Understanding the dominant spatial and temporal modes of variability in global sea surface partial pressure of carbon dioxide (pCO2) is necessary for teasing out the main mechanisms responsible for driving the marine carbon cycle. However, the existing global network of observation-based estimates of this important measure of air-sea carbon dioxide exchange are hampered by data gaps in both space and time.
Landschützer et al.  have developed a new 34-year time series of observationally based global surface ocean pCO2 that uses a sophisticated neural network based-clustering technique along with the application of physical drivers known to influence the sea surface carbonate system. The global multi-decadal pCO2 record is compared to representative indices of the large-scale climate variability, such as the El Nino-Southern Oscillation (ENSO), Southern Annular Mode (SAM), the Atlantic Meridional Oscillation and the Pacific Decadal Oscillation.
Thermal changes driven by ocean circulation and the biological response to the large-scale climate modes drive most regional variability in the global ocean pCO2, although in the North Atlantic the variability is mostly thermally driven, possibly due to solubility changes. The most dominant oscillation periods of the global pCO2 are somewhat patchy but still show large-scale coherent patterns (see figure above). Nonetheless, the authors recognize the shortcomings of resolving 10-year and longer cycles from a 34-year time series and emphasize the need for the ongoing coordination in the collection of surface pCO2 data in the global oceans.
Citation: Landschützer, P., Ilyina, T., & Lovenduski, N. S. . Detecting regional modes of variability in observation‐based surface ocean pCO2. Geophysical Research Letters, 46. https://doi.org/10.1029/2018GL081756
—Janet Sprintall, Editor, Geophysical Research Letters