Juliane Menezes stands atop an observation tower for a project that aims to measure the effects of enhanced carbon dioxide on the Amazon rainforest.João M. Rosa

Apart from the experts, few people realize that climate change could be worse. Every year, trees, shrubs, and every other kind of plant absorb 9 billion tons of CO2—one quarter of what we let loose from our tailpipes and smokestacks—and help slow the gas’s accumulation in the atmosphere. If not for the world’s photosynthesizers, the concentration of CO2 in the air, along with Earth’s temperature, would be rising much faster than it already is.

Our terrestrial plants do us “a fantastic favor” sponging up all that CO2, says Scott Denning, an atmospheric scientist at Colorado State University. That’s especially true in the tropics: By some estimates, the band of jungle that hugs the equator sucks up about half of the carbon absorbed on land. But in the coming years, tropical regions are projected to see steeply rising temperatures and, in some areas, increased drought. That will create less and less hospitable conditions for these crucial equatorial jungles. Nonetheless, published forecasts of the future of climate change, including the marquee results of the Intergovernmental Panel on Climate Change, take it for granted that the world’s forests will continue acting as a terrestrial carbon sink—an assumption that may be disastrously overoptimistic.

There is one slender thread of hope that tropical jungles will keep buffering the amount of CO2 building up in the air. Scientists have long known that plants grow better in CO2-enriched air, and some hypothesize that steadily rising amounts of atmospheric carbon might protect forests’ health despite the otherwise deteriorating circumstances. Most likely, such carbon fertilization has already boosted the productivity of forests all over the world. But many researchers who study tropical forests predict that whatever benefit trees are getting will wane. It may have already started to taper off.

The nightmare scenario is that a hostile blend of heat and drought—not to mention fires like the ones that destroyed more than 2 million acres of the Amazon and reddened Indonesia’s skies this summer—will turn tropical forests into a far less luxuriant habitat, wringing out the carbon sponge. According to Denning, if tropical forests absorb less carbon and release some of what’s already stockpiled in trunks and soil, it could raise global temperatures by roughly 2 degrees Fahrenheit this century. Such a catastrophe would wipe out the benefits of all the climate-change policies that any country so far has put in place or seriously contemplated.

Despite the potential for dramatic consequences, scientists still haven’t settled on what the future holds for these jungles. So I decided to see for myself how they’re trying to figure it out. I loaded up on tropical-disease vaccinations, picked a lightweight mosquito net, and bought an airplane ticket to Rwanda.

At a border crossing between Rwanda and the Democratic Republic of Congo in early 2017, I rendezvoused with Wannes Hubau, a Belgian geographer with the build of a wrestler, a brooding manner, and a taste for adventure. It was only weeks after Joseph Kabila, then the president of the DRC, had refused to step down at the end of his term. To avoid the violent backlash and safely reach our destination, Hubau and I had to slip in from Rwanda on foot, catch a taxi to the nearest airport in Goma, fly to the sleepy city of Kisangani, and drive to the Yangambi Biosphere Reserve, at the heart of the Congo rainforest.

The day after our journey, Hubau stared into the forest of huge buttressed trunks and trees perched on stilts as if summoning the will to hog-tie a spiny palm to the ground. Flies buzzed around his head; he swatted one, and it gave off a sickly sweet scent. He held a piece of paper marked Plot 08, clamped to a clipboard. It listed the previously measured size and location of 384 trees in the thicket encircling us. The painted blazes that marked the height where he hoped to remeasure each tree were gone, chewed off by ants. Identifying tags nailed too closely into the bark had been swallowed up by growth. “They always give me the hard sites,” he grumbled.

Hubau was in the Congo to document the effects of the 2015 El Niño on its rainforest. El Niños temporarily rearrange worldwide patterns of rainfall and temperature, and the 2015 event caused pockets of drought in Central Africa, mirroring the drier conditions that climate change might bring later this century. In 2010, when a swath of forest in the Amazon larger than Alaska and Texas combined suffered a withering drought, it killed more than a billion trees and liberated billions of tons of CO2 into the air. It briefly transformed the Amazon’s carbon sink to a carbon source. Simon Lewis, an ecology professor at Leeds University, wanted to know if the same fate would befall trees in the Congo rainforest.

