loading page

Quantifying errors in observationally-based estimates of ocean carbon sink variability
  • +9
  • Lucas Gloege,
  • Peter Landschützer,
  • Amanda Fay,
  • Thomas Frölicher,
  • John Fyfe,
  • Tatiana Ilyina,
  • Steve Jones,
  • Nicole Lovenduski,
  • Christian Rödenbeck,
  • Keith Rodgers,
  • Sarah Schlunegger,
  • Yohei Takano
Lucas Gloege
Columbia University, Columbia University

Corresponding Author:gloege@ldeo.columbia.edu

Author Profile
Peter Landschützer
Max Planck Institute for Meteorology, Max Planck Institute for Meteorology
Author Profile
Amanda Fay
Columbia University, Columbia University
Author Profile
Thomas Frölicher
Climate and Environmental Physics, Climate and Environmental Physics
Author Profile
John Fyfe
Environment and Climate Change Canada, Environment and Climate Change Canada
Author Profile
Tatiana Ilyina
Max Planck Institute for Meteorology, Max Planck Institute for Meteorology
Author Profile
Steve Jones
University of Bergen, University of Bergen
Author Profile
Nicole Lovenduski
University of Colorado, University of Colorado
Author Profile
Christian Rödenbeck
Max Planck Institute for Biogeochemistry, Max Planck Institute for Biogeochemistry
Author Profile
Keith Rodgers
Center for Climate Physics, Center for Climate Physics, Institute for Basic Science
Author Profile
Sarah Schlunegger
Princeton University, Princeton, Princeton University, Princeton
Author Profile
Yohei Takano
Max Planck Institute for Meteorology, Max Planck Institute for Meteorology
Author Profile

Abstract

Reducing uncertainty in the global carbon budget requires better quantification of ocean CO2 uptake and its temporal variability. Several methodologies for reconstructing air-sea CO2 exchange from sparse pCO2 observations indicate larger decadal variability than estimated using ocean models. We assess these reconstructions’ ability to estimate spatiotemporal variability, by creating the Large Ensemble Testbed using four independent Earth system models. Model pCO2 fields are subsampled as the observations, for each of 100 ensemble members, and the reconstruction is performed as is done with real-world observations. The power of a testbed is that the perfect reconstruction is known from the original model fields; thus, reconstruction skill can be comprehensively assessed. We find that a commonly used neural-network approach can skillfully reconstruct air-sea CO2 fluxes when and where it is trained with sufficient data. Flux bias is low for the global mean and Northern Hemisphere, but can be regionally high in the Southern Hemisphere. The phase and amplitude of the seasonal cycle are accurately reconstructed outside of the tropics, but longer-term variations are reconstructed with only moderate skill. For Southern Ocean decadal variability, insufficient sampling leads to a 39% [15%:58%, interquartile range] overestimation of amplitude, and phasing is only moderately correlated with known truth (r=0.54 [0.46:0.63]). Globally, the amplitude of decadal variability is overestimated by 21% [3%:34%]. Machine learning, when supplied with sufficient data, can skillfully reconstruct ocean properties. However, data sparsity remains a fundamental limitation to quantification of decadal variability in the ocean carbon sink.
Apr 2021Published in Global Biogeochemical Cycles volume 35 issue 4. 10.1029/2020GB006788