This story has been updated.

In a troubling new study just out in Nature Climate Change, a group of researchers says that a warming climate could trigger a “massive” dieoff of coniferous trees, such as junipers and piñon pines, in the U.S. southwest sometime this century.

The study is based on both global and regional simulations — which show “consistent predictions of widespread mortality,” the paper says — and also an experiment on three large tree plots in New Mexico. The work was led by Nate McDowell of the Los Alamos National Laboratory who conducted the research along with 18 other authors from a diverse group of universities and federal agencies, including the U.S. Geological Survey.

“We have fairly consistent predictions of widespread loss of piñon pine and juniper in the southwest, sometime around 2050,” said McDowell. The paper concludes that the consequences could be vast, citing “profound impacts on carbon storage, climate forcing, and ecosystem services.”

The study examined both an extreme warming scenario — which recent climate policies suggest we may be able to avert — and also a more modest scenario that would likely bring temperatures above 2 degrees Celsius above pre-industrial levels by the year 2100, but not necessarily by that much. The more extreme scenario was certainly worse for these trees, but even under the moderate scenario, the negative results were merely “delayed by approximately one decade,” the study found.

The problem is that climate change is expected to not only increase the risk of drought, but will also drive heat up in general. And this could injure trees in two ways — simply drying them out, but also leading to “carbon starvation.” This could occur if, faced with dry conditions, tree leaves or needles close their stomata to keep water in, but therefore cannot bring in more carbon dioxide and thus suffer from reduced or even fully halted photosynthesis.

In the field experiment, conducted over five years, the researchers found that depriving trees of 48 percent of usual rainfall led to 80 percent mortality of piñon pines more than 100 years old – in other words, fully grown trees – and a 25 percent loss for junipers.

The scientists then proceeded to use regional and global models to examine whether they would reproduce this phenomena on a much larger scale in a warming world. The result of regional models was that 72 percent of the U.S. southwest’s needleleaf evergreen forests would “experience mortality by 2050, with nearly 100% mortality of Southwest USA forests by 2100.” Global models were not as grim, but when taken into account, produced the conclusion that “at least 50%” of these conifers “could be lost within the Northern Hemisphere by 2100.”

“Taken together, the validated regional predictions and the global simulations predict widespread conifer loss in coming decades under projected global warming,” the research concludes.

“This isn’t entirely a surprise to some of us who study this, the writing’s on the wall, so to speak,” said McDowell. “On the other hand, no one had ever evaluated these state of the art models for predicting tree death.”

What is perhaps most striking is that even hardy juniper trees were found to be threatened in the research. “They’re really really tough,” McDowell said. “The fact that they are predicted to have heavy mortality is really striking, because that’s sort of the ‘last soldier standing’ species.”

Taking it all into account, then, the study concludes that there is “a high likelihood that widespread mortality” of these types of forests will occur by 2100. So many trees could die and decompose, in fact, that the study suggests that 10 gigatons of carbon — equivalent to 36.67 gigatons of carbon dioxide — could be emitted to the atmosphere as a result, in forests across the globe. That would amounting to a positive feedback that would worsen human-caused climate change; indeed, the number is pretty similar to one year’s worth of the globe’s current fossil fuel emissions.

Overall, the news worsens a prior scientific outlook which had already suggested that U.S. forests, as a whole, could store less carbon in the future than they currently do — and thus, offset fewer U.S. emissions.

Granted, there are uncertainties in these projections, but McDowell says that while the models might overestimate tree death, they could also underestimate it. For instance, he pointedly observes that wildfires are not even included in the models in question — and that’s a major factor that could compound tree loss over the course of the century, especially when combined with drought.

William Anderegg of Princeton University, a researcher who also studies climate change and forests but was not part of the current study, also highlighted uncertainties in the work. While calling the study “an impressive first attempt that raises the point that we should be prepared for more drought stress and probably more mortality in southwestern forests,” he also noted by email that numerous complexities not accounted for in models could lead to tree mortality over-estimates.

“This is definitely an important work, but there are a lot of uncertainties still left to resolve before we can have confidence in regional mortality predictions,” Anderegg said.

Most striking, of course, is the study’s suggesting that even a lesser warming scenario, closer to what the world is now shooting for under the Paris agreement — limiting warming well below 2 degrees Celsius — but still likely beyond it, would still trigger these devastating consequences.

“2 C is huge for trees,” says McDowell.

More at Energy & Environment:

For more, you can sign up for our weekly newsletter here, and follow us on Twitter here.