Models predict the ocean carbon sink, currently slowing anthropogenic climate change by absorbing ~25% of human-derived CO2 emissions from the atmosphere, will become much less effective in the future.

In the Atlantic it is thought this will take the form of declining biological productivity caused by nutrients becoming scarcer, related to either a weaking in the northwards nutrient stream (through a slowed overturning circulation), and / or a strengthening in upper ocean stratification caused by warming temperatures. Our aim in C-Streams is to reveal exactly how the Earth system model projections lead to a weakening in the North Atlantic carbon sink, and the mechanisms that drive it, by utilising a multitude of models and configurations.

Surface currents in NEMO model at increasing resolutions. Credit: Andrew Yool, NOC


Runs from a suite of Earth System CMIP6 models, deployed using different projected emissions scenarios, will be analysed. Looking at the projections for the patterns of the air-sea CO2 flux, we’ll separate them into natural and anthropogenic components, and the area-integrated North Atlantic carbon sink, and connect to their underlying circulations and physical processes redistributing nutrients and carbon. We will reveal whether the models suggest a weakening carbon sink and provide the answer as to why that response occurs, such as via an upstream advective control.

1994-2015 mean NPP in CMIP6 models (left). Change in NPP up to 2080-2099 mean under SSP1-2.6 (middle) and SSP5-8.5 (right). Credit: Kwiatkowski et al. (2020), Biogeosciences. <a href=""></a>


C-Streams will also interrogate two custom configurations of the MIT general circulation model to understand how physical processes alter the Gulf Stream nutrient and carbon streams, and then alter the downstream nutrient and carbon distributions. These will include a global model with higher resolution Gulf Stream and North Atlantic region, and a biogeochemical variant of the ECCOv4 state estimate that is both physically-consistent and observationally-constrained. Experiments will look at the effects of wind-stress and buoyancy forcing (associated with different weather regimes), as well as additional surface heating, on gyre/overturning circulation strength, and their effect on the transports and pathways of volume, nutrients and carbon for the Gulf Stream along the western boundary. We’ll then assess how these vary with changes in the extent of diapycnal mixing and isopycnal eddy exchanges.


Adjoint models provide an efficient method to identify the sensitivity of a quantity of interest (e.g. stratification, meridional overturning strength, area-integrated air-sea CO2 flux) to initial conditions, boundary conditions, and surface forcing (e.g. wind stress, heat flux). Using the adjoint machinery of ECCOv4, in C-Streams we will calculate the local and upstream sensitivities of stratification, surface carbon concentrations, and air- sea carbon flux of selected target regions in the seasonally stratified subpolar North Atlantic, especially in those locations where the air-sea CO2 flux is large, through a combination of adjoint sensitivity experiments and targeted perturbation experiments. This analysis will reveal the upstream physical controls of surface winds and surface heat fluxes, diapycnal mixing, and sub-grid scale mesoscale eddy transport coefficients on downstream surface carbon concentrations.

ECCOv4 adjoint models will be used to quantify the sensitivity of physical and biogeochemical quantities of interest (e.g. stratification, pCO<sub>2</sub>) to upstream physical drivers (e.g. heat flux, wind stress)


The different model approaches will ultimately be synthesised together to identify what the future holds for the North Atlantic carbon sink, and the extent to which it is controlled by mixed layer & stratification changes or upstream nutrient and anthropogenic carbon levels supplied by the overturning & gyre circulations.