Session Information
| Session | Poster Session | | Date | Monday (2008-04-07) | | Time | 5:00 PM - 7:00 PM | | Room | Grand Terrace |
Presentation Information
| Presenter | Clara Grilo | | Title | Sensitivity of carnivores to habitat fragmentation from a multi-scale perspective | | Affiliation | Centro de Biologia Ambiental | | Authors | Clara Grilo, Marta Cruz, John Bissonette, Margarida Santos-Reis | | Keywords | Carnivores, Landscape ecology, Montado, Multiple scales, Variance decomposition | | Presentation Type | Poster | Abstract:
Throughout history, most landscapes of the Mediterranean Basin have been modified by interactions between humans and nature. An example is the montado landscape in the western part of the Mediterranean Basin, comprised of an open formation of cork (Quercus suber) and holm (Quercus ilex) oaks combined with a rotation of crops/fallow/pastures. Over time the montado has experienced significant large-scale changes, including habitat conversion, loss, and fragmentation. In spite of co-evolutionary processes that have allowed species to become resilient to small-scale habitat changes, species have been impacted by the scale of the spatial and temporal changes in the montado. The main goal of this research was to assess the effects of fragmentation of the Mediterranean landscape on medium-sized carnivores. We used Mann-Whitney U tests to test for differences between visitation rates obtained from scent stations located in habitat patches of different size and isolation. We used variance decomposition procedures to specify what proportion of the variance in carnivores occupancy was explained by each of the three spatial scales exclusively and what proportions are attributable to interactions between variables. We found that with the exception of Eurasian badger (Meles meles) carnivores responded negatively to habitat loss. Red fox (Vulpes vulpes), stone marten (Martes foina), genet (Genetta genetta) and mongoose (Herpestes ichneumon) were found more often in continuous sites than in patches, but we found no differences in visitation rates between different patch sizes. When we compared connected patches with those far from the nearest large patch we found significant differences in visitation rates. Only stone marten showed a low tolerance for patch isolation; it showed even lower visitation rates when patches were greater than 1500m apart. Although sensitive to habitat loss, the species responded in different ways. The higher variance was attributable to stone marten (57%), mongoose (43%) and genet (40%); the fox model (most habitat generalist species) explained only 15% of the variance. The effect of patch and landscape factors taken singularly, as well as their interaction was evident in species occupancy. Specifically, single patch factors influenced the overall variability for the fox and mongoose occupancy most whereas genet occupancy variability was mainly explained by the landscape factor. For stone marten occupancy, the patch factor had a substantially high independent explanatory power; nevertheless taken together, more variance was explained by using a three scale extent analysis that combined the effects of the interaction of patch and landscape factors. |
|