Quantifying atmospheric CO2 interannual variability driven by ocean from OCO-2 observations and air transport simulations

Abstract:

The ocean is the key natural sink in the global carbon cycle which regulates the atmospheric CO₂ level: integrated over the industrial period, it has taken up a substantial amount of anthropogenic CO₂ emissions cumulatively. Despite the importance of the ocean in providing a long-term sink for carbon, compared to terrestrial ecosystems, ocean carbon uptake is characterized by lower interannual variability (IAV), and thus leave a smaller imprint on atmospheric CO₂ growth rate variability on short timescales. Given advances both in atmospheric observations, including new space-based measurements that monitor column-mean CO₂ levels, X_CO2, over open ocean, and in methods to upscale sparse surface ocean observations to estimate continuous ocean fluxes fields, there are new opportunities to refine the understanding of the ocean influence on atmospheric CO₂ variability.

In this study, we simulate the atmospheric CO₂ mole fraction originating from ocean fluxes, with a focus on investigating the spatio-temporal characteristics of X_CO2 IAV owing to sea-air fluxes. The variability of simulated X_CO2 driven by ocean fluxes was analyzed in conjunction with observations of atmospheric X_CO2 from NASA’s OCO-2 satellite, which has been obtaining global data since September 2014. OCO-2 measures the column integrated mole fraction, X_CO2, so we vertically integrated atmosphere simulation output to make comparations with the observations. Both modeled and observed X_CO2 were averaged to 10° x10° spatial resolution and to monthly temporal resolution.

To represent the sea-air CO₂ gas exchange for the model simulations, we use fluxes from the Surface Ocean pCO₂ Mapping intercomparison (SOCOM), which are interpolated from sparse air-sea ΔpCO₂ observations through multiple mapping techniques. We use several different SOCOM products as lower boundary conditions in three-dimensional tagged CO₂ GEOS-Chem atmospheric transport simulations to identify the fingerprints of ocean subregions and the whole ocean on atmospheric CO₂ mole fraction. In addition to differentiating the imprints these products leave on the atmosphere, we quantified the effect of interannual variation in atmospheric dynamics by using a seasonal climatology of SOCOM fluxes in an atmosphere CO₂ transport simulation.

We found that the magnitude of X_CO2 IAV generated by the whole ocean varies regionally, with the maximum standard deviation over the 36-year simulation of up to 0.1 ppm. There were substantial differences in the magnitude of X_CO2 IAV resulting from different SOCOM flux products. Variability in the ocean fluxes, rather than patterns of atmospheric transport, plays the dominant role in the overall variability, although IAV in X_CO2 over some regions, especially the tropical Pacific, is strongly tied to IAV in atmospheric transport. Because the IAV in fluxes are coherent across subregions in the Southern Hemisphere, whereas the patterns of IAV differ across ocean basins in the Northern Hemisphere, X_CO2 IAV most strongly reflects Southern Hemisphere fluxes. Our analysis points toward observational requirements to identify signals from the ocean using OCO-2 observations.