Review of “ROCKE-3D 2.0: An updated general circulation model for simulating the climates of rocky planets” by Kostas Tsigaridis et al.

The manuscript describes the new version of the ROCKE-3D model (Planet 2.0). The introduction describes relevant historical developments, before the manuscript moves on to describing model components and some sample simulations made with the model to tune for different scenarios and different combinations of model components. There is also some exploration of model responses from a scientific point of view and differences with the previous ROCKE-3D model version are described.

I think that almost everything that is needed is there and that the paper is quite close to publishable in the journal following the consideration of my minor (but quite lengthy) comments.

General comments:

1. Given that this is a general circulation model description paper, I was surprised that there was no mention of the model’s dynamical core that solves resolved physical equations. Granted this core is taken from the GISS Model E base model, but the same is true for many other model components, which are at least mentioned briefly.

1. There are unfortunately a rather a large number of minor errors in the paper. I have tried to catch as many as I can, but I think the paper would benefit from the authors taking care to go through it again for readability. With luck, this paper should be useful to modellers for many years!

Minor comments:

Line 21: The IPCC is the Intergovernmental Panel ON Climate Change.

L 27: Extra space after “SOCRATES”?

L 31: Think it would be good to say what the atmosphere used in ROCKE-3D 1.0 for readers that don’t know what this is.

L 52: Dynamic oceans also include further processes that are not parameterised, but solved on the model grid -- specifically fluid dynamics of the ocean.

L 57: Why are the modellers unwilling to use dynamic oceans? Presumably computational expense?

L 61: I would include units T = -21.

L 65: I don’t think “demonstrated” is the right word here, as you not actually shown this to be the case in the introduction? Recommend just DELETE “As demonstrated… introduction,”.

L 112: Does the sensitivity of cloud formation to boundary layer height affect cloud outside the boundary layer as well as inside?

L 118: Is the removal of Rossby radius scaling a good thing for planetary science given that ROCKE-3D will be used for planets with different rotation rates? (I have no idea but would like to know!)

L 122: Are you saying that the supply of groundwater is unlimited?

L 146: The prognostic rivers are the ones where rivers are NOT prescribed I assume? Don’t think this was absolutely bolted down where the new river routing scheme is introduced.

L 150: “were not used in this work”. I think you mean that they are not explored in the results shown in the manuscript. They are available in ROCKE-3D?

L 157: I think there may be confusion over the use of the term “radiative forcing” here. In the context of Earth climate change, radiative forcing is the change in top of atmosphere radiative flux compared with background values due to the influence of an external factor, such as greenhouse gases, over the course of the scenario being considered. Hence comparison of a background heat flux (I think the Davies value is a mean geothermal flux with no changes predicted?) to radiative forcing isn’t quite right. For example, the background solar flux for Earth is about 300 Wm^-2, which obviously can’t be ignored, but the solar forcing for the 21^st century is expected to be fairly small compared with anthropogenic greenhouse gas forcing. Perhaps a comparison between background fluxes would be more relevant here?

L 204: Is it possible to say briefly why non-LTE corrections are not always required to assist model users in understanding the validity of the model?

Also – “A thin CO atmosphere…” This sentence does not seem related to the rest of the paragraph. What is the relevance of this?

L 234: Should “apsis” be “apoapsis”?

L 245: I think SOCRATES has moved on a long way since Edwards and Slingo (1996). However, not sure I can find an up-to-date reference in a journal. Perhaps cite something like

Walters, D., Boutle, I., Brooks, M., Melvin, T., Stratton, R., Vosper, S., Wells, H., Williams, K., Wood, N., Allen, T., Bushell, A., Copsey, D., Earnshaw, P., Edwards, J., Gross, M., Hardiman, S., Harris, C., Heming, J., Klingaman, N., Levine, R., Manners, J., Martin, G., Milton, S., Mittermaier, M., Morcrette, C., Riddick, T., Roberts, M., Sanchez, C., Selwood, P., Stirling, A., Smith, C., Suri, D., Tennant, W., Vidale, P. L., Wilkinson, J., Willett, M., Woolnough, S., and Xavier, P.: The Met Office Unified Model Global Atmosphere 6.0/6.1 and JULES Global Land 6.0/6.1 configurations, Geosci. Model Dev., 10, 1487–1520, https://doi.org/10.5194/gmd-10-1487-2017, 2017.

