Pairing high-frequency data with a link-node model to manage dissolved oxygen impairment in a dredged estuary.

Abstract

High-frequency data and a link-node model were used to investigate the relative importance of mass loads of oxygen-demanding substances and channel geometry on recurrent low dissolved oxygen (DO) in the San Joaquin River Estuary in California. The model was calibrated using 6years of data. The calibrated model was then used to determine the significance of the following factors on low DO: excavation of the river to allow navigation of large vessels, non-point source pollution from the agricultural watershed, effluent from a wastewater treatment plant, and non-point source pollution from an urban area. An alternative metric for low DO, excess net oxygen demand (ENOD), was applied to better characterize DO impairment. Model results indicate that the dredged ship channel had the most significant effect on DO (62% fewer predicted hourly DO violations), followed by mass load inputs from the watershed (52% fewer predicted hourly DO violations). Model results suggest that elimination of any one factor will not completely resolve DO impairment and that continued use of supplemental aeration is warranted. Calculation of ENOD proved more informative than the sole use of DO. Application of the simple model allowed for interpretation of the extensive data collected. The current monitoring program could be enhanced by additional monitoring stations that would provide better volumetric estimates of low DO.