Canada’s Forests Are Running Out of Breathing Room 

By Morgane Dendoncker, Christian Messier, Manuel Esperon-Rodriguez & Olivier Villemaire-Côté Published in Climate Change Ecology (2026), with the contribution of Madeleine Gauthier for communication

As temperatures climb and precipitation patterns shift, the trees growing across our 2.3 million km² of managed forest are increasingly pushed beyond the climatic conditions they evolved to handle. Our new study, published in Climate Change Ecology, offers a subcontinental-scale diagnosis of this stress, which can help the forest industry think about the vulnerabilities of the forests they manage. 

Are Canadian trees stressed right now? 

Most tools used to assess how forests will respond to climate change rely on species distribution models (SDMs) which are statistical approaches that predict where a species might live based on where it currently lives. These are useful, but they have a blind spot: they can’t tell you whether a species living in a given location today is already at its climate physiological limit. 

We took a different approach. Instead of asking ”where” species might live, we asked ”how much more change” the trees already in place can tolerate before they run into trouble. To do this, we used climate safety margins (CSMs) which are a measure of the gap between the climate a species is currently experiencing and its realized physiological tolerance limits. 

Think of it as your comfort zone at work. Everyone has a range where they perform well, comfortable enough to focus, resilient enough to handle a bad day. Push someone chronically past that range and wellness decreases, productivity drops, stress accumulates, and it takes less and less to tip them over. Trees are no different! 

What we found: a shift from cold stress to heat stress 

We analyzed 313 North American tree species across Canada’s managed forests, using all available species occurrence datasets for this region, combining Canada’s National Forest Inventory, provincial sampling plots, the U.S. Forest Inventory, and global records from GBIF. In our analyses, we focused on five climatic variables: mean annual temperature, the coldest month’s minimum temperature, the warmest month’s maximum temperature, annual precipitation, and mean monthly precipitation of the driest quarter, capturing both the thermal and water conditions that govern where trees can survive and grow. 

The results paint a clear picture of a system in transition.

Today, an important climatic factor for most Canadian tree species is cold: specifically, the minimum temperature of the coldest month. Given the geography of the Canada, nearly all species are living in the cold edge of their distribution somewhere in the country, particularly in the far north and at high elevations. This is expected; cold winters define range boundaries. 

By the end of the century, that pattern flips. Under intermediate and high emission scenarios, warming winters reduce cold stress, but rising summer temperatures become the new threat. By 2071–2100, up to 97% of Canadian tree species are projected to exceed their tolerance for summer heat (maximum temperature of the warmest month) in at least one part of their range. Forests aren’t just gaining climatically suitable space at their cold northern margins; they’re simultaneously losing it at their hot southern ones (see figure below).

This “cold-to-heat” transition is not unique to Canada; similar patterns have been documented in European and Asian forests. But for Canadian practitioners, the implication is concrete: the species you planted 30 years ago for a cold boreal climate may not be the right bet for the next 30.

The assisted migration puzzle and the cold bottleneck

One response to this challenge is assisted migration, which entails moving species or seed sources from warmer regions into areas where the future climate will suit them better. It’s a strategy gaining traction in forest management circles, but our results reveal a critical and underappreciated obstacle. 

Using climate analogues (places in Canada or the U.S. where today’s climate resembles what a target region will look like in the future), we identified a pool of candidate species that could theoretically grow under future climate conditions. On average, about 30 to 35 such candidate species exist for a given location (from Table 2). 

But here’s the catch: to establish, those candidates have to be in their comfort zone and survive today’s climate first, as seedlings and saplings, before they can match future conditions. And for many of them, the current cold winters in their target region represent a hard barrier, what we call the climate bottleneck. For now, and in the next decades, winter cold temperatures eliminate most candidates. This bottleneck is most severe in northern Manitoba, Saskatchewan, Ontario, and southwestern Québec. 

The upshot: after filtering for species that are safe under both baseline and future climates, only 4 to 13 species per location remain viable candidates for forest diversification, depending on emission scenario and how conservatively you define a species’ climate tolerance.

What this means for forest management

Our results don’t call for alarm, they call for precision. Here are a few practical takeaways: 

Negative safety margins are a warning, not a death sentence. A species operating slightly outside its comfort zone may still grow, but it might be less vigorous and more susceptible to secondary stressors like pests or drought-induced dieback. 

Not all regions face the same risk. Southwestern Ontario, Nova Scotia, and New Brunswick consistently emerge as areas where suitable species, both native and candidate migrants, are most available. Central and western Canada, particularly Alberta and British Columbia, face larger “no-analogue” climate zones where even climate analogue-based planning is limited. 

Assisted migration is promising but not a silver bullet. The climate bottleneck means introduction timing matters enormously. Species that will thrive in 2070 may not survive a harsh winter in 2030. Practitioners should consider microsite selection, shelter planting, and phased introduction strategies to bridge that gap.

The road ahead 

This study provides a broad-scale screening tool, a first filter for identifying which species deserve closer attention, and which regions face the greatest urgency. At the local scale, vulnerability assessments should be refined by taking into account factors such as topography, soil characteristics, genetic differences, seed origin, and forest management practices. These factors can help buffer species that are close to the limits of their climatic range and shape outcomes that are not fully reflected in continental scale analyses. 

But as a framework for prioritizing where to act and which species to consider, climate safety margins offer something most existing tools don’t: a measure of physiological risk, grounded in where trees grow across the globe. 

The full paper is available open-access in Climate Change Ecology: doi.org/10.1016/j.ecochg.2026.100114