In their manuscript “Controls on Early Cretaceous South Atlantic Ocean circulation and carbon burial – a climate model-proxy synthesis”, Steinig and co-authors present a large ensemble of Early Cretaceous (Barremian-Albian) simulations specifically designed to investigate how some key tectonic features control the ocean circulation, and ultimately, carbon burial in the restricted and expanding South Atlantic basin, in which evidence for Early Cretaceous black shales, documenting intense carbon burial and seafloor anoxia, has been found. Such evidence has notably been discussed by the same group of authors in recent papers (e.g. Dummann et al. 2021) and this earth system model investigation is a logical follow-up.

They notably find that the paleogeographic changes throughout the Early Cretaceous led to a shift from restricted basins in the South Atlantic filled by locally-formed warm and saline intermediate and deep waters to increased exchange with the Southern Ocean, thereby filling the Cape-Argentine and Angola Basins with fresher, colder and more oxygen-rich waters. This scenario is consistent with a progressive shift from intense carbon burial in these restricted basins in the Early Aptian to more oxygenated conditions in the Aptian, in agreement with geochemical data.

The authors also show that their evaluation of paleogeographic uncertainties with an ensemble approach, which considers joint changes in boundary conditions, highlights significant non-linearities in the sensitivity of the South Atlantic ocean circulation to said changes. For instance, the sensitivity of the Angola Basin to CO₂ increases with its progressive northward expansion. This ensemble approach also allows the authors to demonstrate that small tectonic changes can lead to very large changes in local water mass properties, which need to be taken into account in the interpretation of the geological record.

I enjoyed reading this manuscript. The methodology is sound and adequately presented. The results are robustly demonstrated and interesting and the figure are clear and informative. I only have minor comments, which should be straightforwardly answered. Nice job!

Comments

Orbital parameters. Though an impressive set of simulations has been performed, I think that stating that a “full” exploration of the uncertainties has been carried out is a bit exaggerated. Recently, Sarr et al. (2022) demonstrated significant orbital variability of the ocean circulation and anoxia. I think the manuscript should be revised to include some discussion about this.

Figure 1. Consider adding “Angola Basin” and “Cape Basin” for clarity.

l. 103-107. It is unclear what portion of the land/sea mask and bathymetry is updated using Sewall et al. (2007). Is it the full Southern Ocean and South Atlantic, with the latter being subsequently refined according the specified sensitivity tests? Or is it only the expanding Southern Atlantic? Please reformulate.

Figure 2f. It is hard to agree that the deep ocean in the DP closed simulations is close to equilibrium because it looks like it is gaining buoyancy. There are chances that further integration will lead to destratification of the deep ocean and overturning. Also it might be useful to show in Supplementary the spin-up of the global deep ocean because closing DP may impact the Pacific as well and we do not know whether the global deep ocean has reached equilibrium.

Section 3. It is unfortunate that no large-scale description of the ocean circulation is given before entering details about the South Atlantic. This is not given either in Steinig et al. (2020). At a minimum, you should provide details about the deep-water formation zones and global overturning. Do any of these change with altering the SA or the Southern Ocean (Drake) bathymetry? Consider also discussing larger ocean circulation changes with respect to orbital variability as shown by Sarr et al. (2022), see above.

l. 186. The subsurface maximum is visible for salinity but much less clear for temperature (Fig. 4a).

l. 198-199. The GBG and CO2-driven anomalies in salinity are of the same magnitude or even larger than that of DP on Fig. S2.

l. 208-209. The subsurface extension affected by warming in response to Drake Passage opening is not the same for the Falkland Plateau (200-800 m) and the South Atlantic (~ 400 – 800 m, not throughout as stated). Also it is Fig. 6g, not 5g.

Figure 7. How variable is the simulated MOC within a single simulation? Judging by its amplitude, and the fact that agreement across all ensemble members occurs over limited regions, I would say quite highly.

Figure 7 / SAIW. Is the intermediate water formation in the Cape Basin somehow driven by downwelling in the center of the subtropical gyre defined in Fig. 3? Is the convection there seasonal?

l. 243. Fig. 7c (and d,e,f) shows difference in MOC intensity but it is unclear whether you computed the difference between absolute MOC amplitude or not? Could you find a way to show in which regions the MOC changes sign? Another example is on Fig. 7d. ∆MOC is -0.3 Sv at 50°S and 1500 m depth but the absolute mean MOC at this location is 0.15 Sv on Fig. 7a. Does it mean the MOC changes sign there? I don’t think so but it is rather unclear.

