In southeastern USA, planted pine forests are increasingly embedded in a complex matrix of urban development and/or extensive agriculture. Although majority of planted pine forests are primarily managed for timber harvest, there are growing interests in management plans to conserve avian species in these habitats. However, most management has focused on local or stand scale. Little attention has given to landscape-scale habitat characteristics. We investigated the relationship between avian species occupancy and three spatial scales: local-scale (50 m-circular area of a sampling point), 500m-circular landscape, and 1 km-circular landscape. During May-August 2010 and 2011, we conducted 10-min 50-m fixed-radius avian point counts and vegetation sampling (Relevé method) at pine stands in the Sand Hill of Georgia. We generated landscape scale variables digitizing aerial photographs and using FRAGSTAT. We used 13 species and 7 species which were detected >15% of total points and whose occupancy models did fit adequately, in 2010 and 2011, respectively. We performed Principle Component Analysis at each scale and used 3 principle component scores of each scale for analysis. We developed 8 occupancy models with two-levels of detection probability and four-levels of occupancy in Program MARK. We used an information-theoretical approach to identify the best supported models.
In 2010, occupancy of 6 species was more associated with local scale variables (gradient of vegetation cover of hardwood, softwood, and herb) and 5 species with 500m- or 1km-landscape variables (gradient of urban development, agriculture, and hardwood and mixed forest). Occupancy of 2 species was not better explained either by local scale variables or 500m- or 1km- landscape scale variables. In 2011, occupancy of 1 species related to local scale variables and 5 species to 500m- or 1km-landscape scale variables; however, occupancy of 1 species was not associated with either of 3 scales. Although avian response to local-scale or 500m- or 1km- landscape scale variables were species-specific, over half of forest interior species and forest edge species tended to be more associated with landscape-scale variables. Most pine-grassland species responded to local-scale variables. Our results showed that landscape-scale habitat features can be important in conservation management plans of some avian species. In particular, when target species for conservation is forest interior or forest edge species, landscape-scale habitat features should be considered.