Survival and reproduction in a biennially-breeding seabird : the Wandering Albatross

Abstract

I studied a long-term data set for Wandering Albatrosses Diomedea exulans collected at sub-Antarctic Marion Island to investigate adult survival, breeding biology and factors affecting reproductive success of this long-lived pelagic seabird. Until recently, the complex adult life cycle has caused biases in conventional mark-recapture modelling analyses that assume that recapture probability is equal for 'observable' and 'unobservable' states of the biennial adult life cycle. During 'sabbaticals' taken after successful and late-failed breeding attempts, the chances of resighting an adult Wandering Albatrosses are reduced, thus affecting estimated recapture rates. I applied a multi-state mark-recapture method to the colony breeding data from Marion Island (1987-2005). This method allows individual albatrosses to move through a network of breeding states according to previous breeding history using transition probabilities to estimate survival, breeding and breeding success probability. The best fit models in the multi-state analysis were those representing constant survival and independent breeding probabilities, suggesting that Wandering Albatrosses constrain reproductive investment to ensure long-term survival. I used the parameter estimates in a transition matrix to calculate the steady state equilibrium for the Wandering Albatross breeding population. For the first time I was able to estimate that approximately 3000 pairs of Wandering Albatrosses use Marion Island for nesting. I also describe the effects of age, sex and experience on Wandering Albatross reproductive success. Birds that were most successful began breeding at the mean age of first breeding and older and younger birds were less successful in 5/10 year periods after the onset of breeding. Disappearance from the population decreased with an increase of age of first breeding. I tested the selection and experience hypotheses on birds with 'complete' histories defined according to an imposed rule on recovery data. Males appeared to increase breeding success as a result of selective removal of low-quality individuals from the population (the selection hypothesis), whereas females appeared to improve their breeding success as a result of increased experience.