Understanding fragmentation: snails show the way

Biodiversity is under assault on a global basis and species are being lost at an increasing rate, mainly due to habitat loss and fragmentation. The disciplines related to conservation have been described as ‘crisis disciplines’ (Doak & Mills, 1994), highlighting both the urgency of the issues and the fact that during a crisis there is rarely sufficient time to assess a situation cautiously before acting. The critical state of global biodiversity, the lack of adequate data and the need for quick decisions have led conservation scientists to rely heavily on theoretical techniques and empirical generalizations (Whittaker et al., 2005). Habitat fragments, terrestrial reserves and national parks are seen as ‘habitat islands’ inside human-altered landscapes, and, like newly formed islands isolated from the mainland by rising sea levels, they initially support a full complement of local species, but should lose species as they ‘relax towards a lower equilibrium’. It may take several generations for this process to play out following habitat destruction and fragmentation, thus delaying many of the species losses (lag time) (Ewers & Didham, 2006). This creates an ‘extinction debt’ – a future cost of habitat destruction that may not be apparent in studies made shortly after habitat fragmentation has occurred (Tilman et al., 1994). The ecological costs of the historically recent spate of habitat disturbance, destruction and fragmentation are yet to be realized. Despite much research on understanding relaxation, we lack universally accepted theoretical grounds for estimating the rate of species loss or the time required for a new (dynamic) equilibrium to be achieved (Doak & Mills, 1994; Laurance, 2008). Precise estimates of the ‘time to extinction’ of each species under threat remain unrealistic; the time lag for relaxation will vary according to circumstances. In untangling the processes underlying relaxation, we need to consider two factors that are critical in determining the net rate of loss across fragments: the rate at which habitat is being lost (taking account of simultaneous regeneration, Wright & Muller-Landau, 2006), and the age of the fragments (the time at which they were separated from others). The process of relaxation in relation to time has received little attention (Whittaker et al., 2005). However, a paper in Journal of Biogeography by Raheem et al. (2009) may set the scene for progress. They studied the land-snail fauna of 21 lowland rain forest fragments, ranging in area from c. 1 to 33,000 ha, in south-western Sri Lanka. The area contains most of Sri Lanka’s remaining lowland rain forest (less than 10% of its original extent). The Sri Lankan land-snail fauna is dominated by endemic species, 240 species out of c. 300, most of them occurring in the lowland rain forest. Raheem et al. (2009) have evaluated the effects of 28 environmental and spatial variables, such as longitude and latitude, fragment age, shape complexity, area, distance from edge, canopy density, altitude, etc., on the species composition in the fragments. Fragment age was quantified as the number of years elapsed since complete isolation from the largest remaining fragment of the region. Multivariate analyses were used to estimate the influence of fragmentation, degradation and pre-existing species turnover. Their findings highlight some crucial aspects of the fragmentation/ relaxation issue. First, fragment age was one of the two key determinants, along with fragment shape complexity, of the fragmentation-related changes in community composition. While other characteristics of fragments, such as area and isolation, might be important factors affecting the occupancy of many species, Raheem et al.’s findings coincide with those of others (Whittaker et al., 2005) supporting the need to include measurements of the temporal scale of fragmentation in order to understand changes in both species richness and assembly. We need to understand the history of the fragment from the time of creation/isolation, and changes in its abiotic and biotic properties, to reveal the spatial and temporal scales of the processes affecting a fragment’s carrying capacity and final equilibrium. Second, in most landscapes it is the productive and/or most accessible areas that are deforested first (Laurance, 2008); fragments that remain after forest loss vary in age. Raheem et al. show that deforestation was initially mainly in coastal and low-lying terrain, but later extended elsewhere. This highlights the non-random spatial distribution of habitat fragments with respect to age. As Raheem et al. (2009) say, ‘topographically diverse landscape may contain older, smaller and more degraded fragments at lower elevations and younger, larger and less degraded fragments at higher elevations’. Hence, time since fragmentation should broadly decline with altitude, especially on oceanic islands. Third, Raheem et al. conclude that the influence of pre-fragmentation patterns of species turnover persists despite fragmentation. Although there must be other contributing factors (48% of the variation in species composition remains unexplained), they show that there was a spatial pattern in species distribution before fragmentation that persists to the present, with many species having geographically restricted distributions. This pattern of restricted endemism is well known for land snails (Stanisic et al., 2007), and can be found in other organisms with poor powers of dispersal, especially where climate changes have altered the size and connectivity of suitable habitats over time. Considering a region, even a small one, as a priori biogeographically homogeneous before fragmentation is unwarranted. Thus, exploring the generality of this pattern for other taxa and/or other fragmented landscapes would advance our understanding of habitat fragmentation. Fourth, Raheem et al.’s study confirms others showing that land snails can be resilient to the effects of forest fragmentation: many species require only very small areas to persist for extended periods. Although the majority (72%) of the native land-snail species present in the study area Journal of Biogeography (J. Biogeogr.) (2009) 36, 2021–2022