Which describes some improvements (although it’s almost 8 years old already).

L 290: I think “small fraction” might give the uninformed reader the idea that oceanic transport is less important than perhaps it is? I agree that clearly the atmosphere dominates, but that is not to say that oceanic transport is unimportant. There is a clear handing over of heat transport to the atmosphere in the mid-latitudes, but the way the paragraph is written overemphasises the role of the atmosphere (and underestimates the potential role of the ocean in other climates).

L 347: Sorry, I don’t understand this sentence. “A secondary goal… not relevant to this work.” How can something be a secondary goal, but also not relevant?

L 350: “the atmosphere might have a hard time to adjust.” The language is a bit colloquial here I’m afraid. What do you mean specifically?

Paragraph starting L 364: Is the comparison here between planet_2.0 and the base GISS Model E or planet_1.0 or something else?

L 379: “disabling of the gravity… with the simpler Rayleigh…” Do you mean “replacement” rather than “disabling”?

Table 2 and Figure 3. The model versions appear in different orders in the figure and table. Consider describing them in the same order to help the reader?

L 435: Consider rephrasing. “Throughout the tropics…” suggests that tropical cloud cover has no effect outside the tropics. Not sure that’s what you mean?

L 437: What is the Earth-centric adjustment?

Figure 4: The individual figure titles are very small and faint and therefore difficult to read. Please improve. Also, what do the dots on this figure signify? I don’t think they are just a plotting artifact, as they seem to be different on different panels.

L 467: “Since no clouds are expected to exist in that model configuration”. Sorry I do not understand this? What simulation are you referring to that has no clouds in it? As I understand it the simulations in table 2 are Earth-like enough that we might reasonably expect some clouds? Perhaps more explanation is needed.

L 482: “+1.0 (M40) and +1.3 (F40)” are these in Wm^-2?

L 487: “A key conclusion is the confirmation that virtually all configurations across model versions produce the same climatology. These are marked with a blue arrow in Fig. 5, where there are practically no differences between the simulations.” This statement does not appear to be supported by the figure? There are more than 30 perturbed simulations in Figure 5, but only two of them have blue arrows.

L 509: “The simulations do not have a stratosphere”.  I had a look into this and I see that you are correct that the Earth literature often defines the stratosphere as a region in which temperatures increase with height. (Even the American Meteorological Society glossary states this!) There is an alternative definition that states that the stratosphere is a region of the atmosphere that is close to radiative balance. This occurs in the absence and presence of ozone and does not require that the atmospheric temperature gradient dT/dz is positive. Only that it is stable. See for example the discussion in the introduction to:

Thuburn, J., and G. C. Craig, 2000: Stratospheric Influence on Tropopause Height: The Radiative Constraint. J. Atmos. Sci., 57, 17–28, https://doi.org/10.1175/1520-0469(2000)057<0017:SIOTHT>2.0.CO;2.

For the present paper I think that this second radiative-balance definition is interesting because it applies across a much wider range of planets than just an Earth-like planet with an ozone layer and highlights the dynamical differences across an atmosphere. Mars, for example, can be said to have a robust stratosphere under this definition. In any case, I think it would be a good idea to state up front what you mean, as initially I was confused. I thought you meant that no part of the relevant atmospheres is close to radiative equilibrium, but analysis of your figures suggests to me that that is not the case.

L 531: I think you are suggesting that a 20 year mean is equivalent to a 20 member model ensemble mean? I’m not really sure what you are driving at. This sentence could be deleted.

Figure 6: The outliers are individual outlier years or something else?

L 565: When you say “is in all climate-relevant respects the same”, do you mean that it produces the same climate or that the code is the same? (I thought the former, but then the next sentence made me reconsider.)

