Two researchers wearing long-sleeved shirts and pants to resist mosquitos, and high boots to block snake bites, gaze at a shattered tree. It lies on the ground garlanded in palm fronds, stretching far into the forest. Until recently, it had towered over most other trees in this vast rainforest.

“It’s lightning, obviously,” says Evan Gora, declaring the tree’s cause of death. He’s a staff scientist at the Cary Institute of Ecosystem Studies in Millbrook, New York.

“You can see burned leaves on the top,” agrees Adriane Esquivel Muelbert. She’s a professor at the University of Birmingham in England. Adriane points up at blackened foliage dangling from 20 trees circling the perimeter of the huge stump. The leaves are seared only on the sides facing the canopy gap left when the big tree toppled — evidence of an electrical strike.

Like Sherlock Holmes unraveling a murder mystery, the tropical forest ecology experts present their reasoning to two postdocs on their team. When lightning hits a tree, Evan says, high voltage flows through intertwined foliage into neighboring trees, killing branches, creating a distinctive pattern. Evan developed this method for determining lightning as a cause of tree death while working in a Panamanian rainforest. Today, he’s identified the same radiating pattern of scorched dead vegetation here around this fallen giant in the Brazilian Amazon. No other cause of tree death looks like this.

“You sometimes can have many trees standing dead together, but not with this centralized [burn] damage,” Adriane comments.

“It’s like a crime scene investigation,” Evan says, with growing enthusiasm.

It may seem odd for two high-powered scientists to spend so much time forensically investigating a single tree’s death in a vast forest. But the implications are weighty. Their research project, Gigante, is exploring the causes of mortality in the biggest trees of the world’s tropical forests. It could help answer a consequential question of climate change science: Will intact tropical forest continue soaking up far more carbon dioxide than it releases?

Intact regions of the Amazon are still storing considerable CO² and slowing the atmospheric buildup of the planet-warming gas humans release when they burn fossil fuels. But if the carbon uptake drops significantly in the Amazon and the world’s other tropical forests, global temperatures could rise faster than today’s already-dire models suggest, making it even harder for humanity to slow climate change.

Ready to fly

The two scientists, along with postdocs Vanessa Rubio and Gisele Biem, have assembled here, in Brazil’s Adolpho Ducke Forest Reserve, for the first time. Their three-year global research project has only recently begun.

The Ducke Reserve, on the outskirts of Manaus, in the central Amazon, covers a 93-square-kilometer (36-square-mile) quadrant of rolling, old-growth rainforest set aside by the Brazilian government for research. Visitors, such as this team and a varied cast of students and collaborators, sleep in tidy whitewashed dormitories and eat in a wall-less dining hall, grounds which they often share with peccaries, vultures, wildcats and jararacas (Bothrops jararaca) — among the world’s most venomous snakes.

The researchers have come to answer these crucial carbon uptake questions and also to hone observational skills, practice collecting and recording data and build esprit de corps. On the first day, the team has focused on identifying lightning strikes. The next day, the topic will be windthrow — trees toppled by wind.

Evan is eager to show off his latest research gadget. In a clearing by a low stucco building that serves as the reserve’s classroom, lab and headquarters, he unzips a huge cloth-covered suitcase.

Eager as a child unwrapping a Christmas gift, Evan lays the halves of the suitcase flat and lifts out a drone fuselage the size of a skateboard. He’s hand-carried this Quantum Systems Trinity Pro drone from Germany. The team watches as he assembles it.

“It’s the coolest toy ever!” Evan exclaims. Then, as if offering holiday turkey, he asks Vanessa, “Do you want to grab a wing?” Evan and Vanessa clip on the plastic and Styrofoam tail and the 1-meter (3-foot) wings.

The featherweight wings and twig-like legs make the drone look fragile. But Evan says it’s a serious research tool at an affordable price. This model will launch vertically, like a helicopter — useful in a forest. On a single charge, it can fly horizontally like a plane and autonomously for an hour and a half at nearly 64 kilometers (40 miles) an hour. Its high-resolution camera distinguishes objects as small as a silver dollar from 120-300 m (400-1,000 ft) up. Without it, the Gigante project couldn’t happen.

