forest change mapping," pp. 1307-1311 in Proceedings of the International Geoscience and Remote Sensing Symposium 2002/24th Canadian Symposium on Remote Sensing, Toronto, ON. IEEE, Piscataway, NJ.
Luther, J.E., R.A. Fournier, R.J. Hall, C.-H. Ung, L. Guindon, D.E. Piercey, and M.-C. Lambert, 2002. A strategy for mapping Canada's forest biomass with Landsat TM imagery," pp. 1312-1315 in Proceedings of the International Geoscience and Remote Sensing Symposium 2002/24th Canadian Symposium on Remote Sensing, Toronto, ON. IEEE, Piscataway, NJ.
National Biological Service and National Park Service, 1994. USGS - NPS vegetation mapping program, standardized national vegetation classification system. USDA. http://biology.usgs.gov/npsveg/classification/index.html
Wood, J.E., M.D. Gillis, D.G. Goodenough, R.J. Hall, D.G. Leckie, J.L. Luther, and M. Wulder, 2002. "Earth observation for sustainable development of forests (EOSD): project overview," pp. 1299-1302 in Proceedings of the International Geoscience and Remote Sensing Symposium 2002/24th Canadian Symposium on Remote Sensing, Toronto, ON. IEEE, Piscataway, NJ.
Wulder, M., M. Cranny, and J. Dechka, 2002a. An illustrated methodology for land cover mapping of forests with Landsat-7 ETM+ data: methods in support of EOSD land cover, Version 2. Natural Resources Canada, Canadian Forest Service, Pacific Forestry Centre, Victoria, BC. 39 p.
Wulder, M., E. Loubier, and D. Richardson, 2002b. A Landsat-7 ETM+ orthoimage coverage of Canada. Canadian Journal of Remote Sensing 28(5):667-671.
Wulder, M. and T. Nelson, 2001. EOSD legend: characteristics, suitability, and compatibility. Canadian Forest Service, Pacific Forestry Centre, Victoria, BC. 74 p.
Wulder, M. and D. Seemann, 2001. Spatial partitioning Canada with the Landsat worldwide referencing system. Canadian Journal of Remote Sensing 27(3):225-231.
Wulder, M.A., 2002. "Mapping land cover of the forested areas of Canada," pp. 1303-1306 in Proceedings of the International Geoscience and Remote Sensing Symposium 2002/24th Canadian Symposium on Remote Sensing, Toronto, ON. IEEE, Piscataway, NJ.
1 Natural Resources Canada, Canadian Forest Service, Pacific Forestry Centre, 506 West Burnside Road, Victoria, BC, V8Z 1M5, Canada. [email protected]; Website: www.pfc.forestry.ca
2 "Land cover" and "land use" are terms that are often used interchangeably; however, they have different meanings. Land cover results from a complex mixture of natural and anthropogenic influences and refers to the composition and characteristics of land surface elements (Cihlar 1999). In contrast, land use is characterized by economic uses of land and people's relationships with the environment (Wulder 2002).