Capture–recapture models were used to provide estimates of abundance, apparent survival and temporary emigration of Indo-Pacific bottlenose dolphins (Tursiops sp.) in a 226-km2 study area off Useless Loop in the western gulf of Shark Bay, Western Australia. Photo-identification data were collected during boat-based surveys in Austral autumn to early spring (April–September) from 2007 to 2011. Abundance estimates varied from 115 (s.e. 5.2, 95% CI 105–126) individuals in 2008 to 208 (s.e. 17.3, 95% CI 177–245) individuals in 2010. The variability in abundance estimates is likely to be a reflection of how individuals used the study area, rather than fluctuations in true population size. The best fitting capture–recapture model suggested a random temporary emigration pattern and, when coupled with relatively high temporary emigration rates (0.33 (s.e. 0.07) – 0.66 (s.e. 0.05)) indicated that the study area did not cover the entire ranges of the photo-identified dolphins. Apparent survival rate is a product of true survival and permanent emigration and was estimated annually at 0.95 (s.e. 0.02). Since permanent emigration from the study area is unlikely, true survival was estimated to be close to 0.95. This study provides a robust baseline for future comparisons of dolphin demographics, which may be of importance should climate change or increasing anthropogenic activity affect this population.