Tropical forest coroners

A study published in the journal Nature in 2015 stunned scientists. It found that intact Amazon forest absorbed 30 percent less carbon dioxide in the 2000s than in the 1990s. The authors suggested that the world’s tropical forest carbon uptake — the tropical carbon sink — is failing. Other studies have since confirmed that result and shown similar declines in tropical forests elsewhere.

“These forests are providing a huge benefit to society for free,” says Simon Lewis, a geographer at Leeds University in the U.K. and a co-author of several of these papers. Like many researchers, he agrees that climate change impacts are a primary cause of the drop in forest carbon uptake. If we don’t halt it soon, he says, “Forests can add to the [climate] problem rather than mitigating it.” Up until now, the Amazon forest absorbed about 12 percent of all the carbon released into the atmosphere by humanity, though the exact amount is a matter of debate.

One reason the tropical sink of intact forest is declining, according to many scientists, is that more trees are dying and/or dying younger. But researchers don’t know enough about why, and when, such trees die. So, it’s not possible to model accurately and predict how these factors will change in the future, creating uncertainty in climate forecasts.

Robust forecasts require estimates of forest carbon uptake. And without accurate climate predictions, people can’t accurately forecast the velocity and ensuing severity of the climate crisis.

Adriane and Evan hope to shed light on the life and death of the largest tropical trees, generally those with trunk diameters wider than that of a large pizza. That’s important because these trees account for a disproportionately high share of a tropical forest’s carbon uptake.

Researchers estimate that large trees soak up about half of the carbon a tropical forest absorbs. The future effectiveness of the tropical sink likely depends on the longevity of these individuals. If increased warming, reduced rainfall or other climate change impacts shorten their lives, the whole forest will become younger and absorb even less carbon than today. The tropical carbon sink could decline or vanish. And as tree death in intact forest intensifies, the world’s remaining tropical forests could even become significant carbon sources. The growth of average and extreme temperatures, the patterns of rainfall and the intensity of storms could meaningfully determine what happens to the large trees of tropical forests. But Adriane says, “Of big trees, we know almost nothing.” They know so little partly because such trees are rare and die infrequently. A 2018 study of a site near the Ducke Reserve found that, of 5,808 trees observed for one year, 67 died. Of these, only one tree was large. You can’t make inferences about how a population behaves by studying one tree. This problem is compounded in the hyper-diverse Amazon with more than 10,000 tree species, with multitudes of distinct life strategies. In order to analyze a sufficiently large set of big trees, scientists need to gather detailed information from more tropical land than has ever been examined. But studying the life history of a rainforest’s individual trees by current methods is painstaking and expensive. Typically, forest workers tag and record statistics such as trunk diameter and species (if known) for every tree within study plots about the size of soccer fields. Like census takers, these workers update their records on successive visits.

In a 2020 paper, Adriane identified 189 such plots in a network of Amazon research sites in a network called RainFor, which she’d deemed big enough, and revisited frequently enough, to include in a study of tree mortality. The plots’ combined area totaled 331 hectares (818 acres), about the same as New York’s Central Park. From this forest sample, she inferred the causes of mortality of an average Amazon tree. But she also concluded that the network’s data “lacked the spatial and temporal coverage to provide information about large trees.”

In other words, so far, too few large Amazon trees have been studied to determine how long they live and what kills them. Adriane says the situation is worse in the world’s other tropical forests in Asia and Africa.

That’s why the team’s budget included the $27,000 Trinity Pro drone. With it, the Gigante project can study more large trees than ever before. We’re “switching our approach from looking at the ground level, measuring trunks of trees, to using a drone,” Evan explains.

For their current work, they’ll monitor a 1,500-hectare (3,700-acre) plot inside the Ducke Reserve. That might not sound huge, but the area contains about 750,000 trees thicker than a fence post and four times more land than in all the small plots studied in Adriane’s 2020 paper.

