Gertrudes Freire and her family came to the great forest in search of land and rain. They found both in abundance on that day half a century ago, but the green wilds of the southwestern Amazon would prove tough to tame.

When they reached the settlement of Ouro Preto do Oeste in 1971, it was little more than a lonely rubber-tapper outpost hugging the single main road that ran through the jungle like a red dust scar.

Sitting on the porch of the family farmhouse in the sweltering heat of the Amazon dry season, Gertrudes, now 79 with neat gray hair tucked behind her ears and a smile that shows half a dozen stubborn teeth, recalls the hardship and hope.

Her children remember the fear. Fear of forest jaguars, indigenous tribes and the mythological Curupira: a creature with backward-turned feet who misleads unwelcome visitors to leave them lost among the trees.

The family carved their home from the forest. They built their walls from the tough trunks of the cashapona tree and thatched a leaky roof from the broad palms of the babassu. There was no electricity, and some days the only food was foraged Brazil nuts. At night, in hungry darkness they would listen to the cascading rain. Life was damp.

Until it wasn’t.

Near the Freire home, there was a stream so wide that the children – aged between 5 and 12 when they arrived – would dare each other to reach the other side. They called it Jaguar’s Creek. Now it’s not a meter wide and can be cleared with a single step.

The loss of such streams, and the wider water problems they are a part of, fill scientists with foreboding.

Covering an area roughly the size of the contiguous United States and accounting for more than half of the world’s rainforest, the Amazon exerts power over the carbon cycle like no other terrestrial ecosystem on Earth. The tree loss from an extremely dry year in 2005, for example, released an additional quantity of carbon dioxide into the atmosphere equivalent to the annual emissions of Europe and Japan combined, according to a 2009 study published in Science magazine.

As more and more of the forest is cut down, researchers say the loss of canopy risks hitting a limit – a tipping point – after which the forest and local climate will have changed so radically as to trigger the death of the Amazon as rainforest. In its place would grow a shorter, drier forest or savannah.

The consequences for biodiversity and climate change would be devastating, extinguishing thousands of species and releasing such a colossal quantity of carbon dioxide into the atmosphere that it would sabotage attempts to limit global climate change.

The Amazon tipping point would mark a final shift in the rainforest’s ability to sustain itself, an inflection point after which the trees can no longer feed traversing clouds with enough moisture to create the quantities of rain required to survive.

Climate models have foreseen other so-called tipping points disrupting Earth’s long-balanced systems, for example warming that causes Siberian permafrost to thaw and release huge amounts of emissions, or Greenland’s Ice Sheet melting at such a rate that annual snowfall can no longer make up for the loss.

Exactly where that point is in the Amazon, science is not yet decided. Some researchers argue that current modeling isn’t sophisticated enough to predict such a moment at all. But evidence is mounting that in certain areas, localized iterations of the tipping point may already be happening.

Reuters has tracked three extended experiences of the Amazon to give a real-world view of degradation once only predicted by computer simulations.

A family that has farmed this once-lush part of rainforest for almost 50 years. A scientist couple who have monitored thousands of individual trees for decades. And an atmospheric chemist who has collected air samples from far above the canopy for years. Their perspectives reveal the long-term impact of deforestation: on rainfall, on the remaining forest and on global emissions. Taken together, they show the dangerous extent of the changes wrought on the world’s largest rainforest, and a possible glimpse of things to come.

Even as science learns more about the far-reaching impact of destruction that began many years or even decades ago, deforestation has surged under President Jair Bolsonaro, who supports further opening the Amazon for mining and agriculture. Last year, an area larger than Lebanon was cut from the rainforest, and though preliminary data for 2021 points to a slight year-on-year decline, deforestation remains at a level not seen in Brazil since 2008.

Ecologist Paulo Brando, one of the leading scientists studying the changing health of the Amazon rainforest, sums it up: “There’s a limit to how much shit the system can take.”

LAND OF THIRST

Year after year, the Freire family hacked and sawed farther into their patch of forest on Brazil’s western frontier.

In 1976, after clearing a couple of hectares and getting permission to use some of their neighbor’s pasture too, they invested in 10 heifer calves and a bull – the start of a dairy business that would over the years grow into a successful herd of about 400 head.

