Modeling population effects of burrowing owls on ashy storm-petrels in relation to house mouse eradication on the Farallon Islands

Background/Question/Methods

Non-native house mice (Mus musculus) were introduced on the Farallon Islands in the 19^th century and are now at extremely high densities.  As a consequence, migratory burrowing owls (Athene cunicularia) now over-winter on the Farallon Islands during the fall and winter, feeding on the mice. However when the mice population crashes during the winter, owls turn to alternative prey, the ashy storm-petrel (Oceanodroma homochroa), which arrives in the winter in preparation for breeding.  The species is designated a Species of Management Concern by the USFWS and Endangered by IUCN, due to recent population declines, and thus high rates of predation by owls on the storm petrels are of concern.  We analyzed intensive field survey data for mice and owls conducted from 2000 to 2012, in conjunction with surveys of predation by owls on storm petrels.  We analyzed results of long-term standardized mist-netting of banded storm petrels to quantify population trends and determine survival rates of storm petrels in relation to variation in the abundance of owls using the program RMARK. We then developed population-dynamic models to project changes in future storm petrel population trends in relation to potential changes in owl abundance as a result of the proposed mouse eradication.

Results/Conclusions

Burrowing owl abundance and predation on storm petrels have significantly increased in recent years.  Predation on storm petrels is positively related to owl abundance and negatively related to mouse abundance, confirming that storm petrels are the owl’s secondary prey. The best-fitting model for the change in storm petrel population size indicated a decline from 2007 to 2012 (at 5.91% /yr; P = 0.050), coinciding with the recent increase in owl numbers. Capture-recapture analyses of storm petrels revealed a significant decrement in annual survival as the owl abundance index increased (P = 0.009).  One owl present on the island for 8 months decreases storm petrel annual survival rate an estimated 1%, a result confirmed through empirical observations of carcasses.  Population-dynamic models projected that under current conditions the storm petrel population will decrease by 5.9% per year. However, a 50% reduction in owl abundance (and associated predation), due to proposed mouse eradication, would eliminate most of the storm petrel decline; reducing owl abundance by more than 70% would produce a nearly stable population.  Thus, reducing owl abundance, through elimination of their primary prey, will likely have a long term effect, promoting stable future population trends for this species of high concern.