Climate change and permafrost thaw-induced boreal forest loss in northwestern Canada
Permafrost distribution throughout the western Canadian subarctic is not well understood due to the remoteness and size of the region, its spatial and temporal heterogeneity, limited data availability, and sparse monitoring networks. These factors severely challenge investigations of how climate war...
| Published in: | Environmental Research Letters |
|---|---|
| Main Authors: | , , , |
| Format: | Article in Journal/Newspaper |
| Language: | English |
| Published: |
IOP Publishing
2018
|
| Subjects: | |
| Online Access: | https://doi.org/10.1088/1748-9326/aad74e https://doaj.org/article/63cc214af67f42a48cb22a8d54c6d9a5 |
| Summary: | Permafrost distribution throughout the western Canadian subarctic is not well understood due to the remoteness and size of the region, its spatial and temporal heterogeneity, limited data availability, and sparse monitoring networks. These factors severely challenge investigations of how climate warming might affect the distribution of permafrost and provide strong justification for new methods of evaluating permafrost extent using remote sensing platforms. This study quantifies forest loss at ten subarctic boreal sites in the southern Northwest Territories and northeastern British Columbia between 1970 and 2010. Historical air photos and optical remote sensing images were assessed using a change detection approach over the ten sites, each 10 km ^2 spanning a north/south transect of 200 km. This study is the first to apply change detection methods to a large-scale gradient and spans the southern margin of discontinuous permafrost where results demonstrate variable patterns of net forest loss at each site ranged from 6.9% to 11.6% over the 40 year study period. Here we show that these differential rates of landcover change can be explained in part through climatic and environmental factors that vary latitudinally across the selected sites. Change statistics—net change, forest gain and forest loss were significantly correlated with an assortment of factors that varied across the ten-site transect. |
|---|