But a fear of drought haunted their work. They had come from the Vale do Jequitinhonha, 2,500 kilometers (1,553 miles) to the east, where decades of slash-and-burn agriculture had dried and degraded the land, plunging its people into poverty. The semi-arid strip became notorious as the “Valley of Misery.” Even while water was plentiful, they sensed the same could happen in their new home.

Soil erosion, like that which plagued the Vale do Jequitinhonha, often follows rapid and chaotic agricultural expansion. Land stripped of native vegetation, especially when transformed into pasture and pounded hard by grazing cattle, loses ability to retain water in soil and foliage. Rain runs off the altered surface in sudden surges, dragging topsoil into streams and rivers that then clog and dry.

Brazil is blessed with the largest freshwater reserves in the world. But the relentless rise of one of the world’s agricultural powerhouses combined with changes in global climate are helping to drive a loss of this vital resource. Data released this year by MapBiomas, a collaboration between universities, nonprofit groups and technology companies, found Brazil lost 15% of its surface water in the three decades prior to 2020.

For the Freires, the last bits of doubt about the drying of the land seeped away on a parched day in 1991. A cowhand told Gertrudes the cattle were so thirsty, they were nuzzling the bottom of dried-out springs, sucking the sand in search of moisture.

She acted swiftly and put in a complex system of pipes and pumps to draw water for the cattle from springs that had not yet gone dry.

Controversially, she began reforesting too. Gertrudes had little idea of what she was doing but trusted her instincts, sharpened by years of drought in the homeland she’d abandoned.

Her neighbors – and husband – thought she was crazy as she planted trees around water sources and along streams and vowed that the last remaining patch of virgin forest, at the far end of the property, should remain intact.

Her words weren’t always heeded. “I came back from one short trip away and my husband had cleared another patch” for pasture, she remembers, shaking her head.

Gertrudes sensed that rainfall was changing too.

Several scientific studies have found the same. Because tropical forests influence rainfall, deforestation can change their pattern. One influential 2011 paper looking at 30 years of precipitation data found that the onset of rains in Rondonia state, where the family lives, had been delayed by up to 18 days.

Research since then has backed up this trend. A major report this year, which brought together around 200 scientists, said available data pointed to a dry season that “has expanded by about one month in the southern Amazon region since the middle 1970’s.”

Antonio Deuseminio, an agroecologist with decades of experience in the rainforest, is helping farmers replant trees and bring water back to their properties. He works for a subdivision of the Ministry of Agriculture focused on cacao, which he says has the oldest weather data in the area. Although total rainfall hasn’t changed significantly in Ouro Preto do Oeste, the dry season has gotten longer and drier, Deuseminio says. For agriculture this is a serious problem, because crops and grasses don’t have long enough roots to find water when there’s no rain.

The drier climate makes reforesting harder too. Twenty years ago, rainforest species could be planted straight into the bare soil. Deuseminio says he must now first plant drought-resistant trees, and only once these have grown enough to provide shade and improve the soil, after five years or so, can he follow up with classic Amazon species. Rainforest saplings now struggle to survive, he says, in this part of the Amazon.

DEATH ON THE BORDER

Decades of farming have made the Freires sensitive to the changing rains. But to the untrained eye, the slow shifts in surviving forests – like the one at the end of the family’s farm – are harder to see. Detecting these changes can require years of methodical study, sweaty painstaking work, with tape measures and walking boots and notebooks.

Ecologists Ben Hur Marimon Jr. and Beatriz Marimon have spent so long in their forest plots that they’ve befriended many of the trees. They’re saddened by those they have lost over the years. Of late, they are losing more and more.

The couple conduct research at the local campus of the Mato Grosso State University in Nova Xavantina, a soy town of 20,000 people located about 1,200 kilometers east from the Freire farmhouse. The surrounding area is a biome borderland, an in-between space where the Cerrado savannah transitions into the Amazon rainforest. The trees that remain, they say, offer a vision of the future.

“This is tomorrow, today,” Beatriz says, crunching through a dry patch of forest on the edge of town. Ben Hur finishes the thought. “This is the border of the Amazon, its protective wall, and it’s dying.”

