Relating trap capture to abundance: a hierarchical state-space model applied to black sea bass (Centropristis striata)

Traps are among the most common gears used to capture fish and crustaceans. When traps are deployed in surveys, the data gathered are often used to develop an index of abundance. However, trap catches are known to saturate over time for various reasons, such as space limitation of the gear or intraspecific interactions, and these features can dissociate the catch from local abundance. In this study, we develop a hierarchical state-space model of trap dynamics that is fit to data in a Bayesian framework. The model links trap catch to estimated local abundance, and additionally provides direct estimates of capture probability. For demonstration, we apply the model to data on black sea bass ( Centropristis striata ), which were collected using chevron traps combined with video cameras to give continuous-time observations of trap entries and exits. Results are consistent with the hypothesis that trap catch is generally proportional to local abundance. The model has potential application to surveys where animals not only enter a trap, but also may exit, such that the apparent trap saturation occurs because the system approaches equilibrium.