Modeling the population-level effects of hypoxia on a coastal fish: implications of a spatially-explicit individual-based model

The northwestern Gulf of Mexico (USA) currently experiences a large hypoxic area (“dead zone”) during the summer. The population-level effects of hypoxia on coastal fish are largely unknown. We developed a spatially-explicit, individual-based model to analyze how hypoxia effects on reproduction, growth, and mortality of individual Atlantic croaker could lead to population-level responses. The model follows the hourly growth, mortality, reproduction, and movement of individuals on a 300 x 800 spatial grid of 1 km² cells for 140 years. Chlorophyll-a concentration and water temperature were specified daily for each grid cell. Dissolved oxygen (DO) was obtained from a 3-D water quality model for four years that differed in their severity of hypoxia. A bioenergetics model was used to represent growth, mortality was assumed stage- and age-dependent, and movement behavior was based on temperature preferences and avoidance of low DO. Hypoxia effects were imposed using exposure-effects sub-models that converted time-varying exposure to DO to reductions in growth and fecundity, and increases in mortality. Using sequences of mild, intermediate, and severe hypoxia years, the model predicted a 20% decrease in population abundance. Additional simulations were performed under the assumption that river-based nutrients loadings that lead to more hypoxia also lead to higher primary production and more food for croaker. Twenty-five percent and 50% nutrient reduction scenarios were simulated by adjusting the cholorphyll-a concentrations used as food proxy for the croaker. We then incrementally increased the DO concentrations to determine how much hypoxia would need to be reduced to offset the lower food production resulting from reduced nutrients. We discuss the generality of our results, the hidden effects of hypoxia on fish, and our overall strategy of combining laboratory and field studies with modeling to produce robust predictions of population responses to stressors under dynamic and multi-stressor conditions.