Absolute abundance estimates from shallow water baited underwater camera surveys; a stochastic modelling approach tested against field data

摘要

Baited underwater cameras are becoming a popular tool to monitor fish and invertebrate populations within protected and inshore environments where trawl surveys are unsuitable. Modelling the arrival times of deep-sea grenadiers using an inverse square relationship has enabled abundance estimates, comparable to those from bottom trawl surveys, to be gathered from deep-sea baited camera surveys. Baited underwater camera systems in the shallow water environments are however, currently limited to relative comparisons of assemblages based on simple metrics such as Max(N) (maximum number of fish seen at any one time). This study describes a stochastic simulation approach used to model the behaviour of fish and invertebrates around a BUC system to enable absolute abundance estimates to be generated from arrival patterns. Species-specific models were developed for the tropical reef fishes the black tip grouper (Epinephelus fasciatus) and moray eel (Gymnothorax spp.) and the Antarctic scavengers: the asteroid (Odontaster yalidus) and the nemertean worm (Parbolasia corrugatus). A sensitivity analysis explored the impact of input parameters on the arrival patterns (MaxN, time to the arrival of the first individual and the time to reach MaxN) for each species generated by the model. Sensitivity analysis showed a particularly strong link between MaxN and abundance indicating that this model could be used to generate absolute abundances from existing or future MaxN data. It in effect allows the slope of the MaxN vs. abundance relationship to be estimated. Arrival patterns generated by each model were used to estimate population abundance for the focal species and these estimates were compared to data from underwater visual census transects. Using a Bland-Altman analysis, baited underwater camera data processed using this model were shown to generate absolute abundance estimates that were comparable to underwater visual census data. (C) 2015 Elsevier B.V. All rights reserved.