Photosynthetically Driven Cycles Produce Extreme pCO2Variability in a Large Eelgrass Meadow and Readily Measured Proxies Can Be Used to Estimate These Changes

Declining ocean pH has spurred research into the effects of marine carbonate chemistry on a variety of organisms, but less work has focused on the potential role of organisms in changing local carbonate chemistry. It has been suggested that photosynthetic activity of macrophytes in coastal areas can decrease pCO₂, increase pH, and may provide areas of refuge for organisms sensitive to ocean acidification. To assess the effect of a large eelgrass meadow on water chemistry, discreet samples were collected hourly over several 24 hour cycles in Padilla Bay, Washington. Calculated pCO₂ ranged from less than 100 ppm to greater than 700 ppm, often over the course of only a few hours. Aragonite saturation, DIC and pH were also highly variable. In –situ sensors, including a YSI glass electrode, a custom built DuraFET sensor and a SeaFET sensor were co-deployed to provide a high frequency record of water chemistry over several months.

These data, (discrete samples and sensors) were used to develop a model that estimates pCO₂ for the summer season based on readily measured parameters. Tidal height, photosynthetically active radiation and pH can predict pCO₂ reasonably well in this environment. We compare the data from the 3 pH sensors and analyze the quality of data and predictions based on each one. A simple theoretical model shows that the large observed and modeled changes in pCO₂ and pH (up to 800 ppm CO₂ or 1 pH unit per day) match the magnitude of changes expected based on experimentally derived photosynthetic rates, measured light and water depth and that CO2 fluxes from gas exchange are expected to be small compared to photosynthetic fluxes in this environment. This study illustrates how eelgrass meadows do have the potential to create favorable carbonate chemistry, and demonstrates both the temporally variable nature of that effect and the possibility of better understanding when and how long it occurs through relatively simple modeling of the system.