Release of constraints on nest-site selection in burrow-nesting petrels following invasive rat eradication

Abstract

Introduced mammals have been eradicated from many offshore islands around the world, removing predation pressure from burrow-nesting seabirds and other affected wildlife. Nest-site selection in procellariiform seabirds is mediated by nesting habitat characteristics and social information, although it is unclear if, or how, nest-site selection will affect post-eradication colony growth. Using a Bayesian hierarchical modeling approach we assessed how nest-site selection differs among burrow-nesting seabird colonies at different stages of recovery after Pacific rat (Rattus exulans) eradication. We compared nest-site selection in a community of seven procellariiform species among six offshore islands in northeastern New Zealand: four designated rat-free over a continuum within the last 26 years, an island which never had rats, and an island with rats present. We hypothesized that, immediately after eradication, birds would be constrained to nesting habitat where they were less vulnerable to predation, and as time since eradication increased birds would eventually spread to new habitat. We found a positive relationship between mean burrow density and time since rat eradication. Soil depth was the most important predictor of burrow presence, abundance, and occupancy in plots among islands, with more burrows found in deeper soil. We found that the relationships between habitat covariates and nest-site selection decreased with increasing time since eradication. The probability of a covariate having a significant effect on nest-site selection decreased with increasing time since eradication and decreasing variability in the covariate across an island. Our results suggest that the eradication of rodents reduced constraints on petrel nesting distribution and that nest-site selection in burrow-nesting petrels may be influenced by burrow density, where selection of particular nesting habitat characteristics may be relatively more important in small recovering populations. We conclude that colony expansion immediately after predator removal is complex, influenced by numerous interacting factors, but may be partly limited by the availability of suitable nesting habitat.