Figure 7: x-axis units are years?

L 627: “dependents” -> “depends”

L 646: Should the idea be that the configuration changes are as limited AS POSSIBLE (to reduce spinup time) rather than that configuration changes MUST be limited?

L 669: Wasn’t the supercontinent called Pangea? (Or Pangaea or maybe the Pangean Supercontinent…?)

L 699: The land at the south pole point in the Venus aquaplanet simulation. Is this incorporated for computational reasons rather than scientific reasons?

L 711 and Figure 10: In the Earth literature at least, the term “cloud radiative forcing” is deprecated. “Cloud radiative effect” is preferred. This is because “radiative forcings” are taken to imply external influences that are applied to the model (such as a change in the composition of an atmospheric absorber, or a change in incoming stellar radiation).

L 779: The “bathtub ocean” has a maximum depth of nearly 3800 meters. Is this a constant depth or is there variation?

L 789: To clarify: The runs have global mean surface temperatures of 20.9 and 22.2 C? Or they are 20.9 and 22.2 C warmer than runs with Earth topography?

L 841: “40x time oceanic increase” Is this a typo?

L 860: “A comprehensive list and explanation of all options is far outside the scope of this work…” Well – this is the model description paper, so we might expect to see some of the major parameters described? As it is, most of the parameters described relate to timesteps, diagnostics and ice and low temperatures. Given that other model parameters were tuned between various set-ups that will be released to the public, it might be good to describe these since that will be useful not only for understanding differences between simulations but also for others who want to do some model tuning themselves? Looks like there may be some description under “variable names” in Appendix D, but this isn’t quite the same thing.

L 863 and table 4: In Line 863, DTsrc is described as the parameter that “tells the model how often its main parts should exchange fluxes” – so the frequency at which coupling between atmosphere and ocean etc occur. However, in table 4, DTsrc is described as the physics timestep (presumably atmospheric processes that are not simulated by the dynamical core?) Is one of these definitions in error?

Table 4: maxctop “can affect model results”. Does this mean that the parameter occurs in the prognostic equations. I recommend that this is stated given that the previous two parameters are given as diagnostic.

L 902: “The first thing one notices when comparing Fig. 5 with Fig. 16 is how much more colorful the latter is, implying that the simulated differences across planetary configurations are much more impactful to climate variables than the changes we did towards the generalization of the Earth configuration.” I’m not sure I follow I’m afraid. I think you are saying that the differences between planetary configurations are larger than the differences between GCM setups for a given planetary configuration?

L 922: Just a check: Is this the tropopause temperature and pressure or that of the temperature inversion?

L 976: Another reference to the stratosphere. Would be good to revisit the definition.

L 1009: “Give way abruptly to” or “abruptly give way to”.

L 1035: Heat that is “deposited near the tropics” because of the star or something else?

Figures 24 and 25: Figure captions / legends are faint here. Would be good if they could be made bold.

L 1116: “Even with this numerical noise… regardless.” I’m not sure this sentence makes sense? Please check.

L 1118: “quantitatively the same”. Do you mean “qualitatively the same” given that values are not – as you say – identical?

Conclusion section: These are worthy aims. However, it would be good if the conclusion section also provided a short summary of the findings of the paper, perhaps with some forward look. I think this is particularly true given that the introduction takes the time to give a nice summary of relevant past work on numerical modelling. What are the most important avenues for future model development?

The paper describes the second version of ROCKE-3D, which includes upgrades derived from both Earth-centric developments (such as physics updates, bug fixes, and structural enhancements from the GISS ModelE2.1) and planetary-centric developments (notably in radiative transfer calculations, atmospheric composition, ocean parameterizations, and model resolution). It also introduces specific options tailored to the modeling of rocky planets, including thin atmospheres, geothermal heat flux, and an updated calendar module.

The paper represents a valuable contribution and a substantial advancement in the modeling of rocky planets. However, the structure should be improved, and some simulation results need a more in-depth discussion, as outlined below (with additional suggestions included in the list of minor comments).

