The early 2000s were a time of exceptional drought in the American Southwest. The year 2002 in particular was perhaps the driest in the last 500 years, according to tree-ring historical reconstructions.

This was bad news for the aspen trees of the Southwest, which died by the millions. But it also raised a scientific question: just how exactly does drought kill trees? A child easily grasps that lack of water will kill about any plant. But the specific biological mechanism by which trees die from drought has not been well-established. It's the difference between knowing that shooting someone in the chest will kill them, and understanding why a bullet puncturing the heart will end a life.

A team of scientists, led by Dr. William Anderegg of Princeton, have been working on the aspen question, and their results were published today in Nature Geoscience. The answer is something called "xylem cavitation." And unless we do something big about climate change soon, it will kill most of the aspens in the Southwest.

Here's what that means. Trees transport water through their xylem tissue (one example of which is regular old wood), basically composed of millions of tiny tubes, or "conduits." Xylem doesn't work like a mechanical pump — instead, water flows up the tree trunk through capillary action. That flow is maintained through evaporation at the leaf surface, removing water at the top so more can replace it, and supplied by the roots drawing water from the soil.

In hot, dry conditions, like the early 2000s drought, water evaporates more quickly from the leaf surface — and there is less water in the soil to maintain supply. Anderegg and his team quantified both of these with a factor they called "climatic water deficit." When the deficit is high, the water pressure inside the xylem decreases due to tension between the top and bottom of the tree.

If the pressure gets low enough, gas bubbles will spontaneously form in the water column — which is called cavitation. A bubble instantly blocks that particular xylem conduit and prevents the water from flowing. Block enough conduits, and the tree desiccates and dies.

It's "like a tree heart attack," says Anderegg. He and his team constructed a model of this cavitation mechanism, calculated a threshold at which aspens should die, and compared it with historical data on the early 2000s drought. They found pretty clear agreement, explaining about 75 percent of the tree mortality during that time (a good result, given how complex forests are). In this image, red and yellow represent when the model correctly predicted whether a tree would live or die, while green and blue are the corresponding wrong predictions:

Xylem cavitation has been understood for years, but this is strong evidence for this being the murder culprit, so to speak. (Note that this model only applies to deciduous tree species. Conifer species like lodgepole pine have also been killed en masse by climate change-fueled drought, but abnormal beetle swarms are what strike the killing blow.) Others had suggested different mechanisms, like starvation. What's more, this threshold model has some disturbing implications. All that is required to kill an aspen forest is a sufficiently hot and dry spell.

According to the big climate models, under a high-emission pathway (that is, assuming world society does little to combat climate change), large sections of current aspen forests will be consistently above the mortality threshold by the 2050s. But since all that is needed to kill an aspen tree is a couple of exceptional years, then the bulk of current aspen forests will likely be dead some time before that decade.

This matters not just for the aesthetic value of forests in themselves, but also for many human interests as well. Besides being a major tourist attraction — they're just about the only fall color in much of the Colorado mountains — aspen is a major commercially harvested species in the Southwest. Aspens also support a large variety of local wildlife, much of it important to local economies, and a variety of other ecosystem services (like water filtration).

Obliterating the aspens would not only be a great ecological crime, but also a terrific blow to local communities across the Southwest.

Therefore, the drought of the early 2000s was a "canary in the coal mine," says Anderegg. If we do nothing about climate change, then by 2050 the average year will be about like 2002 in terms of temperature and precipitation. The aspens, and everything that relies on them, will be dead.