In contrast to the RainFor plots, which fieldworkers visit once every couple of years, Gigante’s drone will survey the study area monthly. In addition, the Gigante team will repeatedly put on those snake-proof boots and walk into select sites. More like coroners than census takers, they’ll only visit newly dead big trees after their monthly analyses of drone images.

Before beginning this data gathering process, they need to launch the drone.

Red tape and a kidnapping attempt keep Gigante grounded

But there’s a hitch. At the Adolpho Reserve’s base camp, Evan dismantles the drone and fits the pieces back into its custom case. He can’t fly it yet. Brazil, like the United States, regulates drones. Despite months of trying, the team hasn’t yet gotten permission.

They’d expected approvals by now, but Evan says there are “some complications.” Partway through applying for permits, the consultant they’d hired to help with the paperwork abruptly stopped responding to texts and calls. After a while, he explained that there’d been an attempted kidnapping. He wouldn’t give further details, but said that he needed more time to get back to work. Weeks have passed without further communication, leaving permission to fly in limbo.

“We might hear from him anytime, or maybe never again,” Evan says. “You need to have a lot of patience,” he explains. “Things will go wrong … they always do.”

Once permitted, the team will fly the drone over Ducke. It will shuttle back and forth over a large rectangular plot, photographing the forest in parallel tracks, like the lanes of a meticulously mowed lawn. The researchers will knit together the pictures, producing a single rendering of the whole area. With help from a computer program developed by colleagues in Panama, they’ll scour this composite for openings in the canopy that appear since prior overflights, each the likely sign of one or more newly fallen tree.

Then they’ll hike out and check each one. Adriane calls each such visit a “necropsy.” The team expects 10-20 new dead tree sites will open in the canopy each month —about 500 trees each year, twice as many as their calculations show are needed to draw statistically significant conclusions about large-tree mortality.

What’s at stake

With $1.7 million in grants from the U.S. National Science Foundation and the U.K. Natural Environment Research Council, Adriane and Evan also oversee a parallel survey in Panama and, starting in the coming year, in Malaysia, Cameroon and a second Amazon site.

Each local crew will use identical methods for collecting data at those locations, following what the researchers call “the protocol,” to enable valid inter-site comparisons. With the fervor of true believers, Adriane, Evan and their acolytes, the postdocs, will teach the protocol to the teams that study the other sites across the tropics.

Since the drone can’t be flown today, and the Gigante crew can’t collect aerial data, team members pull on boots and spritz themselves with mosquito repellent. It’s time for today’s workshop on windthrow. The four scientists and a local forest expert hike several miles of rainforest trails to a gash in the canopy. It looks like a giant hand has swatted three tall, broad trees, knocking them to the ground and wrenching their roots out of the soil. These, in turn, have crushed dozens of smaller trees into a treacherous tangle of branches.

Evan has experienced firsthand the moment when a tree this size topples. “It’s a pretty spectacular sound!” he says. “You hear snapping as the roots are ripping out of the ground, as it’s crushing trees around it.” Imagine a stick an inch thick snapping in two, he says. “Now multiply that diameter by a few feet.”

Before the team can make sense of the jumble, a light rain erupts into a downpour. Soaking wet, the researchers string up a tarp and wait it out. Vanessa passes around a bag of Paçoquitas, bite-size Brazilian peanut candy. They chomp and sing pop songs and the storm rages.

Then the team measures each trunk’s diameter. Two qualify as “giants” — for this study, trees more than half a meter, or about 1.6 feet, across the trunk at “breast height.”

While Adriane and Evan watch approvingly, Vanessa and Gisele note that woody vines known as lianas cling to the overturned trees. On their data sheets, the postdocs rate the liana infestation a factor of two, indicating that vine foliage covers 25-50% of the combined canopy. These creepers block sunlight from reaching a tree’s crown and swipe water from its roots. Lianas sometimes become so heavy that they drag trees down.

The team observes that leaves of the downed trees still dangle, showing that the trees fell while alive. Moving to the base of the trees, they see that fine hairs on the roots are pristine. Such delicate structures degrade quickly in tropical heat and humidity when exposed to air. From this, they conclude that the trees fell within the past month or two.

