Using comprehensive avian survey data (1996-2005) and data layers derived primarily from a National Park Service vegetation map, we developed spatial models of species distribution and diversity within areas managed by the Golden Gate National Recreation Area (GOGA) and the Point Reyes National Seashore. Our objective was to assist GOGA and the Golden Gate National Parks Conservancy (GGNPC) with resource planning and management using landbirds as indicators. We used a focal species approach to select appropriate avian metrics, including the occurrence of disturbance-sensitive species, and species richness of focal species by habitat. Generalized additive models were used to represent non-linear relationships between vegetation and landscape characteristics and the distribution of avian species. Models were used to develop spatial predictions of species occurrence and diversity within the parks. Combining models for different management-sensitive species resulted in four different spatial representations of priority areas, each associated with different management goals.
Results/Conclusions
Model assessment using a subset of the data indicated strong predictive power for most models. Across the species and metrics examined, we found that landscape-level (within a 1-km radius) vegetation characteristics were generally more important predictors than local vegetation type (at the 30-m pixel level), although the latter was an essential component of each model. At the local level, the presence of hardwood vegetation (primarily oak and bay trees) was a positive predictor of many species, while at the landscape level, percent riparian habitat was an important factor. The spatial predictions generated by this study may be used by GOGA and GGNPC to identify priority areas for habitat conservation and potential habitat restoration and enhancement. They may also be used to determine the level and extent of possible impacts of selected planned management activities on landbirds. They are intended to be scalable, and may be used to address management questions at a variety of different spatial scales, but should be used primarily as a filter to identify potential target areas that may be investigated more thoroughly with site visits and surveys.