Wannes Hubau measures the girth of a tree trunk in Yangambi Biosphere Reserve. (Daniel Grossman)

Lewis dispatched teams to seven countries across central Africa to census 100 plots. He sent Hubau to size up the Yangambi forest, and I tagged along. Lewis’s envoys had made measurements at these sites twice before. Each time, they’d recorded the girth of every tree—in Yangambi, more than 4,000 leafy individuals spread over 11 square plots, each the size of a soccer field. They’d used each trunk’s diameter to calculate its carbon content, and added up the total in each parcel. Lewis needed the new measurements to compare normal carbon uptake with that of the El Niño episode.

In a week with Hubau, I learned what should have been obvious before I left home: Tropical-forest surveys are hot, buggy, and mind-numbingly repetitive. That’s why he’d let me tag along: to see “how hard this is,” he admitted one day.

I didn’t learn what Hubau discovered on that field trip until he emailed me months later. He wrote that the trees in his plots absorbed less carbon during the El Niño. Amy Bennett, a graduate student in Lewis’s lab, found that across central Africa, the plots that are normally wettest experienced modest drought during the 2015 El Niño. Plots that are usually dry were largely unaffected. The Congo forest was also slightly warmer than average that year. “The interaction of drought and heating together is incredibly important to understand,” Lewis told me, because the future will bring some as-of-yet uncertain combination of the two.

Lewis said he can’t give me specific numbers until he publishes his results. But he let on that his data show that “the sink declined.” This finding is broadly consistent with results published in a 2017 paper in the journal Science, which used measurements collected by NASA satellites to estimate plant growth and fire-related vegetation loss in the Congo forest. The authors of that paper concluded that the forest absorbed about 3 billion fewer metric tons of CO2 during the 2015 El Niño than it does during a normal year. That’s about as much CO2 as the European Union released by burning fossil fuels over the same period.

A year before I met Hubau, I was in the Amazon, slogging up the steps of a rickety latticework tower. The aluminum lookout was at a site called ZF2, a study area operated by Brazil’s National Institute of Amazonian Research (INPA). Hardly visible in the dim dawn, David Lapola, a biology professor at Brazil’s University of Campinas, beckoned me upward. The thwap of our feet against metal treads drowned out the growls of distant howler monkeys. At sunrise, we reached an observation deck, overlooking the treetops from 10 stories above the ground.

The horizon glowed a fiery orange, but Lapola paid little heed. He was staring straight down, at a disk of experimental forest below us. If only he could get a big enough grant, he lamented, the trees down there could help him forecast the future of the tropical carbon sink.

Not long before, a large team led by a plant ecologist named Roel Brienen had published an ominous paper in the journal Nature. Over decades, the researchers had repeatedly measured the carbon content of 321 plots across the Amazon, using the same method that would later try Hubau’s patience in the Congo. They found that the amount of carbon dioxide absorbed by intact Amazon forest had declined by 30 percent between the 1990s and the 2000s. In a recent phone call, Brienen said that if the trend continues, “in two or three decades the carbon sink could be shut down.” Brienen’s paper speculated that increased CO2 might actually be causing trees to grow faster and die younger, like a campfire doused in gasoline.

In 2016, Lapola was one of two Brazilian scientists overseeing plans for what Richard Betts, a scientist at the U.K.’s government climate lab, called “one of the most exciting experiments on the planet.” The study is a free-air carbon enrichment (FACE) experiment, in which an open-air plot is treated with extra CO2. One of the earliest FACE studies in a forest, which began in 1996, boosted the CO2 concentration in a Tennessee sweetgum plantation by 50 percent—roughly matching the air we’re on track to breathe in 2050. At first, trees grew faster. But they soon stagnated. The soil ran out of nitrogen and the trees couldn’t benefit from extra CO2.

