The Statistical Estimation of Tuna Consumption Rates Using Coupled Bioenergetics and Mercury Mass Balance Models

The quantification of fish consumption rates can give us insight into predator-prey interactions and the environmental and physiological constraints on growth.  However, fish consumption rates are difficult to measure directly, due to the extensive sampling required to account for spatial and temporal variation. We developed a novel approach to generate consumption rates by applying statistical estimation methods to coupled bioenergetics – mercury (Hg) mass balance models.  Specifically, we created coupled models for bigeye (Thunnus obesus), yellowfin (Thunnus albacares), skipjack (Katsuwonus pelamis), and albacore (Thunnus alalunga) tunas and evaluated whether key parameters, including swimming speed and concentrations of Hg in prey species, might be estimated by fitting the models to Hg-at-size data.  Coupling the models allowed us to overcome limitations of each individual model, providing statistical estimates for unknown bioenergetics parameters and linking contaminant bioaccumulation to bioenergetics processes.  Our approach was successful in generating biologically plausible consumption rates and swimming speeds when the estimation procedures were restricted to estimating only one parameter.  The precision surrounding our consumption rates was greatly reduced when estimating three parameters in the models due high covariance in the estimates of swimming speeds and prey Hg parameters, and the absence of highly informative Hg-at-size data to resolve that covariation.  Statistical estimation is a promising tool to generate fish consumption rates and can potentially be improved by using more informative data (including multiple sources of data) and applying Bayesian approach to make better use of pre-existing data to inform and constrain model estimations.