Abstract
•Uncertain parameters limit the ability of models to reliably inform management.•The diet matrix of a West Florida Shelf Ecopath model was highly uncertain.•A more reliable diet matrix was constructed using a robust statistical model.•The revised diet matrix increased complexity and omnivory in the Ecopath model.•It also resulted in more realistic predation mortality rates for many fish groups.
Harmful algal blooms (HABs) are a growing concern in the West Florida Shelf (WFS) region. An Ecopath with Ecosim (EwE) model of the WFS explicitly simulating HABs was previously developed to illuminate the potential impacts of blooms of the dinoflagellate Karenia brevis (colloquially referred to as “red tides”) on the WFS ecosystem. However, the diet matrix of the Ecopath component of this EwE model (referred to as “WFS-HAB Ecopath”) was based largely on sparse, cursory information and not on local survey data. Here, we revise the diet matrix of the WFS-HAB Ecopath model using predictions of a robust statistical model that incorporates local survey data and employs the Dirichlet distribution and maximum likelihood estimation. The relative impacts of both the revised diet matrix and red tide mortality scenarios on model structure are explored by comparing four alternative WFS-HAB Ecopath models: (i) the base model; (ii) a model employing the revised diet matrix; (iii) a model with elevated red tide mortality; and (iv) a model with both the revised diet matrix and elevated red tide mortality. Incorporating the revised diet matrix into the WFS-HAB Ecopath model had a relatively large impact on ecosystem structure (i.e., trophic organization, mortality rates, trophic interaction strengths, and omnivory). Elevated red tide mortality had virtually no impact on ecosystem structure aside from altering the contribution of fishing, natural, and red tide mortalities to the total mortality of functional groups; however, elevated red tide mortality might have meaningful implications in dynamic simulations, which should be explored in future studies. Collectively, results showed that incorporating the revised diet matrix into WFS-HAB Ecopath, which revealed a number of new predator-prey linkages, led to a more complex and interconnected food web. Specifically, prey items were generally consumed by a broader variety of predators, which contrasts with the base WFS-HAB Ecopath model where many prey, particularly juvenile fishes, were subjected to exceedingly high predation mortality rates from specific predators. The incorporation of the revised diet matrix into the WFS-HAB Ecopath model discussed herein is a fundamental step towards increasing the realism of trophic interactions in the model, which is particularly important as these trophic interactions define starting conditions for dynamic simulations.
