The species-area relationship (SAR) gives a quantitative description of the increasing number of species in a community with increasing area of habitat. In conservation, SARs have been used to predict the number of extinctions when the area of habitat is reduced. Such predictions are most needed for landscapes rather than for individual habitat fragments, but SAR-based predictions of extinctions for landscapes with highly fragmented habitat are likely to be biased because SAR assumes contiguous habitat. In reality, habitat loss is typically accompanied by habitat fragmentation. To quantify the effect of fragmentation in addition to the effect of habitat loss on the number of species, we extend the power-law SAR to the species-fragmented area relationship. This model unites the single-species metapopulation theory with the multispecies SAR for communities. We demonstrate with a realistic simulation model and with empirical data for forest-inhabiting subtropical birds that the species-fragmented area relationship gives a far superior prediction than SAR of the number of species in fragmented landscapes. The results demonstrate that for communities of species that are not well adapted to live in fragmented landscapes, the conventional SAR underestimates the number of extinctions for landscapes in which little habitat remains and it is highly fragmented.extinction threshold | habitat conversion | metapopulation capacity | Atlantic forest | Nagoya biodiversity agreement T he species-area relationship (SAR) describes a very general pattern in the occurrence of species, which is fundamental to community ecology (1), biogeography (2), and macroecology (3). Since the 1920s (4, 5), SARs have been applied to describe the occurrence of a wide range of organisms on true islands (6-8), in fragments of distinct habitat (9, 10), and in parts of more arbitrarily delimited contiguous landscapes (1, 3). In the past decades, SAR has become an important concept and a tool also in conservation biology, where it has been used to make broad assessments of species extinctions from habitat loss (11-18). These calculations have been criticized for various reasons (17,(19)(20)(21), but minimally SAR provides a valuable point of reference for the threat that habitat loss poses to biodiversity.SARs are typically applied to a set of habitat fragments within a single landscape, but in conservation, in contrast, the essential question is how many species will persist in different landscapes (regions) with dissimilar amounts of habitat rather than in different fragments within a single landscape. This creates a problem: Habitat loss is virtually always accompanied by fragmentation (22-24), and hence the remaining habitat is not contiguous, unlike assumed by SAR, at the landscape level. In other words, SAR does not account for any adverse effects of fragmentation on the occurrence of species (25, 26). Fragmentation matters whenever individual habitat fragments are small enough to reduce the viability of the respective local populations (27, 28). ...
