Carefully stowed away in a locker, Evlyn Novo keeps a collection of framed plaques honoring her time as a researcher at the Brazilian National Institute for Space Research (INPE). Novo joined the institute as a young remote sensing specialist in 1975 to work on a pioneering effort to use satellite data to monitor deforestation in the Amazon. Over her career, she helped INPE develop into one of the flagships of Brazilian science—a global leader in watching tropical forests from space. For every 5 years Novo spent at INPE, she received a commemorative plaque honoring her service. She was looking forward to getting the 10th one.
But with only 2 years to go until that milestone, Novo, 69, has come to a heartbreaking decision: She has lost faith in the institution’s future and will retire from INPE by the end of this year. “I don’t want to be the one to turn off the lights,” she says.
INPE is in decline, and Novo sees it everywhere. A few years ago, the office lights often stayed on until late at night at INPE’s main campus in São José dos Campos, near São Paulo, where staff and students analyzed remote sensing data, built satellites, and modeled weather and climate. Today, the institute struggles to pay its electricity bills. Potholes pepper the campus streets and sidewalks are broken. They are the physical symptoms of a much larger institutional crisis marked by sharp budget cuts, a shrunken staff, and relentless attacks by Brazilian President Jair Bolsonaro and his supporters. “It is a climate of total dismay,” Novo says.
Other sectors in Brazilian science are backsliding, too. A faltering economy and shifting political priorities have led to steep cuts in funding for science at many universities and federal research institutions. But INPE’s downfall has been particularly painful to watch, many say, because of its international prestige, its role in protecting the Amazon, and the way Bolsonaro has added insult to injury by trying to discredit the institute’s work.
Today, Brazil’s monitoring programs for deforestation and wildfires are imperiled, the supercomputer that runs climate models is aging and unreliable, and INPE’s satellite development program is on hold, lacking funds to advance planned missions and launches. A beacon of Brazil’s scientific prowess has become a symbol of science’s struggle for survival there, in an underfunded and politically menacing environment.
INPE’s director, Clezio De Nardin, acknowledges that his institute is in trouble. In an interview with Science, he says INPE needs at least twice its current operating budget of 92 million reais ($18 million) a year to fulfill its missions. But he blames the cuts on Brazil’s economic problems, not on politics. “I don’t believe any ruler in good conscience would act to destroy an institution that produces essential infrastructure for its own country,” De Nardin says. “Especially because defunding the space sector will have consequences in society for decades.”
In an email to Science, Brazil’s Ministry of Science, Technology and Innovation (MCTI) pointed to a few positive signs: This year, INPE’s discretionary funding went up by 33%, bringing it back to 2019 levels, and grants from the National Fund for Scientific and Technological Development (FNDCT) will add an additional boost.
But researchers say the amount of grant support is uncertain, and the extra funding won’t be nearly enough to reverse years of decline. By now, many staffers have become demoralized, says Gino Genaro, a senior satellite technologist at INPE: “People are distressed without knowing what to do and what the future will hold.”
INPE WAS FOUNDED in 1961, when the space race was in full swing and the Cold War at its peak. Inspired by a visit from Soviet cosmonaut Yuri Gagarin, then-President Jânio Quadros embraced an idea from the Brazilian Interplanetary Society to launch a national space institute. In its early years, INPE created and consolidated research and postgraduate programs in meteorology, astrophysics, and remote sensing. In the 1970s, it became the first space agency in any country to monitor forests using satellite data, obtained by the U.S. Landsat Program.
In 1989, the agency launched the Brazilian Amazon Deforestation Satellite Monitoring Program (PRODES), which provides yearly and historical deforestation trends for the world’s largest rainforest. In 2004, INPE added a system named Real-Time Deforestation Detection (DETER), which uses real-time images from a variety of satellites, some of them developed partly in Brazil, to send daily and monthly updates of fires and other causes of forest loss to enforcement officials.
Data from PRODES and DETER helped Brazil create and enforce policies that were key to reducing annual deforestation in the Amazon by 82% between 2004 and 2014. Both programs are “fundamental for understanding the agricultural expansion in Brazil and carbon emissions from deforestation in the Amazon,” says Douglas Morton, a remote sensing specialist at NASA’s Goddard Space Flight Center. Morton says researchers worldwide prize INPE’s data, which it made publicly available on the internet in 2003, a pioneering step that the United States and other countries would follow. The institute is also an important training center for remote sensing scientists, Morton adds.
Another source of pride at INPE is the Integration and Testing Laboratory (LIT), which assembles satellites. The only lab of its kind in the Southern Hemisphere, LIT has tested and developed many satellites in the past 35 years, most in partnership with other countries. The most successful and long-standing project was a collaboration with China to launch six imaging satellites, the first one in 1999, that gave Brazil its own satellite views for the first time. Two are still used today to monitor fires, deforestation, and land use.
