Estimating habitat-specific abundance and behavior of several groundfishes using stationary stereo still cameras in the southern California Bight

Abstract The increasing use of underwater cameras to estimate fish abundance often does not account for the behavior of target species. These behaviors can affect detectability of fish and bias density estimates. This study estimated abundance and behavior of several rockfishes (Sebastes spp.) and lingcod (Ophiodon elongatus) at Footprint Bank, a small offshore bank using images from randomly deployed stationary cameras. Deployments collected images at 30-second intervals over ∼24 h periods to examine the behaviors of rockfish that might impact abundance estimates. The results showed that time of day and tidal change had a significant effect on the probability of presence, estimated abundance and species composition of fish, with densities highest for most species during daylight hours. The time elapsed since camera deployment did not have a significant effect on fish density. Fish density was significantly affected by habitat composition, an effect primarily driven by speckled rockfish (Sebastes ovalis) which exhibited a 5-fold increase in abundance in bedrock habitats. Speckled rockfish were the most abundant rockfish at depths less than 150 m, with an estimated abundance of 12,994 fish (SE = 6,722) on Footprint Bank. The abundance estimates and coefficients of variation were comparable to surveys conducted in 2011 and 2012 using remotely operated vehicles (ROVs) and manned submersibles. The implications of this study are that habitat and behavior as well as timing of the survey (day/night) are important considerations determining the perceived density of fishes from underwater image surveys.