If the tipping point marks the irreversible march of savannah over rainforest, scientists predict the process would first occur in forests where savannah and rainforest are already intertwined.

Ben Hur is 58, with a neatly trimmed white beard and frayed walking boots – a chewed victim of their dog. Beatriz, 55, has long gray hair tied back in a practical ponytail. The couple met in the 1980s while studying forest engineering in the state capital, Cuiaba. The two have basically worked together ever since. “He likes to talk; I like to do,” jokes Beatriz.

To monitor the forests, the couple tag trees of varying sizes and species across their plots with bits of metal that look like military dog tags. They return at regular intervals – anything from three months to three years – and measure tree circumference, height and carbon dioxide respiration. Trees that haven’t made it are added to a list of the dead.

Rainforests recycle vast quantities of water by returning rainfall to the sky through soil evaporation and plant transpiration, by which water absorbed in the roots is released via a plant’s leaves. In the Amazon, moisture that comes off the Atlantic Ocean is transported for thousands of miles across the South American continent, falling as rain and rising again as vapor as many as seven times until it hits the mountain wall of the Andes. On hot days, after a downpour, the forest can look like it’s steaming.

But large-scale deforestation disrupts this process, reducing the number of trees to such an extent that precipitation levels fall or become more concentrated over a shorter wet season. In some parts of wide-ranging Nova Xavantina over the past 30 years, Ben Hur says, rainfall has fallen by as much as 30%.

As precipitation changes, streams and sources disappear, and the remaining forest turns drier. Local temperatures also increase – particularly on edges where forest and farmland meet. Those vast flat agricultural clearings increase the strength of winds, which can rip through woodland and tear down the tallest, oldest trees.

The drier forest is also more vulnerable to fire, which is still widely used for clearing farmland here. As more trees die – from wind, drought and fire – their deaths increase the likelihood of such extreme weather in the future, creating a deadly feedback loop.

Early experiments that mimicked extreme drought in the Amazon had led scientists to think the drier climate would kill older trees first, but what Ben Hur and Beatriz have found is the opposite. With longer roots, the largest trees are usually the most resilient – at least to drought. Instead, says Ben Hur, pointing to the brown leaves of a nearby plant, it’s the saplings that die. The forest loses its future.

For Ben Hur and Beatriz, the degrading forests around Nova Xavantina demonstrate that the tipping point may already be happening there on a local level. The major question remains whether this same process could occur on a huge scale over entire swaths of the Amazon basin – and if so, when?

Celebrated Brazilian climatologist Carlos Nobre, who has helped popularize the idea of the tipping point over the past decade, puts the precipice at between 20% and 25% deforestation of the original Amazon canopy. We are currently at about 17%, according to the major report with 200 scientists published this year. Nobre believes we could see mass dieback across eastern, southern and central Amazonia within as little as 15 years.

Others aren’t so sure.

Marina Hirota, an earth system scientist who worked on models before switching to field work, says current simulations oversimplify the diverse vegetation, soil type and topography found across the Amazon basin. In her view, there’s not yet enough evidence to say where the tipping point is or even if such a single threshold exists for sure. The models need to be improved first, she says.

Hirota considers it more likely that deforestation would trigger multiple smaller tipping points in different locations across the Amazon, similar to what Ben Hur and Beatriz have seen in Nova Xavantina.

But many scientists think putting a single number on the tipping point is still important as a clarion call, even if it’s too complex to currently prove. Once you’re able to prove it, ecologist Brando argues, it will already be too late.

“We know there’s a cliff out there, and so even if we’re not exactly sure where it is, we need to slow down,” Brando says. “Instead, we’re rushing towards it with our eyes closed.”

VIEW FROM ABOVE

In the decades that Ben Hur and Beatriz were listing trees and wrapping them with tape measures, the atmospheric chemist Luciana Gatti was mastering how to catch carbon dioxide from the skies.

While the view from the ground found trees struggling under hotter and drier temperatures, Gatti wanted to understand what these changes meant for the Amazon’s role in global climate change.

