Multi-scale analysis of atmospheric and oceanic pCO2 time series and of their difference

The oceans play an important role in the carbon cycle by exchanging CO2 with the atmosphere. These exchanges correspond to the biological pump, where the ocean can be sink or source of atmospheric CO2. Our hypothesis is that CO2 concentration, either atmospheric or oceanic, are chemical tracers being strongly influence by turbulence: we thus study separately their dynamics, and also their difference which is giving indication of the direction of the air-sea CO2 flux.

For this we use a publicly available data set of pCO2 simultaneous measurements at high frequency (typically 3 hours time step) at 40 difference places around the globe, from surface buoys (Sutton et al. 2019). We consider here the scaling properties of these quantities in order to characterize their multi-scale fluctuations, which are considered in the framework of passive or active scalars in turbulence.  For each site, this is done by analyzing temperature, salinity, oceanic (pCO2sw), atmospheric (pCO2air) pCO2 and their difference $\delta = pCO2sw - pCO2air$. Power spectral density are estimated in Fourier space and using Hilbert spectral analysis, with adapted methodologies to take into account the missing data problem. Spectral slopes are recorded and are interpreted in relation with the local climatology, depth and other factors.