Net Sea–Air CO$_{2}$ Fluxes and Modeled Partial Pressure of CO$_{2}$ in Open Ocean of Bay of Bengal

Abstract

The key role of oceans in the global climatic system can be accurately quantified by the measurement of CO $_{2}$ exchange across air–sea interface. The direction and magnitude of this CO $_{2}$ exchange are mainly governed by the gradient of partial pressure of carbon dioxide (pCO $_{2}$ ) at the air–sea interface. The sea surface pCO $_{2}$ is highly variable, primarily regulated by seasonal sea surface temperature (SST) and CO $_{2}$ (aqueous) concentration. In this paper, the moored pCO $_{2}$ observations from autonomous pCO $_{2}$ system deployed at 15 $^{\circ }$ N 90 $^{\circ }$ E in the Bay of Bengal (BOB) were used to derive CO $_{2}$ fluxes to examine variability over time scale at BOB. An attempt has also been made to develop a model using multiple linear regression (MLR) and support vector regression (SVR) to estimate the sea surface pCO $_{2}$ from space-based observations of SST and salinity. pCO $_{2}$ variability was primarily attributed to SST and CO $_{2}$ concentration changes. Derived fluxes showed that BOB is a net annual source (0.440 g-C/m $^{2}$ /year for year 2014) of CO $_{2}$ at mooring location with significant seasonality effects. For the estimation of surface pCO $_{2}$ , SVR model showed better results [root mean square error (RMSE) = 7.68 $\mu$ atm] in comparison to MLR model (RMSE = 12.36 $\mu$ atm). The seasonal climatological pCO $_{2}$ maps from September 2011 to February 2018 showed southward increase in pCO $_{2}$ in all seasons possibly due to southward increase in SST and less influence of river water. These pCO $_{2}$ maps can further be utilized in deriving CO $_{2}$ flux maps for BOB.