in English

The world without the Amazon

A climate model predicts the effects of turning the forest into cattle pastures: 25% less rain in Brazil and higher temperatures, with “catastrophic” consequences for agriculture and energy production

João Moreira Salles e Bernardo Esteves

Translated by Flora Thomson-DeVeaux

 versão em português

If the entire Amazon rainforest were razed and turned into cattle pastures, average rainfall in Brazil would go down 25%. That conclusion comes from a study by American scientists, who ran multiple models to simulate the effects on the global climate if the Amazon were to be partially or entirely turned into crops. That drop in precipitation would come on top of the decline in rainfall nationwide already predicted as a result of global warming.

The results of these climate models provide predictive ammunition for the argument that Brazilian agribusiness depends on the rain that the Amazon creates. The forest releases water vapor into the atmosphere, which is carried across the continent on so-called “flying rivers.” Without trees to fuel them, these invisible waterways may dry up.

The simulation demonstrated that, beyond the Amazon itself, the Cerrado regions of the center-west and south – where the bulk of the country’s crops are grown – will be the hardest hit by the reduced rainfall. Brazilian agribusiness relies on rain: only 6% of cultivated land is irrigated. If it fails to curb deforestation in the Amazon, Brazil will struggle to hang on to its status as a global agricultural powerhouse. Rain, however, will not be the only issue. “The increase in temperature will be catastrophic for Brazil,” said Stephen Pacala, the lead author of the study.

Nearly 20% of the Amazonian rainforest’s original footprint has been cut down, of which approximately two-thirds has been turned into cattle pastures. Brazil’s National Institute for Space Research (INPE, in Portuguese) has warned that the territory deforested between January and September of this year, the first nine months of Jair Bolsonaro’s administration, is up 93% from the same period in 2018 – the year in which the annual deforestation rate hit its highest level in a decade.

The Idea

In May 2019, three Brazilians walked into Stephen Pacala’s office in Princeton. Pacala, a professor of ecology and one of the world’s leading specialists in climate change, was standing in a corner of the room making coffee. The three visitors – ecologist Adalberto Veríssimo, forest engineer Tasso Azevedo, and anthropologist João Biehl – had come to invite him to take part in a conference on the Amazon that would be held at Princeton in five months. At the event, they would discuss the future of the forest.

Pacala, his mind apparently elsewhere, kept on fiddling with the coffee machine as Adalberto Veríssimo began talking about the situation in the Amazon. These were difficult times for the region, and the political scenario was nothing short of disastrous – the forest was under attack. There were reports of violence, illegal seizure of public lands, and fires being set. Pacala’s interest was piqued. The coffee forgotten, he sat down, blackboards filled with equations behind him. What the Brazilians were saying was that if this destruction persisted, the forest might collapse and disappear in a matter of decades.

“There’s something most people aren’t aware of,” said the American, addressing his three colleagues and referring to the Paris Agreement. “When you say that to keep temperature increases between 1.5 and 2 degrees Celsius, developed nations need to get to net zero emissions by 2050 and the rest of the world by 2100, you’re assuming that natural carbon sinks” – like the Amazon – “will continue to exist.”

If they disappeared, he explained, cutting emissions drastically wouldn’t be enough. That meant that any hope of meeting the Paris goals rested on ending deforestation and allowing the natural absorption of CO2 to continue. “I think you could show that by deforesting the Amazon, destroying that carbon sink and causing greenhouse gas emissions through forest fires, Brazil alone would probably make the Paris Agreement globally impossible to achieve.”

The Brazilian scientists exchanged looks. This was new. “Interesting,” said Azevedo. “So here’s one thing we could do. We could imagine that Brazilian environmental legislation is altered to allow for another 20% of the Amazon to be cut down. Then we can calculate the effect on the climate.” Pacala, increasingly worked up, shook his head. “Go draconian. Eliminate the Amazon! Turn it all into soybeans. Then you can calculate the value of this resource we’re trying to protect. We get rid of a substantial fraction of the world’s biodiversity and we destroy indigenous peoples’ land. And we get soybeans. What’s the value of that? How important is it for the future of humanity in global warming? Well, let’s turn it into soybeans and find out.”

It would be hard to think of a better place in the world to carry out this thought experiment. Princeton is home to the NOAA Geophysical Fluid Dynamics Laboratory (GFDL), an advanced center for climate modeling. “There’s a hundred-odd PhDs working there,” Pacala explained. Climate and weather models were invented at GFDL back in the 1960s. “The models that predict hurricanes and everything were designed here, just about. It’d be a very simple matter to adapt them for a different scenario – whack the Amazon and see what happens.”

