Predictive Approaches to Habitat Quantification: Dark-Rumped Petrels on Haleakala, Maui

Brandt, C; Parrish, J; Hodges, Cathleen

Physical characteristics of the nesting habitat of the endangered Hawaiian Dark-rumped Petrel (Pterodroma phaeopygia sandwichensis), nesting on Haleakala volcano, Maui, were quantified in order to determine whether birds were choosing nesting sites nonrandomly. Nesting habitat was sampled at three spatial scales: burrow (within immediate vicinity of a petrel burrow), colony (10-60 m from a petrel burrow), and landscape (1,000-m transects through habitat containing colonies). We used a probability model based on logistic regression to discriminate among the three sampling scales. The model was quite successful in reclassifying the locations, correctly identifying over 91% of the burrow sites as such. In general, burrows were located on steep slopes under large rocks in the vicinity of shrub cover. This set of characteristics was not indicative of habitat at either the colony or landscape scales. Our study reveals two important aspects in evaluating habitat preferences for the Dark-rumped Petrel population on Haleakala: (1) habitat must be evaluated multidimensionally, and (2) important aspects of habitat associations may become apparent only at certain scales. The nesting range of the Dark-rumped Petrel is currently severely restricted from historical levels to extreme elevations within Haleakala National Park. Probability models such as ours can be useful in delineating suitable habitat patches where, in this case, Dark-rumped Petrel burrows have a higher probability of being found. This approach is powerful because it can make use of remotely sensed and stored data sets to direct and refine on-the-ground searches.

Predictive Approaches to Habitat Quantification: Dark-Rumped Petrels on Haleakala, Maui

Series