The limit of the forest

Group from São Paulo identifies the signs that precede successive extinction in the Atlantic rainforest

FABIO COLOMBINIAmphibians affected by deforestationFABIO COLOMBINI

The Atlantic rainforest is a forest in pieces. According to recent estimates between 11% and 16% of its original cover remains, most in fragments of fewer than 50 hectares of continuous vegetation, surrounded by plantations, pasture and cities. For some time it has been known that this disjointed architecture makes it difficult for the forest to recover , which is one of the ten most threatened in the world. Now the teams of ecologist, Jean Paul Metzger and zoologist Renata Pardini, both from the University of São Paulo (USP), have found that for each group of animals living in this patchwork landscape, there is a point of no return, a minimum limit of native vegetation cover that needs to remain standing in order to maintain the variety of species in a particular region. When deforestation surpasses this limit, most of the species become extinct in all stretches of forest in the area.

For almost a decade 60 researchers coordinated by Metzger collected information about the abundance and diversity of amphibians, birds and small mammals in dozens of stretches of Atlantic rainforest on the Western Paulista Plain, sloping land that extends from the coastal hills (Serra do Mar) in a westerly direction and cover almost half the state. In comparing data, the researchers observed dramatic falls in biodiversity in the fragments in neighboring and similar regions, which differed only in the total area of native vegetation remaining.

A landscape that has 50% of its native forests preserved, albeit dispersed in fragments, for example, had a diversity of birds that were three times more sensitive to vegetation loss than a similar landscape with 30% of its vegetation. Small mammals, like Bishop’s slender opossum (Marmosops incanus), a marsupial with gray fur, large eyes and a long snout, better resisted the forest being chopped down. However, even for them the end came quickly once the limit had been reached. There was a fall of between 60% and 80% in the number of species when the native area shrank to less than a third of the original.

The study does not supply a single figure for the minimum area of native forest necessary for keeping the biodiversity of the ecosystem intact. “For other groups of animals the sudden loss may happen earlier or later, depending on the dislocation of species and their resistance to disturbance,” explains Metzger, who says the limit fixed for deforestation of Atlantic rainforest established by the current Forest Code is reasonable.

Data analysis allowed Renata and Metzger – with the help of Adriana Bueno and Paulo Inácio Prado, also researchers from the Biosciences Institute at USP, and Toby Gardner, from the University of Cambridge, England – to arrive at a plausible explanation as to why the biodiversity of fragments of Atlantic rainforest reduced in some regions, but in others remained similar to the stretches of continuous vegetation on the coastal hills. In an article published in October 2010 in PLoS One journal, they present a conceptual model of how this could happen.

According to the model, the collapse of populations would be caused by a combination of processes that occur on two leveles: local and regional. Processes with a regional effect are linked to the difficulty of migrating from one forest fragment to another. Conditional upon the total area of remaining forest in the region this difficulty increases as deforestation advances, because the distances separating the sections of forest grow exponentially, and many species, and even birds like the white-collared foliage-cleaner (Anabazenops fuscus), do not move from one fragment to another when there are pasture or roads in the way. Trapped in restricted areas these species become more susceptible to processes that influence extinction on the local scale, like a reduction in the area of the fragments, which reduces the size of populations.

ARTHUR GROSSETWhite-collared foliage-gleaner: unable to cross pastures and roadsARTHUR GROSSET

The most important thing is that this model can guide decisions about the best way of applying resources for conserving and recuperating the Atlantic rainforest. According to the researchers, it forecasts, for example, that the events that precede extinction would provide clues of its arrival beforehand. The way in which the species are distributed in the fragments in a region signals when the biodiversity is on the brink of declining abruptly, but there is still a good chance of it being helped to recuperate. “Under these conditions small investments in restoration that facilitate the flow of animals between fragments would produce a great return,”  says Metzger. “If we want to increase Atlantic rainforest vegetation cover, with rapid gains in biological diversity, it is in this band [regions with 20% to 40% remaining] that we have to attack.”

One of the fundamentals of the model is the evolution of the fragments, which Metzger and other researchers started understanding better, when they carried out computer simulations of how deforestation cuts down a virtual forest. These simulations suggest that as the native vegetation in a region decreases a fundamental transformation occurs in the characteristic of forest fragments: the distance between them, which initially falls gradually, starts to increase exponentially.

Starting with these results, confirmed in real landscapes by Metzger and researchers from other countries, the group from USP began to question if the evolution of the geometry of the fragments could affect the biodiversity of a region by having an influence on two phenomena that are well known by ecologists.