Gatti, 61, first specialized in reactive gases and began her career at Brazil’s Nuclear and Energy Research Institute. After the 1992 Earth Summit in Rio de Janeiro, she turned to the Amazon, joining a crop of home-bred scientific talent striving for a bigger Brazilian role in global climate research. She now works for INPE, Brazil’s space research agency, where her narrow office is cluttered with family photographs and lanyards from past conferences. On her desk sits a stress ball shaped like planet Earth.

Since the Industrial Revolution, scientists estimate that roughly a quarter of all fossil fuel emissions have been absorbed by forests and other land vegetation and soils, chief among them the Amazon.

Through the 1980s and 1990s, as mass human migration to the Amazon was just beginning, the rainforest drew down some 500 million tons of carbon from the atmosphere every year, more than the current annual emissions of Germany, Britain, Italy and France combined. Photosynthesis by the forests’ billions of trees, using carbon dioxide to live and grow, served as a vital buffer against climate change.

As migration increased and more of the Amazon was cleared for agriculture, scientists knew the forest’s ability to suck in carbon would be hit. But no one knew quite how much.

To try and get an answer, Gatti squeezed into a roaring single-engine four-seater plane armed with a padded suitcase packed with glass flasks. From up over the canopy, she could sometimes see the scale of destruction, the gray smoke billowing from burning trees and the yellow patches of earth shorn of the forest green.

Gatti’s earliest air samples date back to 2000, from a single point in the eastern Amazon. But she found the data too narrow and volatile to give a picture of the carbon balance for the whole basin, so over the following years she expanded the work, training teams and contracting light aircraft to fill flasks of forest air from four parts of the Amazon: Santarem and Alta Floresta in the east and Tefe and Rio Branco in the west.

Since then, the aircraft have taken more than 600 vertical profiles – a series of samples taken at different altitudes over a given spot. At one point Gatti doubted her results. She grew depressed. The data didn’t make sense. It couldn’t be true. It showed the southeastern Amazon was releasing more carbon that it was absorbing, even in rainy years when scientists had expected the forest to be in better health. It meant a part of the rainforest was no longer helping to slow climate change, but adding to the emissions driving it.

She changed her methodology. Changed it again. And again. In total, she went through seven methodologies before eventually accepting what had seemed impossible. The southeastern Amazon is not only a net producer of carbon, but even when you strip out the fires, the forest alone – or the non-fire net biome exchange – is a carbon source. Scientists widely regard the results, recently published in Nature, as the most definitive so far on the changing carbon fluxes of the rainforest.

The western part of the Amazon, protected by its remoteness, is in better health and can still absorb substantial amounts of carbon, the study shows. But it’s not enough to compensate for the polluting east, where ranching and soy farming have cut deep into the rainforest. The so-called lungs of the Earth are coughing up smoke. “We are losing the southeastern part of the forest,” Gatti says.

Gatti thinks her numbers show that certain parts of the Amazon may already be at their tipping point. She believes the data points to the same process that Ben Hur and Beatriz have witnessed, but on a greater scale: Rainforest species such as the brazil nut and the ironwood giving way to trees like mabea fistulifera and ouratea discophora that are more tolerant of the drier, hotter climate. Such regime change releases huge quantities of carbon and would help explain the forest’s flagging ability to draw down emissions.

“It is a path without return,” Gatti says.

WATER A CONSTANT WORRY

Back in Ouro Preto do Oeste, the Freires bemoan the driest dry season any of them can remember. It is mid-August, and the first rains used to come by now, they say. The dry season, once just three months, now stretches for four or five. Across the whole country, reservoirs are dangerously low as Brazil suffers one of its worst droughts in a century.

The family is diversifying to try and shield their business from drought, building out capacity in breeding and beef cattle to complement their milk production. They’ve also started an organic soap business and want to plant corn.

Water is a constant worry. Some nearby farmers have already sold their land – mostly to larger cattle ranchers who address the problem by digging deep wells or piping water over long distances.

“It’s going to get even drier,” says Gertrudes, looking out over her farm’s yellow grass as two cats laze comatose in the stifling afternoon heat. In the distance, smoke hazes the horizon as newly slashed forest burns. “The water will finish.”

(Reporting by Stephen Eisenhammer; editing by Kari Howard)