Melbourne BA, Davies KF, Margules CR, Lindenmayer DB, Saunders DA, Wissel C, & Henle K (2004)
Species survival in fragmented landscapes: where to from here?
Biodiversity and Conservation 13(1): 275-284.
Key words: Extinction, Focal species, Habitat fragmentation, Habitat modification, Metapopulation, Modelling, Population viability analysis, Rules of thumb, Traits
We summarise the contributions of empiricists, modellers, and practitioners in this issue of Biodiversity and Conservation, and highlight the most important areas for future research on species survival in fragmented landscapes. Under the theme ‘uncertainty in research and management’, we highlight five areas for future research. First, we know little about the effects of density dependence on the viability of metapopulations, a requirement for fragmented landscapes. Second, successful early attempts suggest that it is worth developing more rigorous calibration methods for population viability analysis with spatially explicit, individual-based models. In particular, the balance between model complexity, ease of calibration, and precision, needs to be addressed. Third, we need to improve methods to discriminate between models, including alternatives to time-series approaches. Fourth, when our ability to reduce model uncertainty is weak, we need to incorporate this uncertainty in population viability analysis. Fifth, population viability analysis and decision analysis can be integrated to make uncertainty an explicit part of the decision process. An important future direction is extending the decision framework to adaptive management. Under the theme ‘tools for quantifying risk and predicting species sensitivity to fragmentation’, we highlight three areas for future research. First, we need to develop tools to support comparative approaches to population viability analysis. Second, population modelling can be used to find rules of thumb to support conservation decisions when very little is known about a species. Rules of thumb need to be extended to the problem of managing for multiple species. Third, species’ traits might be useful for predicting sensitivity but predictions could be further refined by considering the relative importance of population processes at different scales. Under the theme ‘tools for reassembling fragmented landscapes’, we consider the ‘focal species’ approach, and highlight aspects of the approach that require more rigorous testing. Finally, we highlight two important areas for future research not presented in the previous themes or papers in this volume. First, we need to incorporate the deterministic effects of habitat modification into the modelling framework of population viability analysis. Second, an avenue of research that remains largely unexplored is the combination of landscape-scale experiments and population modelling, especially using data from existing fragmentation experiments and from experiments designed to test the effects of defragmenting landscapes.