Heterogeneous patterns of availability for detection during visual surveys: spatiotemporal variation in sea turtle dive–surfacing behaviour on a feeding ground

摘要: Summary 1. During aerial or boat-based surveys for large-bodied diving taxa (e.g. marine mammals and turtles), a proportion of animals present will be missed because they are submerged out view, leading to ‘availability bias’ in abundance indices. Information on dive–surfacing patterns can improve corrections availability bias. However, as dive data typically limited, correction factors often based poorly resolved surface times, heterogeneity is not considered. 2. We collected records green turtles Chelonia mydas, Linnaeus 1758, loggerhead Caretta caretta, foraging ground Shark Bay, Western Australia quantify assess potential correlations with easily measured environmental features: habitat depth water temperature. Bayesian regression models were used predict interval durations across temperature–depth gradients their uncertainty. We these predictions variation factors, which multipliers designed, this case, adjust sightings incorporate animals. 3. Dive both species varied positively negatively temperature, consistent priori expectations, although temperature effects always significant. Dive metrics predictable, uncertainty increased deeper few observed dives. 4. Availability highly heterogeneous, larger necessary colder, conditions (long-diving, infrequent surfacing behaviour) smaller required warmer, shallower (short-diving, frequent-surfacing behaviour). 5. Predictable the behaviour chelonid sea reveals that site-specific knowledge important mitigate bias during population surveys. Accounting such trends may reliability ecological inferences spatiotemporal distribution trends) efficacy applications conservation planning) survey data.