Figure 10. Are the contours are the E-P computed over the ocean only?

l. 277. “contradicting a simple dry gets drier, wet gets wetter paradigm”. That the ∆CO₂ E-P be different between stage 1 and stage 2-3 over continental areas is hardly visible on Fig. 10. And the mean ∆(P+E) over subtropical South America and Africa on Fig. S6h between 1200 and 600 ppm looks quite similar to me as well.

l. 297-298. Not sure that the reference to Wallmann et al. (2019) is relevant. Their mechanism has been demonstrated for the mid-Cretaceous, with a significantly deeper configuration of the Central American seaway and more restricted North Atlantic than in the Early Cretaceous paleogeography used here. It is unclear whether it would apply in the Early Cretaceous.

l. 345. Is there any evidence for a downward positive temperature gradient in the proxy record?

l. 420-421. Unclear. Fig.12a (stage 3) shows (northward) export of southern-sourced waters across the GBG at the depth of DSDP Site 361 but it is stated that “the simulations are characterized by a weak southward export of SA waters via the whole GBG depth during stage 3”.

l. 427-434. As it stands, change deep-water formation by intermediate-water formation. Fig. 3e shows max MLD of ~ 200 m, which suggests that convection occurs across a limited depth. It is very possible that dense waters are formed in shallow coastal areas and that the density anomaly is then propagated on the seafloor by the bottom boundary layer scheme of NEMO. But I am not sure the latter is convection as people generally think. Alternatively, it would be interesting to output daily variables to see if the MLD monthly max of ~ 200 m is the result of one or a couple of deeper convective episodes intertwined with the absence of convection. In this case, the term “deep-water formation” would probably be more justified.

In their manuscript Controls on Early Cretaceous South Atlantic Ocean circulation and carbon burial – a climate model-proxy synthesis’ by S. Steinig et al. used an ensemble of 36 coupled ocean-atmosphere simulations to evaluate South Atlantic Ocean circulation during the early Cretaceous. They setup the simulations in order to account for the uncertainty in basin configuration, regional seaways (Drake passage and Walvis ridge) and pCO2. To do so they account for all possible combination of forcing. They then carefully analyzed the contribution of each factor on oceanographic changes in the Southern Atlantic basin.  In the last part of the paper, they integrate simulated scenarios with eNd data and sedimentological information to discuss evolution of Southern Ocean oceanography during the Barremian-Aptian.They notably highlight that paleogeography changes, happening at higher temporal resolution than those commonly investigated with paleo-climate modeling might have very strong impact locally, which would be reflected in the proxy.

I find the methodology very interesting. Those ensemble simulations, that are now accessible with classic ESM/GCM thanks to improved model performance seems to be one of the steps forward for deep time paleo-climate modeling, though they are not often used yet. The choice of boundary conditions and forcing are well justified in the text and relevant with regard to existing literature. The manuscript is overall super easy to read, and I appreciate the care that have been given to figures. I really enjoyed reading this paper and find remarkable that I am almost left with nothing to say. So congrats !

I only have minor comments:

• Authors analyze the dynamics of Southern Atlantic as isolated from the global circulation. I wonder whether the boundary conditions they implemented have also an impact of global circulation and whether global changes could affect the regional signal. For example, do the water masses transported through Drake Passage and entering the Southern Atlantic always have the same characteristics? Or does changes in source water characteristics might also impact the paleoceanography of the Southern Atlantic basin?

Figure 1 – 2 : It could be useful to have the location of basins indicated on the maps as well so it’s easier to visualize. While it’s not essential for understanding what the authors are writing it would help reader not always familiar with local geology structures to better understand.

Figure 1 : Color distinction between sites 249 and 693 is not strong enough.

Figure 2 : would be good to have a plot showing global time series for the sensitivity as well.

Figure 4c-d : This one is a very minor suggestion but changing the color scale for c) and d) to a scale that have various shade of the same color (like white – light blue to dark blue for example) would help better emphasize where temperature/salinity are highly variable and where variability is small.

L.185 – Note that the maxima in salinity and temperature are not strictly located at the same depth: one looks to be in the subsurface while the other is at the surface in the figure. Also “the southern South Atlantic and the Angola Basin both show pronounced” – do the authors mean Cape-Argentine basin instead of southern Atlantic ? Because this would make more sense in the sentence?

L. 197-198 – authors should also refer to figure 5c in this sentence.

L.275 – “somewhat surprising result is that the doubling of atmospheric CO2 does not only increase evaporation over the ocean but also reduces local precipitation, contradicting a simple dry gets drier, wet gets wetter paradigm.” This sounds that it is a general response while looking at figure S6 at first order region of E-P > (<) 0 have increased (decreased) E-P with CO2 doubling no matter if this is driven more by changes in precipitation or evaporation. I suggest that the authors reformulate this sentence.

L337 - Anagola Basin --> Angola Basin ?

L385 - Walvis Ridge Overflow Water – can you plot that on you maps/cross section to improve readability?

Figure S1 : to a horizontal grid of 0.5◦ x 0.5◦ horizontal grid  --> to a horizontal grid of 0.5◦ x 0.5◦ . Please also specified that this is the case for Perez-Diaz & Eagles, 2017 but that KCM plots are at model resolution.

Figure S6 vs Figure 10. Please chose between P-E or E+P notation.