If this weren’t just a training exercise, the team would also check the fallen trees for heart rot damage, probing into each with a tool called a Resistograph. The device looks like an assault weapon; holding it by a pistol grip, the researchers would press the tip of the barrel, which houses a needle, against a trunk and pull a trigger. A drop in resistance to the probe suggests punky wood caused by a fungal infection, another possible cause of tree death.

If a tree falls … what killed it?

Wind clearly blew down at least one of these big trees. Others might have been knocked down by a toppling neighbor. The postdocs mark “W,” for wind, in their notes. But did wind kill the trees? Wind does knock over Amazon trees frequently; half of all trees that die of natural causes are blown down by wind, according to Adriane. By “natural,” she means not killed with a chainsaw. But in fact, it’s getting harder all the time to draw a line between natural and anthropogenic tree death.

One recent study estimates that by 2100, greater storminess in the Amazon brought on by human-caused climate change will generate an increase of 43 percent in death by windthrows. And though scientists don’t know for certain, wind might have the most impact on the biggest trees, since their vulnerable tops rise well above the surrounding canopy, which slows wind.

The researchers are intent on determining whether one factor rather than another “causes” mortality. Evan explains that though wind is the likely immediate, or proximate, cause of these trees’ dying, it might not be what really killed them — the ultimate cause. Investigating a tree death is just as fraught forensically as finding a person dead at the bottom of a staircase: Did he die from a blow to the head while falling, or from the stroke that led to his fall?

A lightning strike is among the few diagnoses of dead rainforest trees made with a high degree of certainty. Evan discovered that 40 percent of the large dead trees that he’d studied in a Panamanian rainforest had died immediately after a lightning strike.

If lightning strikes increase by as much as some climate projections suggest (a rise, globally, of up to 50 percent by 2100), Gora and colleagues estimate that large-tree mortality in Panama could increase by between 9 percent and 18 percent. Such a change would in turn reduce forest carbon uptake. But Evan cautions that lightning might not play a similar role in every tropical forest. The Gigante research should help determine that role.

To investigate ultimate causes of mortality, the new Gigante protocol requires collecting information on multiple risk factors — conditions that could bring a tree to the brink of death before another factor strikes the final blow.

For instance, Evan explains, “We could conceivably find that every tree that died by wind has this massive load of lianas. It would come out that lianas are driving mortality, even though the proximate cause that we wrote down was wind.” Other risk factors to consider include heart rot, insect infestations and water stress (too much or too little), all which could be intensified by climate change.

Adriane says that because assigning an ultimate cause of tree death is difficult, their research will always carry some degree of uncertainty. “We never get to the cause of death. We get to the potential cause of death.” Still, she’s motivated by the urgent need to forecast the effectiveness of the tropical carbon sink in the coming decades.

Many changes in tropical forests happening now, or projected, could increase tree mortality and degrade the tropical carbon sink. Climate change is altering patterns of precipitation, extreme winds and lightning, while lianas are becoming more abundant in the Amazon and some other tropical forests as areas of human disturbance and intensifying heat increase.

Sorting out all these factors requires days of drudgery. At the end of this day, the team trudges single file back along the muddied and slippery trail to the Ducke headquarters. Glistening water droplets hanging off leaves refract miniature fisheye replicas of the dimming forest. The foliage exudes invisible sweet floral scents, oddly with hints of garlic.

Approaching base camp, they halt by a living big tree rising like a thunderhead above the surrounding canopy. A harpy eagle sits inside a stick nest big as a bathtub that she’s propped in a crook of the crown. Harpies (Harpia harpyja), the Amazon’s largest bird of prey, perch atop the trees and the food chain. The giant raptor swivels its head toward the researchers. Entranced, Adriane abandons her scientific dispassion. “This is why giant trees are important,” she exclaims.

Reporting for this story was supported by the Pendleton Mazer Family Fund and Abby Rockefeller and Lee Halprin.