A biodiversity Vision for the upper Paraná Atlantic Forest Ecoregion

To the three edge effects described above, we add a fourth and very important edge effect in our ecoregion – that of human activity. Hunting, illegal logging, illegal harvesting of non-timber forest products are more pronounced near the forest edge, but human activities penetrate up to a thousand meters into the forest. Hunting tends to decrease population sizes of most large vertebrate species in the Neotropics and to produce changes in the structure of mammal communities (Bodmer et al. 1997, Peres 2001, Bennett & Robinson 2001). Hunting in small forest patches can completely extirpate some species over the short term. For example, heavily hunted forest fragments of about 2,000 ha in the Upper Paraná Atlantic Forest in the western portion of the state of São Paulo, Brazil, were emptied of tapirs, white-lipped peccaries, and brocket deer (Cullen et al. 2000, 2001).

In forest fragments with highly irregular shapes, perimeter-to-area ratios are large and thus edge effects include a larger proportion of the fragment (Davies et al. 2001). For similar reasons, smaller fragments have a larger proportion of the area affected by edge effects than larger ones (Furlan et al. 2000). Very small forest fragments are entirely affected by edge effects and as a consequence are unlikely to preserve intact communities of the Atlantic Forest (Tabarelli et al. 1999).

Size effects. Ecologists have long recognized that there is a direct relationship between fragment size and number of species (Rosenzweig 1995). Just by chance alone (i.e., “sampling error”), a small fragment may fail to include individuals of rare or sparse species. In like fashion, sampling theory predicts that small forest fragments will include a smaller number of ecological communities. Because ecological communities are composed of unique sets of species, forest fragments that are missing communities will have decreased species diversity. The risk of a species´ local extinction within small fragments is also greater because of several factors that contribute to the extinction risk of small populations. First, random environmental variation such as fires or severe droughts can wipe out a small population. Second, deterministic threats (e.g., continuous deforestation or habitat degradation) can also decimate a population. Third, random demographic effects (e.g., a pronounced bias in the sex of new offspring) can drive small populations to extinction. Fourth, inbreeding and the loss of genetic variation are more common in small populations and render these populations less responsive to environmental change and more prone to extinction (Davies et al. 2001). In tropical rainforest fragments of about 100 ha, substantial numbers of under-story bird species are lost within two decades following fragment isolation. For many tropical bird species, forest fragments of less than 100 ha will have little conservation value (Ferraz et al. in press).

Some species have large habitat requirements and small fragments cannot fill those requirements. Chiarello (2000) estimated that only forest fragments in excess of 20,000 ha can sustain viable populations of medium-size to large mammals in the Atlantic Forest. A literature search of the habitat requirements of a small set of birds and mammals from the Upper Paraná Atlantic Forest shows that, even for species with relatively small habitat requirements (e.g., squirrels, armadillos, agoutis, and monkeys), a forest fragment of less than 1,000 ha is not large enough to maintain a viable population. For species with large habitat requirements (harpy eagles, jaguars, tapirs), it is necessary to maintain forest fragments of at least a few hundred thousand hectares (Table 2).

The disappearance of vertebrate species from forest fragments has a cascade effect on the ecosystem with consequences affecting other animal guilds, and even ecological processes like dung decomposition (Klein 1989), pollination and seed dispersal. In forest fragments, the absence of predators can result in an increase in herbivores, which can in turn have a dramatic effect on the forest structure and overall species diversity (Terborgh et al. 1999, 2001). The lack of top predators may allow for an increase in mid-sized predators, which may result in higher predation rates on birds and small mammals (Davies et al. 2001, Terborgh et al. 1999). This effect may explain the sudden increase in predation on the highly-endangered golden lion tamarin of the Atlantic Forest (J. Dietz pers. com.).

Our ability to preserve umbrella species, those with large habitat requirements, will thus be a good indicator of our ability to preserve intact biodiversity and healthy ecological processes. In order to maintain intact ecological communities and processes, it is essential to preserve the largest forest fragments that still contain individuals of umbrella species such as jaguars and tapirs. In the Upper Paraná Atlantic Forest, isolated forest fragments of about 2,000 ha have already lost jaguars, and those that are heavily hunted have lost many other large mammals as well (Cullen et al. 2000). However, forest fragments with several tens of thousands of hectares still have umbrella species and most of its biodiversity, including Morro do Diabo State Park in São Paulo Brazil with 35,000 ha (Cullen et al. 2000, Valladares Padua et al. 2002) and Mbaracayú National Park in NE Paraguay 59,000 ha (Zuercher et al. 2001, D. Ciarmiello pers. com.).

Based on species’ home range requirements in this ecoregion (Table 2; op. cited above), we may thus use 10,000 ha of well-protected forest as the lower limit for what we will consider a large forest fragment. The figure of 10,000 ha also corresponds to the minimum area requirement for a male jaguar (P. Crawshaw 1994 and pers. com.). A block of about 10,000 ha of well-preserved forest can contain one adult male jaguar and 1-2 adult females, thus constituting the area required by a minimum breeding unit for this species. For these reasons, we have chosen the jaguar as an umbrella species for this analysis and we will use this species to monitor the effectiveness of our Biodiversity Conservation Landscape design in the future.

Isolation. Considerable evidence suggests that isolated areas are difficult to re-colonize once their species are lost. Many forest species find it difficult or impossible to traverse the cattle pasture that often separates forest islands. The lack of gene flow into small and isolated forest populations contributes to the deleterious effects of inbreeding and increased probability of extinction (Dobson et al. 1999). The maintenance of biological corridors connecting forest fragments and allowing the movement of individuals and consequent gene flow, can reduce the deleterious effects of genetic isolation (Mech & Hallett 2001).

Forest fragments are not oceanic islands that have precise boundaries with the surrounding ecosystem, but they are usually surrounded by other terrestrial ecosystems. The matrix in which forest fragments are embedded may facilitate or preclude connectivity among forest patches. The more similar the matrix is to the original forest, the more opportunities for the native species to disperse to other forest fragments. The matrix can even provide alternative habitat for generalist species if the structural differences between the matrix and the original forest are small (Gascon et al. 1999, Davies et al. 2001). For example, scientists studying tropical rainforest dung beetles living in forest fragments of Amazonia near Manaus, found the rainforest beetles in only one surrounding clearcut - one containing extensive second growth vegetation (Klein 1989). However, to allow all native species to disperse among forest fragments those forest patches should be connected through corridors of native forest.