Fighting fire with fire in the Amazon

[SÃO PAULO] Sparks often fly when environmentalists and farmers come together, especially in Brazil. And that is exactly what happened last month at Fazenda Tanguro, a farm in Querência, Mato Grosso state.

But the sparks were real, not metaphoric. They came from fires ignited in the name of science and conservation as part of the ‘savannisation’ experiment, the world’s largest study of tropical forest fires.

Conducted by an international group of researchers and Brazilian farmers, this unusual bonfire at the edge of the Amazon rainforest was part of a US$660,000 project of the Large Scale Amazon Atmosphere-Biosphere Experiment, sponsored by the US space agency NASA, the US National Science Foundation, the Gordon and Betty Moore Foundation, and the United States Agency for International Development.

The focus of the study is to determine the impact of fire on the ‘transition’ forests that form a fragile boundary where the vulnerable rainforest meet the savannas — or cerrado — of central Brazil. The scrub-like cerrado is relatively well adapted to repeated burning. This happens naturally when lightning strikes, but local farmers also contribute to the process with fires called queimadas, which are used to control pests and weeds. Too many fires, however, could result in the runaway expansion of the savanna, triggering a process of biological erosion that threatens the edges of the Amazon rainforest.

During an El Niño — a cyclical warming of the Pacific Ocean, which has far-reaching effects on weather in other regions — up to 25,000 sq km of Brazilian forest are affected by low-intensity fires. Travelling 10-20 kilometres an hour at a height of a few centimetres, these fires are responsible for what has been termed ‘cryptic deforestation’.

Since the flames do not reach the canopy they do not kill trees directly. But they do increase the rate at which leaves fall, and therefore, the depth of the layer of dead leaves on the ground. The patchier forest canopy that results also lets more sunlight through, drying the fallen leaves and making the forest even more vulnerable to fires.

In 1998 — an El Niño year — fires in the northern state of Roraima destroyed 11,000 sq km of forest and shocked the world. But in the same year, almost no one noticed when low-intensity fires devastated nearly 30,000 sq km of the transition forest along the south-eastern fringe of the Amazon.

Logging, cattle ranching and soybean farming, are the main driving forces behind the fires, in that order. The farmers and ranchers start fires to clear the forest in order to cultivate the land, while loggers use fire to clear land to establish their camps. Each of these activities has been growing steadily in Brazil as a result of increasing international demand for timber, beef and soybeans. The US Department of Agriculture estimates that Brazil has 150,000 square kilometres of land suitable for mechanised agriculture. Standing between farmers and increased profits, however, is the transition forest — the interface between the precious, biodiversity-rich rainforest and the fire-adapted cerrado.

“It’s the decade’s major environmental issue”, says ecologist Daniel Nepstad, one of the savannisation experiment’s coordinators, who is based at the Woods Hole Research Centre in Massachusetts, United States, and Brazil’s Instituto de Pesquisas Ambientais da Amazônia (IPAM).

According to predictions by IPAM and the Federal University of Minas Gerais, at the present rate of recurring fires, only 15 per cent of the transition forest will remain by 2050. But the impacts of the fire are poorly understood because satellites cannot ‘see’ the scars they leave. This means they have to be studied on the ground.

“How frequent do fires need to be before a forest stops being a forest?” asks Nepstad. This is, in essence, the central question of the study, and reflects the concern that unless the fires become relatively infrequent, the transition forests are doomed to perish and be replaced by savanna.

Nepstad is leading the research jointly with biologist Claudia Azevedo-Ramos, also from IPAM. They hope that by ‘designing’ fires, they can measure what happens after the fires burn out. To do this, they have formed a unique partnership with export-oriented farmers. The experimental plots of transitional forest are located on an 820 sq km property belonging to the Maggi Group, Brazil’s major producer of soya beans, which is owned by the family of Mato Grosso governor Blairo Maggi.

Along with researchers from the Brazilian federal universities of Pará and Brasilia, and Yale and Stanford in the United States, Nepstad and Azevedo-Ramos had been camping at Fazenda Tanguro for a full month before August’s fire was started.

In an effort to mimic accidental fires, which usually start when sparks blow in from adjacent fields where farmers have lit queimadas, the researchers burnt a square kilometre of land. During the next six years, an additional two sq km will be burnt, with some sub-plots being treated this way every year.

The researchers are focusing on measurements of air temperature and humidity. In a dense rainforest, relative humidity reaches at least 65 per cent. But in the transition forest at Fazenda Tanguro, it is only 28 per cent along the edge, and hits a peak of 45 per cent deep in the woods.

Conditions in the area are being recorded before, during and after the fire at hundreds of points in the study area, which make up a grid 50 metres square.

The data collected will feed into a super-matrix of information on the dynamics of the fires, revealed by their speed and the amount of energy released from burning matter at any one point. Comparing the wealth of information measured on the spot with existing satellite images will allow the team to calibrate the satellite instruments so they are capable of reading the ‘signature’ of ground fires in the transition forest.

When combined, the satellite and ground-level data will be used to make up an image of a one sq km square of land, striped with the marks of different fire lines. Comparing fire intensity and speed with other variables such as the time of the day, type of vegetation and the presence of bare surfaces around ant nests, the researchers hope to make reliable computer models of the forests’ vulnerability to fire and the likelihood of their conversion to savanna in the near future.

The transition forest shares much of its biodiversity with the denser Amazon rainforest, and IPAM’s Oswaldo de Carvalho Júnior says that 65 per cent of the 46 species of wild mammals said to live in the region have already been spotted in the experimental area.

As the composition and density of animal populations after a fire is bound to influence the future structure of the surviving forest, the researchers also want to find out which animals suffer most when there are repeated low-level forest fires. Some animals, for example, are important seed dispersers but other species feed on seeds. Which survive will affect whether forests can regenerate.

But flames, not data, were the most visible result of the strange gathering that mid-August Monday morning at Fazenda Tanguro. At one relatively open spot, where the sunlight had dried the mat of decaying leaves on the forest floor, the flames shot 20 metres high into the canopy. The noise was deafening.

The fire’s horizontal movement is also a concern. To control the fires, the researchers have the gruelling task of using machetes to cut safeguard firebreaks known as aceiros, which are about 30 centimetres wide.

When the day’s burning is over, the tired researchers drag themselves to their 4×4 trucks, their faces dark with soot. A few will ride back to their base perched on a tractor, just in case one needs to cut a five-metre wide aceiro to prevent fire from jumping into the neighbouring pasture. Unlike the carefully monitored experimental fires, that would be a blaze that nobody wants — and a reminder of just how open to damage this unique forest borderland is.