Predicting water quality effects on bay anchovy (Anchoa mitchilli) growth and production in Chesapeake Bay: linking water quality and individual-based fish models

Water quality in the Chesapeake Bay and the Patuxent River has decreased since the 1950s due to an increase in nutrient loadings. Increased nutrient loads have caused an increase in the extent and duration of hypoxic conditions. Restoration via large-scale reductions in nutrient loadings is now underway. How reducing nutrient loadings will affect water quality is well predicted; however the effect on fish is generally unknown as most water quality models do not include trophic levels higher than zooplankton. I combined two water quality models with bay anchovy models (Anchoa mitchilli) to examine the effects of changes in nutrient loadings on anchovy survival and growth. An individual-based predation model was statically linked to Patuxent River watershed land-use and water quality models, and used to simulate the effects of changes in watershed land-use, Chesapeake Bay boundary condition nutrient loadings, and water year types on the summertime survival of daily anchovy egg and larval cohorts. I found that changes in Patuxent watershed land-use had little effect on egg and larval survival, while reduced nutrient loadings at the Chesapeake Bay boundary condition increased egg survival but reduced larval survival in June. The second analysis dynamically coupled a spatially-explicit, individual-based population dynamics model of juvenile and adult anchovy to the 3-dimensional Chesapeake Bay water quality model. Growth rates of individual anchovy within water quality model cells were calculated using a bioenergetics equation. Zooplankton densities from the water quality model provided prey for anchovy consumption, and anchovy consumption was an additional mortality term on zooplankton. Anchovy mortality was size-dependent. Anchovy movement depended on water temperature, dissolved oxygen, and zooplankton concentrations. Multi-year simulations with fixed annual recruitment were performed under decreased, baseline, and increased nutrient loadings scenarios. Increasing nutrient loadings had small effects on survival, but increased anchovy growth and therefore biomass. Anchovy growth exhibited compensatory density dependence. The results of both analyses showed that anchovy responses to changed nutrient loadings were complex and depended on life stage. Full-life cycle, spatially-explicit population models that are dynamically coupled to water quality models are needed to truly predict the effects of changes in nutrient loadings on fish populations