An empirical model using dissolved oxygen as an indicator for eutrophication at a regional scale

Simple empirical models can sometimes capture salient patterns without sacrificing predictive capacity when compared to more complex models. Herein we examine dissolved oxygen as an indicator of eutrophication status for shallow estuaries. Dissolved oxygen was measured hourly in the upper estuary of 15 watersheds along a nutrient-loading and geographic gradient. Metrics describing hypoxia and supersaturation were devised and then analyzed using multivariate statistics. Results revealed independent responses for hypoxia and supersaturation with hypoxia-related metrics correlating strongly with water residence. A metric integrating hypoxia and supersaturation effectively discriminated between seagrass and algae-dominated habitats and was significantly correlated with both water residence and nitrate-N loading. Chlorophyll, measured bi-weekly, was not correlated with our predictor variables likely because it does not account for benthic production. Over 70% of the variability in hypoxia was explained by water residence and nitrate-N loading indicating that this model can be of use for managers.