First, for completeness, a brief description of the atmosphere, sea ice, continental ice, and vegetation components used in ROCKE-3D should be provided in Section 2.

As shown on p. 33, when using the SOCRATES radiation scheme with Qflux = 0 at 1xCO₂, the simulation results in a snowball state. This implies that the SOCRATES runs at 1xCO₂ and 2xCO₂, starting from the same initial conditions, lie below and above the unstable branch that separates the warm and snowball states, respectively. This behavior is both interesting and expected, not only in slab-ocean models but also in GCMs (e.g., Ferreira et al., J. Clim. 24, 992 (2011); Brunetti & Ragon, PRE 107, 054214 (2023)). The fact that the GISS radiation scheme reaches the snowball state at a different CO₂ concentration (0.5xCO₂) simply indicates that the bifurcation diagrams for SOCRATES and GISS differ—particularly regarding the location of the unstable branches.

A similar behavior is expected when using a dynamic ocean. Therefore, it would be useful to expand the discussion at this point and, if possible, include additional simulations showing the occurrence of the snowball state also in the case with a dynamic ocean.

As you mention in lines 802–804, sea ice was removed in Rencurrel & Rose (2018) because their study specifically focused on ocean–atmosphere interactions. However, more generally—and depending, of course, on the scientific question being addressed—sea ice should not be excluded from the model. Its role in the full coupling between atmosphere and ocean is crucial for climate dynamics, particularly for the emergence of steady-state climates characterized by large ice caps, waterbelt configurations, or snowball conditions (Rose, J. Geophys. Res. Atmos. 120, 1404 (2015), Brunetti et al., Clim Dyn. 53, 6293 (2019), Hörner & Voigt, Earth Syst. Dyn. 15, 215–223 (2024)), all of which are highly relevant in exoplanet studies. The discussion in this section would benefit from being expanded and refined to better reflect the importance of sea ice in such contexts.

Figure 18 is particularly interesting for understanding how key mechanisms depend on the choice of radiation schemes in dynamic ocean simulations. The GISS scheme produces a warm blob in the northern hemisphere when using atmosphere A, associated with reduced total cloud cover. This result is somewhat counterintuitive and may be linked to processes occurring in the stratosphere. Could the warm blob also be influenced by ocean circulation? The paper does not discuss the strength of the overturning circulation, and including such results would be valuable.

When considering atmospheres x and N, the warm blob disappears, suggesting that it may be related to cloud physics—although it is difficult to draw firm conclusions, given that simulations A, x and N are equilibrated using different parameters, as shown in Table 3.

In the x simulations (central column in Fig. 18), the southern hemisphere becomes warmer with GISS and features increased cloud cover. However, the northern hemisphere becomes colder—also with more clouds. Could you provide an explanation for this behavior?

In the N simulations (right column in Fig. 18), the equatorial regions become warmer and less cloudy with GISS. Similarly, in Figure 19, the qN simulations show that GISS yields a warmer equatorial region with reduced cloudiness. These findings raise important questions about the cloud formation mechanisms at play, which would benefit from a deeper discussion.

Minor comments and corrections:

l30, Abstract: specify which is the atmosphere used in ROCKE-3D 1.0

l54: the sentence is not clear in my opinion, it seems that slab-ocean (Q-flux) can be considered AOGCM while they do not have a dynamical ocean (thus, not OGCM by definition). The sentence needs to be reformulated.

Section 2, Model description: I suggest to use only ROCKE-3D 1.0 and 2.0, instead of adding also planet_1.0 and planet_2.0, which are redundant (and confusing) definitions of the same models. (Note that at l. 98 another name is used: ROCKE-3D Planet 1.0).

l83: `in section 2.1’ instead of `in the same section’

l122: `small increases and decreases’ : how much water goes in groudwater recharge? How well is the water mass conserved?

l130-131: use \citet for Rosenzweig, Russell and Schmidt refs.

