Fig 1.
Bird survey sites on the Albemarle-Pamlico Peninsula in eastern North Carolina, USA shown according to BBS route, along with perimeters of all fires that burned between 2001 and 2014 (red) and the extent of publicly-owned forest within 1.5 m of sea level (blue).
Table 1.
Common names, 4-letter codes, and scientific names of the 91 bird species detected on the Albemarle-Pamlico Peninsula in eastern North Carolina, USA (2013–2015).
Table 2.
Stepwise logistic regression model selection procedure to determine the variables that best predicted ghost forest in eastern North Carolina (2014).
Fig 2.
Posterior median species richness (estimated number of species at a given site of the 56 species we analyzed), shown with 95% credible intervals (error bars) and linear relationship (blue line), as a function of canopy density (vegetation from 20 to 30 m) at 156 points in coastal forests of eastern North Carolina, USA (2013–2015).
Fig 3.
Posterior means for the effects of forest canopy and midstory vegetation density on occupancy of 56 species of birds detected in coastal forests of eastern North Carolina, USA between 2013 and 2015.
The farther above/below or left/right of the zero lines (solid lines), represent a stronger response to canopy density and midstory density, respectively. Codes for bird species are located according to the magnitude and direction of the corresponding posterior means and are colored according to known habitat associations (generalist = multiple habitat associations, marsh/shrub = low statured vegetation, mature forest = closed canopy forest, open forest = open canopy forest, snag = standing dead trees). Common and scientific names of birds are in Table 1.
Table 3.
The number of bird species whose posterior means were less than 0 (more likely to occur in saltwater affected forest) and greater than 0 (more likely to occur in unaffected forest) for each of 2 vegetation density covariates on occupancy in eastern North Carolina, USA (2013–2015).
Fig 4.
Posterior distributions of species-specific parameter estimates for each of 3 covariates (midstory density, canopy density, and burned/unburned) on occupancy probability for 56 species of forest birds detected in coastal forests in eastern North Carolina between 2013 and 2015.
Fig 5.
Estimated habitat losses (blue) and gains (green) for hooded warbler on the Albemarle-Pamlico Peninsula in eastern North Carolina, USA resulting from the LiDAR-derived changes in canopy (vegetation 20–30 m tall) and midstory (vegetation 4.5–10 m tall) density between 2001 and 2014.
Changes in occupancy probability (modeled using a multi-species, hierarchical occupancy model) less than -0.3 were assumed to equate to habitat loss, whereas increases greater than 0.3 were assumed to equate to habitat gain. Prescribed fires and wildfires that burned between 2001 and 2014 shown in red and unburned, publicly owned forests within 1.5 m of sea level are shown in brown. Inset maps show the location of the study area (left), a smaller scale view of the Pain’s Bay wildfire that burned in 2011 (middle), and a smaller scale view of a portion of the Alligator River State Game Land.
Fig 6.
Estimated habitat losses (blue) and gains (green) for northern bobwhite on the Albemarle-Pamlico Peninsula in eastern North Carolina, USA resulting from the LiDAR-derived changes in canopy (vegetation 20–30 m tall) and midstory (vegetation 4.5–10 m tall) density between 2001 and 2014.
Changes in occupancy probability (modeled using a multi-species, hierarchical occupancy model) less than -0.3 were assumed to equate to habitat loss, whereas increases greater than 0.3 were assumed to equate to habitat gain. Prescribed fires and wildfires that burned between 2001 and 2014 shown in red and unburned, publicly owned forests within 1.5 m of sea level are shown in brown. Inset maps show the location of the study area (left), a smaller scale view of the Pain’s Bay wildfire that burned in 2011 (middle), and a smaller scale view of a portion of the Alligator River State Game Land.
Table 4.
The estimated hectares of net habitat change from 2001 to 2014 for the 25 bird species whose occupancy was significantly predicted by at least one of the vegetation metrics related to ghost forest.