Greater nutrient loads since the 1950s have increased the extent and duration of hypoxic conditions in the Patuxent River and the Chesapeake Bay (Maryland, USA). We linked watershed, water quality, and individual-based predation models to predict how changes in local (Patuxent River watershed) and regional (bay-wide) nutrient loading rates would affect bay anchovy Anchoa mitchilli egg and larval mortality rates in the lower region of the Patuxent River. Nutrient loadings affected hypoxic volume and the degree of spatial overlap between anchovy and their predators. Mortality rates were simulated during June and July under combinations of a wet or dry year, reduced and increased nutrient loadings from the Patuxent River watershed, and increased and decreased nutrient loadings into the Patuxent River at the Chesapeake Bay boundary. Chesapeake Bay water quality at the downstream boundary had a much larger effect on egg and larval mortality rates than nutrient loading rates from the Patuxent River watershed, and these responses were consistent with the downstream boundary condition having a greater effect on hypoxia. Water column structure, year type (wet/dry), and location within the lower Patuxent River had smaller effects on egg and larval mortality. Due to indirect effects, the effect of the Chesapeake Bay boundary condition on larval mortality rates during June was opposite to that predicted for egg mortality rates. Our results illustrate that statements and justifications about the benefits of nutrient loading reductions on estuarine ecosystems should avoid oversimplification, be specific, and recognize that species responses to changes in environmental conditions can be complex and variable.