At the end of their meeting, Pacala warned that the model could only be run if he could convince one of his colleagues – the Russian-born Elena Shevliakova, a climate scientist at GFDL– to drop everything she was doing and work on it.“She’s always very busy,” he explained, “and she’s very serious. But if I ask her nicely…”

Two weeks later, Biehl, Veríssimo, and Azevedo were walking through campus when they spotted Pacala heading their way on a bicycle. He dismounted and delivered the news with a grin: “Elena’s in. We’re going to run the model.”

The results were presented on October 17th, 2019 at 4 p.m., during the conference on the Amazon put on by João Biehl and his colleagues at Princeton.

The Model

To run the model, Pacala and Shevliakova broke the globe down into small analytical units to understand how the climate would shift in each of them. On the continents, they applied a grid where the units were 1° by 1° (latitude and longitude), corresponding to squares of 110 km by 110 km. The oceans were divided into smaller blocks, 55 km by 55 km.

The world on a grid: based on a model that divided the world into small blocks, a supercomputer analyzed the effects of climate change in each quadrant.


The model accounts for the specific conditions in each of these quadrants. If one cell represents a slice of the ocean, it will calculate for currents, water temperature, and the geochemical interactions of that particular region; if it represents land, it will consider methane emissions (from agriculture and animals), carbon from fires, evapotranspiration from plants, carbon aerosols (particles suspended in the air, produced by fires), mineral aerosols (sand particles blown by the wind), condensation, and evaporation. Using these data, the model goes back to the preindustrial period, in the mid 19th century, and calculates what has happened and will continue to happen to temperature and carbon from that date through 2050. Each round of calculations steps forward thirty minutes. And so, in half-hour intervals, the model advances to the mid-21st century. Eight years of climate transformations, divided into half-hour chunks, covering every square on the grid – from the Amazon to Siberia, from the Horn of Africa to an empty spot in the South Pacific – meant a full day’s worth of processing from the supercomputer operated by Shevliakova. The calculations required 3,456 processors for the land regions (the smaller ocean quadrants took just half that).

Pacala and Shevliakova ran the model on two scenarios foreseen by the Paris Agreement. The first, more optimistic set of parameters, dubbed SSP1 – for Shared Socio-Economic Pathways  – assumes that the world will take the route of sustainability. The pessimistic scenario, SSP5, imagines that we will continue emitting greenhouse gases at current rates through 2050.

In both SSP1 and SSP5, as envisioned in Paris, the Amazon still exists. When the scenarios were developed, there was no reason to imagine that the forest would face an existential threat. Pacala and Shevliakova’s exercise, meanwhile, includes three situations: the SSP1 world with no Amazonian rainforest whatsoever, and SSP1 and SSP5 without half of the Amazon. The researchers ultimately decided to simulate the Amazon being turned into cattle pastures, not soybean plantations, since most deforested land is used to raise livestock.

In terms of temperature and rain, what would that do to Brazil and the world?


The Results

The results are dire.

If the Amazon disappears altogether, even in the rosier SSP1 scenario in which the world is able to slash its carbon emissions, average temperatures worldwide will rise 0.25°C beyond the expected increase for SSP1, which is between 1.5° and 2.5°C. Just thirty years stand between us and 2050. If the model went out as far as 2100, we would probably see a larger increase in temperature.

Crucially, these are average temperatures, derived from a calculation of a vast variety of temperatures across the globe. It so happens that people do not live in average places, but specific ones. What matters for their lives is the local temperature. In the Amazonian region, for example, the model indicates that completely eliminating the forest would lead to an additional rise of up to 2.5°C, above the 1.5° to 2.5° already predicted in SSP1. The region would get up to 4.5°C hotter, making it practically uninhabitable. Nearly all of South America would be affected. Only parts of Patagonia and the Andes would remain relatively unscathed.

In SSP5 – the “business-as-usual” model in which we do not alter our current rate of carbon emissions – eliminating just 50% of the rainforest is nearly as disastrous as eliminating it altogether in the rosier scenario. Parts of northern Brazil would see a 2.0 to 2.5-degree increase on top of the 4°C to 5°C produced by an SSP5 scenario with an intact forest. Center-western Brazil, meanwhile, would see the heat go up 1 degree.

The red areas on the map above will see the largest temperature increases. The first scenario depicts the world with lower carbon emissions (SSP1) and no Amazon rainforest; the second models the world with lower carbon emissions (SSP1) and only 50% of the Amazon; and the third shows the world if current carbon emissions continue and 50% of the Amazon is destroyed. The map below shows variations in precipitation under the same three scenarios, with red indicating areas where rainfall volume will decline.

