Quantification of Ocean Acidification and Flow Properties in the US North East Coast with Applications to Aquaculture Monitoring and Marine Debris Mitigation

Lead PI: Themistoklis Sapsis, Massachusetts Institute of Technology

The overarching goal of this proposal is to develop a regional ocean flow-ocean acidification model that utilizes all the available sources of data, including historical data, and leverages physical laws, possibly in simplified form, that can be used for: (i) monitoring and forecasting ocean flow and currents (ii) improve monitoring and forecasting capabilities of OA QoIs and (iii) predict marine debris position and support cleaning efforts. Therefore, the objectives are focused around topics 2 (Ocean acidification) and 3 (Mitigating marine debris).

We plan to derive reduced order models in the form of variational autoencoders using reanalysis data for the ocean flow. Using this reduced order representation of the flow we will formulate a Bayesian inference scheme in terms of the wind fields, rivers discharge and available drifter information. With the inferred flow we will improve modeling of ocean acidification quantities as well as setup a probabilistic prediction scheme for ocean debris.

Currents and movement of water masses are key for anything at sea, from navigation to cleaning of debris and microplastics and to patterns of biodiversity. An efficient model for movement of water masses will not only improve our understanding for the above topics but will also allows to improve modeling of temperature and salinity fields, both key variables in ocean acidification modeling. Ocean acidification is a global problem but particularly exacerbated in New England and adjacent Nova Scotia shelf where coastal waters have reduced ”buffering capacity”.