There are few better places than Frijoles Mesa to study the mortality of trees. This tongue of land lies partly within the grounds of Los Alamos National Laboratory in northern New Mexico's Jemez Mountains. To the west rises Cerro Grande, a mountain riddled with the charred skeletons of fir and pine trees. To the southwest are the lingering scars of another fire, one so intense that its heat alone killed trees that weren't consumed by the flames themselves.

The mesa itself is an exceptionally tough place to be a tree, even where the land has escaped conflagration. This summer, many ponderosas were so short of water that their weakened limbs snapped like pretzel sticks. The trees that sit behind a padlocked gate off State Road 4 were also struggling. This is tree physiologist Nate McDowell's outdoor laboratory. Here, he's enclosed piñon and juniper trees in transparent silos, cranked up the heat and deprived many of water – in order to watch them die.

McDowell spent his early career studying the towering conifers of his native Pacific Northwest and came to Los Alamos in 2003, eager to begin a U.S. Department of Energy job that would allow him to set his own research agenda. But looking out his office window at New Mexico's characteristic piñon-juniper woodlands, he had second thoughts. "This is not a forest," he scoffed. The stout, pear-shaped junipers – one of the most common species here – resembled ill-kept hedges more than trees, all arms and twisted torsos, barely showing any leg. "They were like a weed to me," he remembers.

Like weeds, junipers are durable. Those outside McDowell's window were still green, but the piñon around them were dead. During the deep drought of 2002 and 2003, piñon died throughout the Southwest in historic numbers. Had the Old Testament told stories of forest die-off, as it did of floods, the carnage around Los Alamos would have been called "biblical": More than 90 percent of the area's piñon succumbed. "What a bummer," McDowell sulked. "I'm a tree physiologist, and the trees are all dead. What am I gonna do?"

At first, the cause of the trees' demise seemed obvious. The punishing drought badly weakened them, and when beetles bored through their bark, the trees couldn't muster enough sap to pitch them out. Once inside, the beetles mated, multiplied, dug tiny tunnels and spread a fungus that cut off the flow of water and nutrients, killing the tree.

But Dave Breshears, a University of Arizona professor and arid lands ecologist who had studied the woodlands for years, suspected that the truth was more complicated. During the 1950s drought, tree death seemed less extensive, even though that drought was longer and drier than the more recent one. What was different about this drought was temperature: It was a degree or two hotter.

Breshears' observations inspired McDowell to take a second look at the struggling forest. It's common knowledge that trees die during and after a drought, McDowell says, but "nobody can predict where it will happen, when it will happen, what trees it will happen to. That means we don't understand it. That was exciting to me – there's a science question there."

Why do trees die? It's a deceptively simple question in urgent need of answers: Trees are dying at alarming rates not only in the Southwest but in Colorado, the Northern Rockies, Alaska and elsewhere. This summer in northern New Mexico, even junipers began to expire in droves.

It might seem surprising that, in 2013, we don't know how trees die. We understand tree growth so well that we can decipher its code – tree rings – and reconstruct droughts thousands of years in the past. So why is tree mortality such a mystery?

"There has been a long tradition in plant science where, if your plant died during your experiment, you were bummed out," McDowell explains. "It was like, 'Ugh, we've gotta start over.' The question was never, 'Why did it die?' " Besides, he adds, tree death didn't seem particularly pressing. "I think people inherently look at trees as these stable things in our lives, like mountains. We didn't know there was a problem."

Western forests are confronting new versions of familiar foes. In the 1990s, a series of warm winters and summers in south-central Alaska allowed bark beetle populations to explode and kill millions of old spruce trees. Beetles gained similar strength in the Rockies during mild winters in the late '90s and early 2000s, killing not only their usual victims but also entire hillsides of ancient whitebark pines, which live at altitudes once too frigid to support the insects.

Farther south, piñons were also attacked, but by a beetle that, unlike its fellows in the Rockies, typically preys only on the weak. Here, scientists believed the industrious insects were less the cause of death than the final straw: a strong shove to trees with one foot already dangling over the cliff.