The effect on rainfall is equally catastrophic. If the Amazon goes, even in the optimistic SSP1 scenario, the state of Goiás and the northern regions of the state of Mato Grosso and Bahia, as well as much of the Brazilian southeast – all areas of intense agribusiness activity – would lose between 0.6 and 1.8 millimeters of rain per day. In the heart of the forest, that drop may be as much as 2 millimeters, or nearly 30% of annual precipitation. On average, it will rain 25% less in Brazil.

On a global scale, in the SSP1 scenario with complete deforestation of the Amazon, we see a significant rise in temperature in the American Midwest, a significant but milder increase in Europe and Oceania, and a sharp rise of nearly 2 degrees in the Arctic. This will mean accelerated melting of the polar ice cap, and quite probably a radical shift in wind patterns. The destruction of the forest will, in a single sweep, cause an increase of 30 parts per million in the atmospheric concentration of carbon dioxide, the equivalent of fifteen years of burning fossil fuels at the current rate.

In the map on the left, red areas will see the biggest increase in temperature; in the map on the right, red areas will see the starkest decreases in rainfall.


Changes in rainfall will mainly affect South America and Indonesia, which will feel a similar impact in terms of precipitation volume.

Without the Amazon, it will be virtually impossible to limit the rise in global temperature to 1.5°C, as the Paris Agreement seeks to do. But the greatest suffering will be reserved for Brazilians and their South American neighbors.

The Consequences

Replacing the Amazon rainforest with pastures would have a direct impact on hydroelectric plants, which are a major part of Brazil’s energy mix. In a future with less rainfall, rivers will dwindle and water flows will slacken, leading to seasonal energy collapses.

According to an energy specialist who asked to remain anonymous because his professional contract does not allow him to speak to the press, reduced output from hydroelectric plants, which provide Brazil with 75% of its energy needs, would have to be compensated by the use of other resources. Brazil might opt for renewable energy sources such as solar or wind, but it would likely also build new thermoelectric plants (coal-fired or natural gas, for example) that would make its energy mix dirtier and increase greenhouse gas emissions, worsening the cause of the country’s water crisis.

The severest effects, however, would fall on agriculture. piauí asked Eduardo Assad, an agricultural engineer who has studied the effects of climate change on Brazilian agribusiness, what the 25% drop in average rainfall would mean. “A disaster,” said Assad, who works at Embrapa, the Brazilian Agricultural Research Corporation, and did not participate in the study. He explained that water will become scarcer, making for more unseasonal hot spells and dry periods during the rainy season, which in turn would affect the commodities that Brazil exports. “It’s already happening, and it’s on track to get increasingly worse.”

Studies carried out by Assad’s team since 2007 have shown that soybeans, corn, and coffee, in that order, would be the crops most affected by the decrease in rainfall. “The problem for soybeans won’t be a rise in temperatures, but a decrease in the water supply,” Assad said, suggesting two approaches to the issue. “The first is stopping the deforestation and replanting the areas so we can get the hydrologic cycle going again.” The second would be to adopt integrated crop-forest-livestock systems, which are better able to retain water in the soil, thus reducing the damage.

Assad said that the results of the study announced in Princeton are in line with projections made by other Brazilian research groups. “We’ve been saying this for over fifteen years, but right now, the country is hell-bent on dismissing scientific knowledge. These people keep on denying that climate change is happening and saying that there’s no problem with deforestation,” he said. “If we keep on this way, we’ll be faced with lots of problems, sooner rather than later.”

João Moreira Salles

Documentarista, é editor fundador da piauí. Dirigiu Santiago, Entreatos, Notícias de uma Guerra Particular e Nelson Freire

Bernardo Esteves (siga @besteves no Twitter)

Repórter da piauí, é autor do livro Domingo É Dia de Ciência (Azougue Editorial)

Leia também


Ovo indiciado

Delegada qualifica arremesso de ovo como tentativa de homicídio e prende acusada de atacar Marcha pela Família Cristã

Quem é o terrorista?

Dez questões para ajudar a entender por que o novo projeto de lei contra o terrorismo ameaça seu direito de manifestação

Descontrole total

Documentário mostra como Trump desdenhou da Covid – inspiração seguida pelo ocupante do Planalto

Covid rouba dois anos da expectativa de vida dos brasileiros

Mortalidade associada à doença diminui esperança de vida de 77 para 75 anos; Rio e Porto Alegre registraram mais mortes que nascimentos em 2020

Desenvolver vacina no Brasil é dose

Além do coronavírus, cientista batalha contra negacionismo, orçamento curto, burocracia e exaustão

Memória de uma dor sem fim

Mãe da menina Isabella Nardoni, assassinada em 2008, troca mensagens de apoio com o pai do menino Henry Borel

“Queremos testar medicamento contra Covid no Brasil”

Brasileiro reproduz em laboratório de universidade americana anticorpos que combatem coronavírus

Mais textos