mCDR 2023: Data requirements for quantifying natural variability and the background ocean carbon sink in marine carbon dioxide removal (mCDR) models

In order to combat the expected impacts of climate change, active removal of carbon dioxide (CO2) from the atmosphere and oceans will likely be needed in addition to emissions reductions. There is a growing consensus that at least some of society’s needs for carbon dioxide removal (CDR) will have to come from the ocean. An important requirement of any ocean CDR strategy is the ability to demonstrate that it has “worked,” i.e., that it has resulted in the uptake of additional CO2 from the atmosphere beyond what is already naturally occurring due to rising atmospheric CO2 concentrations, termed here the “background” ocean carbon sink. Ocean models are expected to play a key role in this effort. In the interest of validating these models, this project will determine the natural background carbon uptake, its variability, and the degree of certainty with which it is known, in areas of the ocean where CDR deployments are likely to take place. Requirements for additional sampling needed to improve understanding of the background ocean carbon sink and to confidently measure the additional signal from CDR will be determined. This work will support future observing system development, and ultimately the future development of observation-based benchmarks against which proposed marine CDR models can be evaluated. The project will provide salary support to an early career researcher to become an expert in ocean carbon cycling and machine learning, skills critical to ocean science and the marine CDR (mCDR) workforce. This project is being jointly supported by the National Oceanic and Atmospheric Administration, through the National Oceanographic Partnership Program.The objectives of this project are to 1) quantify uncertainties in air-sea CO2 flux variability and the integrated background ocean carbon sink on regional scales, and 2) set requirements for additional data collection that will reduce these uncertainties. Following successful prior work at the global scale, these objectives will be achieved by developing and applying a ‘testbed’. This testbed will be a high-resolution (1/10°) ocean model that will be sampled with the spatio-temporal pattern of existing surface pCO2 observations in regions on the West and East US Coast, Hawaii and the Bering Sea. Machine learning reconstructions will be performed based on these samples to reconstruct full field, time-varying pCO2. The unique advantage of a testbed is that the fidelity of the reconstructions can be evaluated based on comparison to the original full model fields. This approach allows for assessment of how well sparse data and state-of-the-art machine learning techniques can be combined to constrain surface ocean carbon fluxes. In a second phase, observing system simulation experiments (OSSEs) will establish optimal observing designs that can further reduce reconstruction uncertainties.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.