SEARCH

SEARCH BY CITATION

Keywords:

  • burned coniferous forest;
  • cavity-nesting birds;
  • Colaptes auratus;
  • habitat suitability;
  • Melanerpes lewis;
  • nesting habitat;
  • Picoides spp.;
  • receiver operating curves;
  • remotely sensed data;
  • Sialia spp

ABSTRACT  Models of habitat suitability in postfire landscapes are needed by land managers to make timely decisions regarding postfire timber harvest and other management activities. Many species of cavity-nesting birds are dependent on postfire landscapes for breeding and other aspects of their life history and are responsive to postfire management activities (e.g., timber harvest). In addition, several cavity nesters are designated as species at risk. We compare the ability of 2 types of models to distinguish between nest and non-nest locations of 6 cavity-nesting bird species (Lewis's woodpecker [Melanerpes lewis], black-backed woodpecker [Picoides arcticus], hairy woodpecker [P. villosus], northern flicker [Colaptes auratus], western bluebird [Sialia mexicana], and mountain bluebird [S. currucoides]) in the early postfire years for a ponderosa pine (Pinus ponderosa) forest in Idaho, USA. The 2 model sets consisted of 1) models based on readily available remotely sensed data and 2) models containing field-collected data in addition to remotely sensed data (combination models). We evaluated models of nesting habitat by quantifying the model's ability to correctly identify nest and non-nest locations and by determining the percentage of correctly identified nest locations. Additionally, we developed relative habitat-suitability maps for nesting habitat of black-backed and Lewis's woodpeckers from the best models. For all species except Lewis's woodpeckers, model performance improved with the addition of field-collected data. Models containing remotely sensed data adequately distinguished between nest and non-nest locations for black-backed woodpecker and Lewis's woodpecker only, whereas models containing both field-collected and remotely sensed data were adequate for all 6 species. Improvements in the availability of more accurate remote sensing technology would likely lead to improvements in the ability of the models to predict nesting locations. External validation with data from other wildfires is necessary to confirm the general applicability of our habitat-suitability models to other forests. Land managers responsible for maintaining habitat for cavity-nesting birds in postfire landscapes can use these models to identify potential nesting areas for these species and select areas in burned forests where postfire salvage logging is most likely to have minimal impacts on cavity-nesting bird habitats.