Matrix habitat and species richness in tropical forest remnants

Abstract

The abilities of species to use the matrix of modified habitats surrounding forest fragments may affect their vulnerability in fragmented landscapes. We used long-term (up to 19-year) studies of four animal groups in central Amazonia to test whether species' abundances in the matrix were correlated with their relative extinction proneness in forest fragments. The four groups, birds, frogs, small mammals, and ants, had varying overall responses to fragmentation: species richness of small mammals and frogs increased after fragment isolation, whereas that of birds and ants decreased. For all four groups, a high proportion of nominally primary-forest species were detected in matrix habitats, with 8–25% of species in each group found exclusively in the matrix. The three vertebrate groups (birds, small mammals, frogs) exhibited positive and significant correlations between matrix abundance and vulnerability to fragmentation, suggesting that species that avoid the matrix tend to decline or disappear in fragments, while those that tolerate or exploit the matrix often remain stable or increase. These results highlight the importance of the matrix in the dynamics and composition of vertebrate communities in tropical forest remnants, and have important implications for the management of fragmented landscapes.

Introduction

Deforestation has led to striking fragmentation of natural landscapes throughout the world (Whitmore, 1997, INPE, 1998). Fragmented landscapes are composed of a patchwork of habitats of differing quality for fauna. With the exception of islands in hydroelectric reservoirs, human-fragmented landscapes maintain some degree of terrestrial connectivity via the matrix of modified habitats surrounding fragments. It has become increasingly apparent that understanding how species are affected by fragmentation requires information on their responses to all components of the landscape, i.e. forest patches, remaining continuous forest, and the intervening matrix (Harris, 1984, Laurance, 1990, Laurance, 1991, Laurance, 1994, Malcolm, 1991, Fahrig & Merriam, 1994, Aberg et al., 1995, Lidicker. 1995, Flather & Sauer, 1996, Tocher et al., 1997).

The matrix is likely to be important in the evolution of fragment dynamics for several reasons. First, the matrix will often act as a selective filter (not an absolute barrier) for the movements of species across the landscape. The type of vegetation in the matrix will determine the pore size of the filter for movements of individuals. Tall second-growth forest, for example, would be analogous to a filter with large pores allowing more faunal movement because of its structural similarity to primary forest, while a pasture would have small pores that impede much movement (Malcolm, 1991). If this filter analogy is correct, one would expect a gradual increase in species richness, at least of those associated with primary habitat, from fine-pored (i.e. pasture) to larger-pored (i.e. second-growth) habitats. Forest species would be expected to vary in their pore-size requirements, with species that tolerate smaller pores being better able to persist in forest remnants (Laurance, 1990, Laurance, 1991, Laurance, 1994).

Second, the matrix may exert a strong influence on within-remnant community dynamics (Janzen, 1986, Laurance, 1990, Malcolm, 1991). For example, species associated with the matrix may invade forest remnants, altering the species composition of some taxonomic groups (Hutchings, 1991, Brown & Hutchings, 1997, Laurance, 1994, Tocher et al., 1997). Finally, different matrix types may influence fragment edge effects. Edge-related gradients in physical and biotic variables are likely to be less pronounced when the matrix is structurally similar to the original habitat, and this may reduce the ecological impacts of edge effects (Laurance & Yensen, 1991, Mesquita et al., 1999).

To investigate the importance of the matrix habitat in explaining species richness in forest fragments, we use results from long-term surveys of four taxonomically diverse groups (ants, small mammals, frogs, and birds) from the Biological Dynamics of Forest Fragments Project (BDFFP) in central Amazonia. To our knowledge, this is the first test of matrix effects using such a large and taxonomically diverse assemblage of species.