Empirical Models for Estimating the Carbonate System Off the Northeastern US from Basic Hydrographic Data: An MLR Approach

The decline in pH of the global ocean is a predictable consequence of rising atmospheric carbon dioxide concentrations from human emissions, but in the coastal ocean, local processes can modulate or mask this trend. Nevertheless, constraining the carbon budget of the coastal ocean is especially worthwhile because the coastal ocean plays a disproportionately large role in the global carbon cycle and is home to several of the most vulnerable species to ocean acidification. Observations of carbonate system parameters are not spatially resolved enough nor are they frequent enough to capture regional trends in local processes off the northeastern coast of the U.S. Local carbonate chemistry is controlled by a balance of physical and biological processes, which are captured to a degree in observations of temperature, salinity, oxygen, and nitrate concentration. Here, a multiple linear regression (MLR) model has been developed to estimate carbonate system parameters total alkalinity (TA), dissolved inorganic carbon (DIC), pH, and aragonite saturation state (Omega) from measured temperature, salinity, oxygen, and nitrate off the northeastern coast of the US. Empirical models have been developed using this approach for other coastal regions, but because local processes are region-specific, different models are required for each region. Calibration data used here included measurements of TA and DIC collected during the Gulf of Mexico and East Coast Carbon cruises in July-August 2007 and 2012 and the East Coast Ocean Acidification cruise in June-July 2015. Resulting empirical relationships for TA, DIC, and are robust with > 0.96, but the variability in pH is not as well captured by the MLR = 0.89), which differs from the MLRs developed for other regions. A data set collected in the East Coast Ocean Acidification cruise in June-July 2018 was used to test the performance of the MLR for summer data. Comparing estimated and measured carbonate chemistry values indicates the MLR performs well with > 0.9 for TA, DIC, and , and > 0.7 for pH. These newly developed empirical models can be used to extend records in space and time, to reveal the regional carbonate system variability driven by local processes.