Phytoplankton response to nutrients was examined with a 26-yr database from the Delaware Estuary. Biomassnormalized primary production did not increase linearly with increasing nutrient concentrations and instead showed saturation at comparatively low nutrient concentrations and decreased at high concentrations. To separate the effects of light availability and temperature on primary production from those of other environmental variables, we developed an empirical model of areal primary production. The model equation was obtained for the entire dataset and the effect of the residual variables was expressed as correction factors of observed primary production to the model estimates. The model accounted for 67% of variability of observed primary production overall, indicating that production of the estuary was mainly controlled by light availability and temperature. In contrast, a similar model applied to a Chesapeake Bay database had shown a poorer fit, indicating consistent light limitation in the Delaware Estuary and varying strengths of light and nutrient limitation in the Chesapeake Bay. The relationships between nutrients and correction factors for the Delaware Estuary showed that the model underestimates primary production at low and mid nutrient concentrations and overestimates it at high concentrations. The model fit and correction factors for five regions of the estuary indicate a high-nutrient, low-growth situation in the Delaware Estuary because of varying influences of light limitation, proportions of nutrients, and probably toxic contaminants in areas with large anthropogenic inputs-including high nutrients.Much of the classical thinking about eutrophication has arisen from lake studies. In lake systems, a relatively simple model appeared to successfully describe phytoplankton response to the effect of a single limiting nutrient, phosphate (Hutchinson 1969). In estuaries, early evaluation of eutrophication was made by Ketchum (1969) and was popularized by Jaworski in the Potomac River (Jaworski et al. 1972). During the evolution of conceptual models of estuarine eutrophication, cautions about oversimplification of this problem have been made for a number of years (e.g., Schindler 1981;Nixon and Pilson 1983;Vollenweider et al. 1992). However, there is still a tendency to think of eutrophication in estuarine waters in terms of a simple linear response to a single limiting nutrient: what Cloern (2001) has described as a Phase I model. To construct more complex models, Cloern addresses system-specific responses; complex ge-1 Corresponding author (kyoshi@udel.edu).
AcknowledgmentsNumerous graduate students, research associates, and postdoctoral researchers, as well as many undergraduate and high school student volunteers, have assisted over the years; we thank all of them for their contributions. We also thank James E. Cloern and two anonymous reviewers for their useful comments on the manuscript.