Importance of the bacterial dynamics in model simulations of seasonal hypoxia

The occurrence and spread of hypoxia in coastal waters is known to depend strongly on nutrients, primary production, water column structure, wind and tidal mixing. Accurate prediction of the onset, intensity and areal extent of hypoxia remains a challenge. Previous modeling efforts have needed to "tune" vertical mixing or phytoplankton respiration in order to obtain results that agree with field observations of dissolved oxygen (DO). In this study, we use a one-dimensional physical model coupled with a biogeochemical model to establish mechanistic links between factors involved in the evolution of seasonal hypoxia in western Long Island Sound. The coupled model includes bacterial dynamics, which allows accurate prediction of the onset of late summer hypoxia and subsequent recovery. Model results indicate that a hyperbolic temperature response function represents temperature control on bacterial community growth rate better than the traditional Q₁₀ function. We also find that onset of hypoxia's late summer relaxation is sensitive to factors beyond wind mixing alone. Wind mixing contributes at most 30% to the late summer DO variability. The cost of not including the bacterial dynamics in the model is an overestimation of bottom DO by as much as 700% in summer. By including bacterial dynamics, we eliminate the need to distort vertical mixing or phytoplankton respiration rate to simulate observed seasonal variability in DO. In order to accurately model DO dynamics under normal wind forcing, therefore, coastal ecosystem models need to explicitly include terms for major components of the microbial loop.