This is an original study on the effect of large valley incision on lower crust exhumation in the context of a continental plateau underlain by a thick continental crust with very low ductile resistance. The authors conclude that, under certain conditions (high plateau elevation, thick crust, large incision rate and prolonged incision history), river incision can lead to exhumation of the ductile lower crust beneath the valley axis. The article is well-written, has high-quality illustrations, and I agree with the main interpretation of the model results. However, I have several criticisms that should be addressed prior to publication, this is why I ask for major revisions although I think that they will be relatively easy to address. My comments to the authors are as follows:

• I understand that in nature, crustal thickness and plateau elevation can vary widely. However, in your case, since you assume constant crust and mantle densities, there should be a linear relationship between these two parameters due to isostasy. This means that you cannot arbitrarily choose both crustal thickness and plateau elevation independently. For example, if we assume a mean crustal thickness of 35 km for sea-level elevation, then local isostasy (neglecting density changes due to temperature and pressure and including the plateau in the total crustal thickness) would give approximately the following values: a 1 km-high plateau corresponds to a total crustal thickness of 42 km, 2 km corresponds to 49 km, and 3 km to 56 km. How, then, can you justify a 65 km-thick crust with a 3 km-high plateau, as shown in Figure 2a, using your chosen densities, without introducing a significant initial isostatic imbalance? Am I missing something here?
• I find one of your results particularly interesting — that surface uplift can be decoupled from the lithospheric response, and that crust-mantle decoupling explains why the Moho remains stable while the lower crust migrates toward zones of lower pressure. Could this behavior explain the very low effective elastic thickness often inferred from the isostatic response of the lithosphere to surface processes? In other words, could this be explained by a situation where only the upper crust effectively responds?
• Yield stress envelope in Figure 2b: there appears to be no strength in the mantle, which seems surprising. With the dry olivine rheology you use, I would expect some resistance.
• I'm generally not in favor of requesting additional model runs in modeling papers, as this can easily become an endless process. However, I am somewhat puzzled by the fact that you don’t really discuss your choice of an extremely weak and thick lower crust, which leads to strong convection and very rapid ductile flow. While I understand this may be a deliberate choice, I think it would be helpful to include a comment on how this specific rheology — which possibly resembles that of an orogenic crust — may not represent the "average" continental crust. Out of curiosity, I would be very interested to see how the system behaves with a more resistant (mafic) lower crust and/or a colder lithosphere. For instance, you could add another dimension to your parameter space in Figure 11, for instance by representing the effect of the thickness and/or average viscosity of the ductile crust and the comparison to natural settings.
• By the way, you should clearly define in the main text what you mean by "lower crust." In your model, you designate crust below 10 km depth as the lower crust, but it shares the same rheological properties as the upper crust. In the literature, "lower crust" can refer either to the ductile portion of the crust — as you do here — or to the more mafic and mechanically stronger part of the continental crust. While this is briefly explained in a figure, it would be helpful to clarify this choice explicitly in the main text to avoid confusion.
• Along the same lines, I’m not sure that such an overthickened and weak crust could remain stable without collapsing, unless it's being artificially supported by the model boundaries. This issue is not visible in your setup because you impose a constant crustal thickness and therefore remove any lateral pressure gradients (except the ones due to valley incision). But from a large-scale geodynamic perspective, the configuration might not be entirely realistic — especially if we consider that real-world plateaus are not laterally infinite. That said, since your model already shows lower crustal flow driven solely by valley incision, I can only imagine how much flow would occur if this plateau were adjacent to a region of much lower elevation and much thinner crust.

Review of "Deformation and exhumation in thick continental crusts induced by valley incision of elevated plateaux", by Thomas Geffroy, Philippe Yamato, Philippe Steer, Benjamin Guillaume, and Thibault Duretz. 

This paper presents a comprehensive numerical study of the impact of valley incision on crustal deformation. Using coupled 2D thermo-mechanical & surface process models the authors present the evolution of a hot and weak crust topped with an orogenic plateau subjected to constant river vertical incision down to a predefined base level. A total of 48 models have been used to explore the role crustal thickness, initial plateau elevation and incision velocity in controlling the relief evolution, crustal exhumation and strain distribution in the crust. Although the authors insist the model aims to reflect tectonically inactive regions, the settings tested (especially where crustal thickness is ≥ 50km, that is for all but 9 models) appear more representative of orogenic plateaux or orogenic systems in general. The comparison with natural examples in the Nanga Parbat and Namche Barwa Massifs in the introduction points that way.  

The paper briefly reviews published literature on the role of valley incision, insisting on the role of erosion potential in controlling crustal deformation over long periods of time (here up to 10 Myr). I really liked the synthesis proposed in Fig 2a showing some statistics about crustal thickness and surface elevation. The physical description of the model is detailed in the paper and the appendices, but needs to be updated here and there (see details below). Results and discussion are well organized but the importance of the partially molten lower crust (and the very high geotherm) is not discussed in enough details, in particular with respect to the exhumation pattern.

