Multiscale habitat suitability index models for priority landbirds in the Central Hardwoods and West Gulf Coastal Plain/Ouachitas Bird Conservation Regions

Abstract

Ecoregional conservation planning for priority landbirds requires methods that explicitly link populations to habitat conditions at multiple scales. We developed Habitat Suitability Index (HSI) models to assess habitat quality for 40 priority bird species in the Central Hardwoods and West Gulf Coastal Plain/Ouachitas Bird Conservation Regions. The models incorporated both site and landscape environmental variables derived from one of six nationally consistent datasets: ecological subsections from the National Ecological Unit Hierarchy, National Land Cover Dataset, National Elevation Dataset, National Hydrography Dataset, State Soil Geographic Database, and Forest Inventory and Analysis data. We initially defi ned potential habitat for each species from unique landform, landcover, and successional age class combinations. Species-specifi c environmental variables identifi ed from the literature were used to refi ne initial habitat estimates. We verifi ed models by comparing subsection-level HSI scores and Breeding Bird Survey (BBS) abundance via Spearman rank correlation. To validate models, we developed generalized linear models that predicted BBS abundance as a function of HSI score and Bird Conservation Region. We considered models that included a signifi cant (P ≤ 0.100) positive coefficient on the BBS predictor to be valid and useful for conservation planning.

Brown-headed Nuthatch Occupancy in Central Florida and Its Relationship to Forest Type, Forest Structure, and the Presence of Red-cockaded Woodpeckers

“…In the ONF, occupancy and detection probabilities were not strongly influenced by hardwood basal area, but our results are consistent with those reported previously because hardwood basal area in the ONF was well below the maximum values the nuthatch tolerates elsewhere (4.6 m^ ha"'; Watts 1999, Tirpak et al 2009). Conversely, snag densities in the ONF were generally lower than those recommended by McComb et al (1986;>2.3 ha') and Tirpak et al (2009;>8 ha"')-However, we found that snag densities bore no strong relationship with \\i andp and suggest other factors had greater importance in this setting. Approximately 68% ofthe sites sampled in mesic flatwoods had estimated snag densities <2.3 ha"', while probabilities of detection and occupancy at these sites were high (> The Brown-headed Nuthatch exhibits limited dispersal in many settings Slater 2009, Haas et al 2010), so patch isolation coupled with variation in the nuthatch's abundance in different forest types could influence patch occupancy.…”

Comparison of statistical and theoretical habitat models for conservation planning: the benefit of ensemble prediction

“…This approach assumes that each variable is equally important to the species and that if any model variable is absent or does not have a value in the suitable range for the species (i.e., received a suitability value of 0) then the location is unsuitable (i.e., HSI value is 0). We sent the models to two to five experts on each species for review, and models were revised based on their comments (Tirpak et al 2009b). The models were applied to NLCD, FIA, NED, and ecological subsection boundary (Cleland et al 1997) data sets to depict the spatial configuration of suitable habitats (Table 1).…”

“…We used the FS models to calculate habitat suitability for each species on each FIA plot. To estimate habitat suitability we assembled FIA data tables for each state within Microsoft Access (Microsoft, Redmond, Washington, USA) and (1) derived forest structure variables for each FIA plot condition , (2) applied suitability relationship equations developed for those variables in the HSI model building process (above and Tirpak et al 2009b), (3) calculated the geometric mean of resulting suitability values across model variables for a species for each condition on each plot, and (4) used FIA area expansion factors (U.S. Department of Agriculture Forest Service 2010) to calculate an area-weighted relative suitability value across all plot conditions for counties and subsections. As with the HSI models, the FS models predict the relative suitability of habitat on a scale from 0 (nonhabitat) to 1 (optimal habitat).…”

Dynamic‐landscape metapopulation models predict complex response of wildlife populations to climate and landscape change

“…In addition, HSI models captured each species sensitivity to habitat patch size and the predominance of forest in the surrounding landscape (Tirpak et al. ). See Appendix for a full description of the HSI models.…”