“INPE is certainly one of the most important research institutions in Latin America,” says computer scientist Gilberto Câmara, a career researcher who led the institute from 2006 to 2013. During Câmara’s tenure, in 2010, Brazil invested $13 million in a new Cray XT6 supercomputer, for use by INPE teams and researchers elsewhere in Brazil. Nicknamed Tupã, after an Indigenous South American god of thunder, it was one of the 30 most powerful computers in the world at the time—and it was hailed as another milestone for Brazilian science.
But the string of successes wouldn’t last.
INPE’S FORTUNES TURNED about a decade ago, during the first term of leftist President Dilma Rousseff. Federal support for science, including at INPE, suffered in a maelstrom of economic and political troubles. The turmoil culminated in Rousseff ’s impeachment and removal in 2016, and Bolsonaro’s election in 2018.
The decline accelerated after Bolsonaro took office, despite his campaign promises to prioritize science. His government cut M CTI’s overall budget by 35% in the first 3 years of his administration, to 8.3 billion reais ($1.7 billion). Despite a partial recovery this year, after Congress banned the government from freezing FNDCT funds, MCTI’s overall budget for this year is still 34% smaller than 5 years ago, not counting inflation.
INPE’s annual budget, meanwhile, plunged by 63% from 2010, to a record low of just 76 million reais ($15 million) in 2021. Even after the recent boost, it is still “obviously behind what is needed,” De Nardin tells Science.
The consequences have been felt at every corner of the institute. The budget dedicated to the two rainforest monitoring programs declined by 70% in the past decade, to 2.7 million reais ($540,000) last year—about 0.1% of what New York City spends on its Department of Parks & Recreation annually. “If there are more budget constraints, we will have to stop doing something; maybe reduce the area we monitor,” says INPE’s senior technologist Cláudio Almeida, who coordinates the two programs.
Keeping an eye on other corners of Brazil’s rich biodiversity has also been challenging. One program tracks deforestation and wildfires in the Cerrado, 2 million square kilometers of shrubland bordering the Amazon forest that covers most of central Brazil. Among the world’s most biodiverse savannas, the Cerrado is under heavy pressure; more than half has already been cleared for crops and cattle. Last year, Almeida’s team avoided shutting down the program only by transferring money from another project at the last minute. They hope a FNDCT grant of 15 million reais ($3 million), released in late April, will keep the program going another 3 years.
The funding crisis is “very worrisome,” says Julia Shimbo, an ecologist at the nongovernmental Amazon Environmental Research Institute. INPE’s data are not only used for public policies on deforestation, she notes, but are also a reference for international agreements on greenhouse gas emissions to which Brazil has committed.
The cash crunch creates other problems as well. After 12 years, Tupã is on its last legs. Processors frequently get overheated, circuit breaks are not uncommon, and the institute can barely afford the electricity to run the computer. “This obsolescence affects the country’s climate research, weather forecasts, and INPE’s mission itself,” says environmental physicist Paulo Artaxo of the University of São Paulo’s main campus, one of many external scientists who have stopped using Tupã.
There is no money for a replacement. As a makeshift solution, INPE bought a second, smaller and less powerful machine in 2018 that took over weather forecasting and other everyday operational processes, while the old supercomputer is mostly used for research. Last year, with funds from the United Nations Development Programme, it bought a set of additional processors for Tupã. But if the old machine dies, Brazil’s climate modeling program might shut down. The agency has submitted a request to the Brazilian Innovation Agency for 200 million reais ($40 million) to upgrade its entire computer system. Even if the project is approved, it will take at least a year until the new machinery is ready to work.
LIT, the satellite development facility, has also suffered heavily. Its last major project was the Amazonia-1, the first satellite entirely developed in Brazil, launched from India in February 2021. It added an extra eye in the sky to watch over the Amazon, increasing the frequency at which images of the region are generated and allowing faster deforestation alerts.
Amazonia-1’s launch was a cause for celebration, but the success masked many problems. A lack of funds had delayed the mission by 3 years. Most researchers working on it had lost their grant by that time; the grants had to be reinstated temporarily to make the launch possible. And funds are lacking to develop two long-planned companion satellites. “The discontinuity of this project is a shot at the heart of the Brazilian space program,” says former INPE Director Ricardo Galvão. As a continental country with huge areas of native vegetation and crops, Brazil needs a whole constellation of new satellites to meet its remote sensing demands, Galvão says.
To save on LIT’s formidable electricity bill, INPE has reduced its operations, shutting down cleanrooms from time to time. The room where spacecraft were assembled now houses the skeleton of a satellite, built from spare pieces of the Amazonia-1, “to show off to visiting politicians and journalists,” an employee tells Science. LIT has long tested products for industry on the side, such as cars, phones, and even bathtubs; today that is practically all it does.
De Nardin says the institute is looking for funds to develop new Amazonia satellites. But meanwhile, Genaro, who has worked at the institute’s space systems division for 20 years, says, “All we have are loose ideas on paper, without any budget or a team to execute them.”
