Modeling Forest Bird Species' Likelihood of Occurrence in Utah With Forest Inventory and Analysis and Landfire Map Products and Ecologically Based Pseudo-Absence Points

Phoebe L. Zarnetske
Thomas C. Edwards, Jr.
Gretchen G. Moisen


Estimating species likelihood of occurrence across extensive landscapes is a powerful management tool. Unfortunately, available occurrence data for landscape-scale modeling is often lacking and usually only in the form of observed presences. Ecologically based pseudo-absence points were generated from within habitat envelopes to accompany presence-only data in habitat classification models (HCM) for the northern goshawk (Accipiter gentilis atricapillus). We built models at two resolutions, using predictor variables derived from 250-m Forest Inventory and Analysis map products, 30-m U.S. Department of Agriculture Landfire map products, and digital elevation models. Cross-validation provided an assessment of models' predictive capabilities. Use of ecologically based pseudo-absence points to accompany extant presence points in HCM can be a powerful asset for species conservation.