• capture–mark–recapture;
  • Fiordland;
  • hydroelectric power;
  • information theory;
  • Leslie matrix model;
  • population viability;
  • tourism impacts;
  • Tursiops sp


  • 1.
    The bottlenose dolphins of Doubtful Sound, New Zealand are a declining population at the southern limit of the species' range, exposed to impacts from tourism and habitat modification. Patterns in apparent annual survival were analysed from photographic resightings of naturally marked adults (1990 to 2008) and calves within the first year of life (1994 to 2008) using capture-recapture models.
  • 2.
    The most parsimonious model for adults provided a time-invariant, sex-invariant estimate of survival (ϕa(1990–2008)=0.9374; 95% CI: 0.9170–0.9530), marginally lower than prior estimates for wild bottlenose dolphins.
  • 3.
    The most parsimonious model for calves indicated a significant time-variant decline in survival from an estimate similar to other populations (ϕc(1994–2001)=0.8621; 95% CI: 0.6851–0.9473) to a current estimate that is, to our knowledge, the lowest recorded for free-ranging bottlenose dolphins (ϕc(2002–2008)=0.3750; 95% CI: 0.2080–0.5782).
  • 4.
    Information theoretic evidence ratios suggested that observed patterns in calf survival were 22 times more likely to be explained by a decline coincident with the opening of a second tailrace tunnel for a hydroelectric power station than by a decline in any other year or across multiple years.
  • 5.
    Projections using an age-structured stochastic population model indicated that the current level of calf survival was unsustainable (population decline: 100% of model runs; population extinction: 41.5% of model runs) and was a key factor in the observed population decline in Doubtful Sound.

Copyright © 2008 John Wiley & Sons, Ltd.