MONEY IS NOT THE ONLY resource in short supply. There’s also a dearth of brains, especially young ones. INPE has seen a striking reduction in personnel over the past 2 decades as researchers retired with no one to replace them. The last time INPE was allowed to add permanent staff was in 2012. The cash-strapped government is reluctant to allow hiring, in part because Brazil’s generous pension system has become a huge financial burden. Other federal research institutes and universities face the same problem.
As a result, the number of full-time employees at INPE has dropped from some 2000 in 1990 to about 753 today. (Of those, 146 are researchers, 460 work in technical positions, and the remainder in management and administration jobs.) Like most other groups at INPE, Almeida’s Amazon monitoring team relies mostly on temporary researchers with external grants who often don’t stay long because of payment delays and the instability of living without a job contract.
“Back in the day, it was superhard just to find room to put up a desk at INPE,” Almeida recalls. Nowadays, he says, “any fellow [grantee] has their own office.” It’s not just empty, he says. “The institute is growing old.”
Younger scientists see little future there. Digital ecologist Thiago Silva did his postdoctoral research at INPE between 2010 and 2013, using satellite data to study wetland dynamics in the Amazon. Working under Novo, he was one of two contenders selected to be hired by the institute in 2012. He turned down the offer, fearing he might end up isolated and without enough money—or even colleagues—to continue his work in the long term. Silva moved to Scotland in 2019 to become a professor at the University of Stirling. “We end up having to leave Brazil to keep our research when it would be much more beneficial to stay in the country,” he says.
Novo’s last four graduate students have left the country as well. Most of their research lines were discontinued. “The situation is nerve-racking,” Novo says. “You spend years building a lab, buying equipment, and forming human resources. And all that can disappear overnight because those in charge of policies lack long-term vision.”
The government says INPE can’t complain. Marcos Pontes, who left as Brazil’s minister of science, technology, and innovation in April to run for Congress, told a local radio station in February that the institute was “one of the most privileged” units within the science ministry. Pontes, best known as the country’s first and only astronaut—he spent 9 days on the International Space Station in 2006—said every institute supported by MCTI had suffered restrictions and no one could accuse him of “making deliberate cuts” to INPE. De Nardin agrees that the government isn’t singling INPE out for cuts.
Many critics see it differently. They say Bolsonaro’s government has intentionally targeted INPE, angered by its leadership and transparency in monitoring deforestation and environmental crimes, which critics say the government’s pro-development policies have encouraged. Since Bolsonaro came to power in 2019, deforestation and fires in the Amazon have risen to their highest levels in more than a decade. “INPE is paying the price of the dismantling of the environmental sector in Brazil,” says biologist Izabella Teixeira, who served as environment minister from 2010 to 2016. To her, INPE’s crisis reflects the Bolsonaro government’s “ideological contempt for the environment and science.”
The unruly president has not hidden his dissatisfaction with INPE. “What happens with a lot of INPE reports … is that they just copy the previous year’s reports,” he said in 2019. Bolsonaro accused INPE of “lying” about an uptick in forest destruction and had Galvão—a renowned physicist—fired from the director’s post for confronting him in public about it. (Galvão was replaced by Darcton Damião, a retired air force colonel with a master’s degree in remote sensing from INPE, who was succeeded by De Nardin in October 2020.)
Many INPE employees and researchers who spoke with Science believe the institute is a victim of politics, but did not want to say so on the record. “You see that people have a scream stuck in their throats because they don’t dare to express their ideas, fearing retaliation,” Novo says.
“Little by little, they are making INPE disappear,” says Thelma Krug, who worked at the institute for 37 years, where she helped create the Amazon monitoring programs and had a leading role as an environmental data analyst. Krug, one of the three vice presidents of the Intergovernmental Panel on Climate Change, retired in 2019. “I left because I could not agree to not speaking about what was going on at INPE,” she says.
IF INPE STOPS monitoring deforestation and fires, companies and nongovernmental institutions based in Brazil and other countries can fill the gap. But the weakening of the agency is also a weakening of Brazil’s sovereignty, Câmara says: “It is fundamental for the state to have the competence to produce its own data.” Many scientists say INPE can only recover if Bolsonaro is defeated during the presidential election in October and a new government takes over. The campaign has yet to begin, and Bolsonaro’s main competitor in the polls, former President Luis Inácio Lula da Silva, has not announced specific plans for science, although he has alluded to prioritizing investments in science and education. INPE flourished and Amazon deforestation went down sharply during Lula da Silva’s previous presidency, between 2003 and 2010.
Novo is less optimistic. “Even if there is a change in government, it will take a long time for INPE to recover,” she says. Novo could have retired 18 years ago but stayed because she loves her job. Now, she is just waiting for her last students to finish their theses. “I want to cry when I think about INPE’s destiny,” she says, her voice cracking.