One of them is the influence that the area of a fragment has on the probability of survival of a population. The greater the area, the larger the populations of species that live there – and, therefore, the smaller the risks of being extinct because of an accidental event, such as only females being born in a particular year or the occurrence of a natural disaster.

By this reasoning, the biodiversity of a fragment should be proportional to its area. However, this is only part of the story. “It’s not only the size of the fragment that matters, but also the landscape in which it is found,” explains Renata. After all, fragments are not perfectly isolated islands. If they are sufficiently close to each other, many species of animal can travel between them, thus avoiding the extinction of populations in smaller fragments. “This is the rescue effect,” she says. “Although a fragment is small and the risk of extinction great, the population is maintained because of migration.”

The researchers imagined, therefore, that in the deforestation process, before the accelerated increase of the distance between fragments occurred, these stretches of forest would still be sufficiently close to each other for the rescue effect to keep biodiversity high in the region. With the reduction in remaining forests, however, this effect loses strength and the diversity of the small fragments is reduced, although the total biodiversity of the region is preserved, with most of the species being concentrated in larger fragments. At this stage, it is possible to observe the effect of the size of fragments in the region, which makes the diversity of species in a fragment proportional to its area.

This effect predominates until deforestation starts increasing the distance between sections of forest exponentially. Once this limit has been passed rescue stops and the risks of population extinction increases for a large number of species, which disappear from both the large and small fragments.

FABIO COLOMBINITesting the model
The next step was to test whether the model could forecast the distribution of the species that the group from USP had observed in the field work it had done between 2000 and 2009 in the State of São Paulo, with the support of FAPESP, the National Council for Scientific and Technological Development and the Federal Ministry of Education and Research of Germany. In the project, the researchers carried out a survey of amphibians, birds and small mammals in three areas of 10,000 hectares that had native forest with different degrees of preservation (50%, 30% and 10%) and in three areas of continuous Atlantic rainforest on the coastal hills.

After capturing the animals and identifying their species the researchers separated them into two groups: specialist species that only live in stretches of Atlantic rainforest, and generalists that are capable of surviving in the forest as well as in areas modified by human action, such as plantations and pasture. The classification was essential to compare the survey data with the theoretical forecasts on the effect of fragmentation that should only be observed for specialist species.

In the case of small mammals, 27 of the 39 species found were specialists. For these the diversity patterns observed were as expected. In the region with 50% native cover, both large and small fragments contained almost all the species found in the region of neighboring continuous forest. These same species were also found in the region that had 30% forest, but were concentrated in the larger fragments. In the region with 10% forest, on the other hand, the deforestation threshold had been exceeded and diversity was uniformly low: its fragments, regardless of the area, sheltered from three to five times fewer specialist species than the region of continuous forest.

The researchers also noted that, in the absence of specialist species, the populations of generalist species exploded in the region with 10% forest. In areas with 50% forest, 63 black-footed pygmy rice rats (Oligoryzomys nigripes), a generalist species, were caught, while the number soared to 409 in the region with less forest. This fact is worrying. This rodent is the main carrier in the Atlantic rainforest of the virus that causes human hantavirus disease and its presence in pasture and plantations may increase the risk of contagion in people.

This is just one example of the impact that the loss of biodiversity may have on the health and quality of life of humans. Other services provide by natural ecosystems, like the pollination of plantations and the control of agricultural pests, may also disappear. “We don’t want to preserve biodiversity to maintain living museums, but to maintain the services that the remaining ecosystems provide,”  says ecologist, Thomas Lewinsohn, from the Public University of Campinas, who did not take part in the research.

For Lewinsohn, the work of Renata’s and Metzger’s groups represent a qualitative leap in ecology, because it combines difficult to carry out field study, with a theoretical model that explores the final consequences of different effects, that were previously discussed in a separate way by researchers investigating the reduction and fragmentation of natural environments worldwide. “They have made an important contribution to our understanding of the consequences of forest loss on biodiversity,” comments ecologist, Ilkka Hanski, from the University of  Helsinki in Finland, a pioneer researcher into the impact of transformations in habitat in plant and animal communities. “This study is likely to become highly influential in the biology of conservation.”

The project
Diversity of mammals in fragmented landscapes of the São Paulo Atlantic plateau (nº 2005/56555-4); Type Young Researcher Grant Program; Coordinator Renata Pardini – IB/USP; Investment R$ 264,307.22 (FAPESP).

Scientific article
PARDINI, R. et al. Beyond the fragmentation threshold hypothesis: regime shifts in biodiversity across fragmented landscapes. PLoS One. Oct 23, 2010.