Ecosystem Classification of the Hudson Bay Lowlands
Description
Ecosystems in the Hudson Bay Lowlands (HBL) form one of the most ecologically significant regions in Canada, shaped by the interplay of tundra, taiga, and one of the world’s largest peatland complexes. These ecosystems support diverse wildlife including migratory birds, polar bears, boreal species and sustain the cultural, subsistence, and spiritual relationships of Cree, Dene, and Métis Nations. The ecological functions of the HBL are globally important, containing immense carbon stores in deep peat deposits. The HBL has become increasingly vulnerable to rapid warming, altered hydrology, and shifts in permafrost conditions. As these changes accelerate, understanding ecosystem extent and distribution becomes crucial for assessing risks to biodiversity, carbon stocks, and climate feedbacks.
Accurate mapping of ecosystem extent in Hudson Bay is also essential for Canada’s international reporting obligations and global conservation commitments. Ecosystem extent is a foundational indicator across major frameworks, including the UN Sustainable Development Goals, UNFCCC greenhouse gas reporting, and especially the Kunming-Montreal Global Biodiversity Framework, where it underpins progress toward the 2050 Vision for Biodiversity. Because models used for national carbon accounting (e.g., CBM-CFS3 and GCBM) rely on precise ecosystem-type inputs, consistent and spatially explicit mapping of the HBL is necessary to quantify carbon sinks, monitor ecosystem change, and report transparently to bodies such as the CBD and IPCC. In a rapidly warming region with globally significant carbon stores, comprehensive mapping ensures that Canada can track ecosystem change, safeguard biodiversity, and meet its international environmental commitments.
This product was produced by using Classification by Progressive Generalization (CPG) which uses enhanced false-colour Landsat Earth-observation imagery, specifically Red, NIR, and SWIR bands as the input for a multistep clustering and merging process. Each band is first contrast-stretched to widen the dynamic range of key features such as vegetation and soils while compressing dark (water, shadow) and bright (cloud, ice) extremes. K-means clustering is then applied to the enhanced 3-band composite to produce 200 spectral clusters, which are colour-coded using the average RGB brightness of the corresponding Landsat bands. This initial reduction from millions of possible spectral signatures to 200 clusters forms the first “generalization.” Further generalization is achieved by merging clusters based on thematic and spectral similarity and spatial adjacency. Once merging is complete, a lookup table converts the 200 clusters into the final mapped classes.
Resources
| Name |
Format |
Description |
Link |
| Ecosystem Classification of the Hudson Bay Lowlands |
HTML |
|
https://data-donnees.az.ec.gc.ca/data/water/scientificknowledge/Ecosystem-Classification-of-the-Hudson-Bay-Lowlands?lang=en |
| Classification des écosystèmes des basses terres de la baie d'Hudson |
HTML |
|
https://data-donnees.az.ec.gc.ca/data/water/scientificknowledge/Ecosystem-Classification-of-the-Hudson-Bay-Lowlands?lang=fr |