Estimates of productivity and detection probabilities of breeding attempts in the sooty shearwater (Puffinus griseus), a burrow-nesting petrel

Monitoring of breeding success in burrow-nesting seabirds is problematic, owing to the difficulties of detecting occupants in complex burrow systems. We summarise 6 years of monitoring the breeding success of sooty shearwater (titi, muttonbird, Puffinus griseus) on two southern New Zealand islands, The Snares and Whenua Hou, with a portable infrared camera system. Breeding attempts were monitored three times during the breeding season, i.e. egg laying, hatching and. edging. Overall breeding success was calculated in two stages. First, we estimated breeding success for each island-site-year combination with a model that allowed for imperfect detection of an egg or chick and accounted for the proportion of the breeding season that was covered by monitoring. The resulting estimates for each island were then analysed with a linear model, to provide a single estimate for that island. Breeding success was found to be highly variable and non-synchronous between islands, with the average proportion of eggs successfully. edging on The Snares (0.35, 0.20-0.52; mean and 95% creditable interval) being considerably lower and more variable than that on Whenua Hou (0.76, 0.70-0.82). Probability of detecting a breeding attempt was higher on The Snares whereas correcting for the proportion of the season monitored had a variable effect, reducing The Snares and Whenua Hou estimates by 27% and 7% respectively. The implications of these findings with respect to the demographic modelling of burrow-nesting species are discussed.