Top: Sabrina Garcia, a postdoc on AmazonFACE, prepares to measure a tree’s photosynthesis rate through its leaves. Bottom: Automated dendrometers measure the trunk diameter of trees in AmazonFACE, while diamond-shaped devices measure sap flow. (João M. Rosa)

Richard Norby, who ran the sweetgum experiment, says it’s hard to predict in any given case whether carbon fertilization will make a difference. Scientists have run FACE experiments in temperate forests in the U.S., Europe, and Australia, and sometimes the extra CO2 helps. Other times, a dearth of water, too much heat, or some other factor gets in the way. “That’s why we need more experiments,” he says.

During my visit to ZF2 in 2016, an international crew of researchers at the base of the observation tower scurried around the circular plot, monitoring each tree’s health like so many intensive-care nurses. Lapola boasted that the experimental forest and a nearby twin—the control—are the most studied jungle patches in the Amazon. I watched a grad student gather litter for sorting and weighing. Lapola fiddled with tubing on a study tree outfitted with an electronic thermometer, a growth gauge, and a sap-flow sensor. Norby, who’d flown in from Tennessee to help, tweaked a contraption that measures soil microbes’ CO2 emissions.

Lapola was eager to build the plumbing that would spray CO2 into the experimental site, but he was out of money. He needed $3 million for equipment and another $1.5 million for each year’s worth of CO2. He said the trial should run for 10 or more years to produce robust results.

When I visited the site again this past February, Lapola still hadn’t raised the money he needs. “You are visiting a more pessimistic David,” he said. This time, the plot was still. After three years, the commotion of setting up equipment and training its operators has mostly given way to routine checks on automated gear. Lapola fretted that he might have to make do with a much smaller—and cheaper—experiment that only tests saplings. The behavior of juvenile trees is important, he said, but they probably don’t respond to rising CO2 the same way adults do. Two months after my visit, frustrated by the dim funding prospects, Lapola dialed back his management role in the experiment.

During that February trip, Lapola and I again climbed up to the aluminum tower’s observation deck. A warm breeze caressed the canopy just beneath our perch, jiggling leaves tinged in countless hues of green. A pair of Cricrió birds sang a melodic duo. I recalled something Richard Norby had said on my first trip to ZF2, when I’d asked him what he thought enhanced CO2 would do to this forest. “Diddly,” he’d replied without hesitation. Central Amazon soils are notoriously low in phosphorus, he explained, which will probably prevent the forest from taking advantage of carbon dioxide fertilization. “I hope I’m wrong,” he added.

A rendering of a fully funded AmazonFACE project (Pedro Lorenzo and Rogério Lupo)

Unfortunately, new evidence suggests he’s exactly right. In August of this year, the team planning the Amazon FACE experiment released the results of a computer simulation of CO2 fertilization in a forest with phosphorous-deficient soils. The study, published in Nature Geoscience, concluded that extra CO2 will not protect such a jungle from a changing climate. The ecosystem, its authors wrote, “is less resilient to higher temperatures and changing rainfall patterns than previously thought.”

The fate of nations might rest on the uncertain behavior of the trees in these remote jungles. More scientific research could help to predict whether carbon-dioxide fertilization will safeguard the carbon sink, slowing the advance of climate change, or do “diddly.” But neither of the two countries best poised to conduct necessary studies—the United States, with its large science budget, and Brazil, which occupies the majority of the Amazon—seems likely to sponsor such research. The chief executives of both countries question the very existence of climate change, and both want to slash science funding. (So far, congressional support for science has frustrated proposed cuts in the U.S.) Jair Bolsonaro, the president of Brazil, has made clear that he supports more cutting, not preservation or research, in the Amazon.

If projects like AmazonFACE don’t get off the ground, we may be forced to simply wait to see what comes. If humanity keeps burning forests and pumping CO2 into the atmosphere, time will give us the verdict for free.

Support for this article was provided by a grant from the Pulitzer Center.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.