Run-off from agricultural and urban areas pollutes many water bodies in the US, such as in the Chesapeake Bay, driving harmful algal blooms that impact water quality and threaten public health. One major consequence of nutrient pollution is the development of oxygen-free (anoxic) or reduced oxygen (hypoxic) dead-zones that deteriorate the habitat for many aquatic animals. Although inhospitable for aquatic animals, the dead-zone hosts many unique microbial processes. Unfortunately, observations suggest that the volume of hypoxic water generated per unit of total nitrogen has doubled in the past 25 years. Microbial processes may be responsible for exacerbating the size of dead-zone, through more efficient nutrient recycling, yet current biogeochemical models do not include many microbial processes known to occur in anoxic environments. When are more complex models warranted? How can we validate complex processes when they are added to the model? We propose to complement biogeochemical model development with sequenced-based observations of the microbial community in the Chesapeake Bay to improve predictions of the size and duration of the dead-zone.
We surveyed the microbial community throughout the dead-zone of the Chesapeake Bay with high-throughput amplicon sequencing. We found microbial community structure is responsive to biogeochemical changes in the Bay, but that dispersal can act to homogenize the community in distinct biogeochemical zones. Dissolved oxygen levels remained relatively high in June 2015, while July and August samples displayed hypoxia below 7 m and 12 m and became anoxic below 9 m and 15 m, respectively. Despite differences in water chemistry along depth profiles, bacterial communities were more similar within a sample time point than between time points. Overall, bacterial communities were part of 11 clusters that segregated by sample month and depth. Only 5% of all OTUs were observed in each of the 11 clusters; yet, these persistent OTUs comprised about 88% of the bacterial community in each cluster. Thus, the majority of the autochthonous community was persistent and distributed throughout the water column even during stratification. Comparing the dynamics of these persistent OTUs revealed four distinct subpopulations that were most strongly correlated with different environmental parameters. We are evaluating the best way to integrate amplicon and shotgun metagenomic data into a biogeochemical model (ChesROMS) of the Chesapeake Bay dead-zone.