Section 2.2.1: Dynamical lakes are tested elsewhere. However, it would be useful to know which is the impact of including them, in particular at which resolution one needs to include their dynamics.

l184: millibar is sometimes denoted as mb and other times mbar (the second is prefereed, however should be consistent everywhere)

l189: `stellar’ instead of `solar’ (check also elsewhere in the text)

l210, 213, 222…: `integer’ instead of ‘integral’

l258: speed reduction: you can point the reader to Sec. 4.5 and Fig. 15

Sec. 2.5: As you discuss at p. 33, it is possible to simulate aquaplanets with flat bathymetry and full dynamical ocean. You should specify this here and point to the paragraph at p. 33.  

l339: `In this work, the GCM variable names of the model output are in italics for many quantities…’

Sect. 3.2: specify that these simulations are performed with prescribed temperature (p).

l369: sect. 3.1 instead of 4.1

l384: Can you specify which is the `geographic adjustment’ needed in the Earth-oriented version?

Sect. 3.3: is the 1850 sim the same as in sect. 3.2?

Fig. 3 caption: `explained in Sect. 3’  instead of 4

l425: `sect. 3.1’ instead of 4.1

l437: which are the `Earth-centric adjustments’ that you mention here? It is not clear if you refer to irrigation, etc mentioned at the beginning of this section, or to cloud parameterizations mentioned later, at p. 17. Please specify and, in case, reoganise this section accordingly.

l440: which is the atmospheric composition in these sims?

l467-470: These sentences are not clear, please reformulate.

l479: which is the longwave correction you mention here?

Sect. 4.1, first paragraph: Give definitions of U00a, U00b and the other two parameters.

Caption Fig. 6: how long are the simulations with dynamical ocean? 2000 yr as you say later, at p. 23? Specify how long are these runs for obtaining the values in the figure. 

l566: for regional differences across model simulations (with prescribed sst) you can refer to fig. 5

l580: in Table 2 for intermediate simulations, and in Table 3 for the final simulations

l610-611: in the caption of fig. 7 you say that planet 1.0 is shown in green, while in the main text you say it is the faster to equilibrate (while it is the blue)… Please check, since you mixed up the colors in the main text.

Fig. 8: at which depths levels 11 and 13 correspond? Deeper levels should require more time to stabilise.

Sec. 4.4.1 and fig. 9: Have these maps for the ancient Earth  a constant ocean depth? Which is this depth?

l692: Say that the modern-day Venusian topography can be found in Fig. 5 of Way & Del Genio 2020

Fig. 10: I think that this figure is not necessary, since just the final values at the equilibrium are used, which can be put in a table.

l738: \citet

l740: … and albedo, and conditions at \sim 3.5 Ga.

l748: Here, it is…

l813: this paragraph on `Deep ocean worlds’ is describing a future update, I suggest to include it later when you talk about `Other updates’, which could become `Other updates and future developments’

l823: h/R in math font

l827: parenthesis should be removed

l866: give a definition of DT and Nisurf or point to Table 4

l919: you should point the reader to the right variable in Fig. 16 for the low cloud fraction (pcldl?). The same for tropopause temperature and pressure (cldtpp and cldtpt?)

Fig. 17: Add S, G, p, q, o at the top (as for example in Fig. 16) for clarity.

Fig. 18: points are not visible, try to increase marker size or type

l987: … smoother transition in specific humidity.

l987-988: I would eliminate the last sentence. All this is better explained in the following section.

l1002-1004: Reformulate the sentence: `This is directly …. A atmosphere,…’

l1024: I would introduce panel labeling and say: `see, e.g., Fig. 19a in comparison with 19d and 18c for the changes …’

l1029: (Fig. 18c and 19d), for dynamycal ocean and Q-flux, respectively.

l1042: … invalid, we decided …

Fig. 24: last two panels are inverted with respect to the description in the caption.

l1111: … in Sec. 4.3 and App. B

l1023, Outreach: add links

l1046: gridbox^{-1} on the same line

l1048: `… 680 mbar. It also avoids…. ‘

l1058: add link for ROCKE-3D spectral files