Room: 5-314
Speaker Name:
Bianca Champenois
Affiliation:
Sand-lab Doctoral Student
Abstract:
A significant portion of atmospheric CO2 emissions is absorbed by the ocean, resulting in acidified seawater and altered carbonate composition that is harmful to marine life and ecosystems. Despite detrimental effects, monitoring and forecasting indicators of ocean and coastal acidification (OCA) is difficult due to the scarcity of in-situ measurements and the high costs of computational modeling. We develop a parsimonious data-driven framework to model properties that drive OCA, and we test the framework in the Massachusetts Bay and Stellwagen Bank, a region with industries affected by OCA. First, we trained a neural network to predict in-depth fields for temperature and salinity (x,y,z) using surface quantities from satellites and in-situ measurements (x,y). The relationship between 2D surface and 3D properties is captured through the in-depth modes and coefficients obtained from principal component analysis applied to a high-resolution historical reanalysis data set. Next, we used Bayesian regression methods to estimate region-specific relationships for in-depth total alkalinity (TAlk), dissolved inorganic carbon (DIC), and aragonite saturation state (Omega_Ar) as a function of in-depth temperature, in-depth salinity, and surface chlorophyll-a concentration. Lastly, 4D field predictions are made from surface measurements using the neural network followed by the regression models. The model’s performance is evaluated using withheld measurements at multiple depths, locations, and seasons, and the near real-time predictions for temperature, salinity, TAlk, DIC, and Omega_Ar are useful for understanding the impacts and evolution of OCA. Each step of the framework includes uncertainty quantification which can be used for future planning and optimal sensor placement.