Local and global canopy height models reveal consistent large-tree forest conservation gaps despite low spatial agreement

Canada's coastal temperate rainforests are home to some of the world's largest trees. These forests support disproportionate carbon storage, high structural complexity, and habitat for a wide range of species. Historical logging focused on the most productive stands, combined with the protection of less productive areas, has transformed much of this ecosystem into younger, more homogeneous forest landscapes. Effective conservation therefore requires identifying and protecting remaining stands that support exceptionally large trees. Medium- and high-resolution canopy height products offer a practical proxy for detecting these ecosystems. Here, we evaluated the ability of locally calibrated and global canopy height products to detect forests supporting the largest trees. We developed an approach for identifying large-tree forests from canopy height models and applied it across ∼2.7 million ha of forests on Vancouver Island, British Columbia, Canada. We identified approximately 133,000 ha of large-tree forests, with nearly half occurring within highly modified landscapes. Spatial agreement among local and global canopy height models was low (9–40%), indicating substantial disagreement in the mapped location of large-tree forests. Despite these differences, all models consistently revealed major conservation gaps, with 70–77% of large-tree forests occurring outside formal protected areas. The protected portion of large-tree forests was unevenly distributed, with fewer than 25% of protected areas containing more than 1 ha of this forest type. Together, these results show that large-tree forests are under-protected, fragmented, and vulnerable to human-caused loss. Our study provides an operational framework for identifying and monitoring these rare and vulnerable ecosystems at broad spatial scales.