Overall this contribution is of broad interest and has the potential to create impact in both the tectonic and geomorphology communities, worth publishing in Solid Earth journal. The manuscript is well written and nicely illustrated, and some reworking should make this a solid contribution. The references seem adequate too. I would recommend accepting this manuscript after moderate revisions. 

Below, I outline specific points for improvement, ranging from minor corrections to more critical issues:

+ line 35: you mention that "first order results align well with observations from natural systems", please provide at least a couple examples here in the abstract. The comparison with natural systems is one of the main concern I had when reading the manuscript, so this should be strengthen. Based on what you provide in the introduction (l. 74-75) it is not clear to me the setting really applies for "inactive regions".

+ Line 73-74: Capitalize "Massif" (e.g., "Nanga Parbat Massif")

+ line 80: you briefly mention the type of constitutive laws used in the thermo-mechanical model, but there is no detail regarding the surface processes here. I suggest adding that the surface process model is a simple erosion law coupled with diffusion. More details are provided later l137 to 144.

+  line 83: Why did you choose this 10 Myr cutoff value? I would assume that in inactive regions processes can be much slower... supposedly the river incision velocity wouldn't be as large as tested in your experiments. 

Regarding the equations in general, please use bold symbols for vectors and tensors.

+ line 99, eq 1: g_i --> should be g, the gravity acceleration as defined l. 104

+ line 100, eq 2: v_i, this term is not defined. Please mention v_i or rather *v* is the velocity.

+ line 101, eq 3: C_P^{eff} is written with an upper case P (as for the equations in Appendix B), but it is spell with a lower case p line 104. Fix it on line 104. There is another major issue with the Heat source terms Q_L and Q_r in Equation 3 which should be in W.m^-3 but they are provided in J.kg^-1 (Table B1, l606) and W.m^-2 (Table 1, p 7) respectively. Please fix the units and make sure you've used appropriate values.

+ Figure 2: I am totally frustrated not to see the geotherm plotted along the vertical strength profile. Moho temperature in all these models is super high (>1000˚C) and important for explaining the convective regime in the partially molten lower crustal domain. I suggest you provide this information for the different models. In fact in Figure 4 it seems that T_Moho is decreasing through time.    

+ Table 1: You provide here diffusion creep parameters for the mantle, yet in Appendix A you write that viscous strain is the sum of dislocation and Peierls crop only. There is not mention of diffusion. Please revise the Appendix and you should also mention that Peierls creep is apply to the lithospheric and asthenospheric mantle materials only.

+ line 209: you mention the computation of the effective erosion rate (E_eff), could you precise what discretization is applied through time for this calculation? 

+ line 233: the time thresholds provided appear to be exact round numbers... Is that because of the model time stepping or output intervals? 

+ line 253-257: To me the "two distinct  high strain rate zones" are not clearly visible. Strain appears diffuse in the viscous crust (between ~6km and the partially molten region). I suggest you provide close up views of these objects, zooming in the crustal region where T < ~ 650˚C. As it appears now the lower crustal convection attracts most of the attention.

+ Figure 4: It would be nice to have the melt fraction in the crust displayed, either with contours or along a vertical profile as you did for the for the strength profile in Fig. 2. The lower crust temperature rises more than 300K over the solidus so I suspect there is a quite large melt fraction... is that reasonable to assume?  

+ lines 277-278: I'm a little confused here. If the plateau height can vary independently from the crustal thickness it means the model isn't at isostatic equilibrium. That seems like a problem for an inactive tectonic setting. Because of the left and right BC applied to the models I assume any configuration will be "stable", the free-slip BCs kind of mimic a lateral stress. Another way to write this is that the assumption of isostatic equilibrium is unclear. The free-slip boundary conditions may artificially stabilize the system, warranting further discussion. As such I have some doubts about the reasoning in this section. 

+ line 295: why is the threshold limit exactly 3km for h_P? Is this number related to the model configuration (i.e. model-dependent)? Could you please elaborate on this in the discussion?

+ line 347: "after 1-2 Myr" --> it seems ∆h reaches a plateau as soon as h_min hits the base level before 1 Myr. After that it gently increases and slightly oscillates. Could you rephrase this part?

+ line 374: Could you please explicitly cite in the manuscript which of the models have the valley reaching the imposed base level?

+ section 4.2. In the model presented here the upper crust if completely decoupled from the mantle because of the partially molten lower crust. Moreover the Moho appears flat for every model. Line 420 you propose that in the models presented vertical flow is associated with curtail isostatic rebound rather than lithosphere response. I would argue that the bottom BC applied in the model doesn't allow for spatial variable lithosphere readjustments. So the discussed behavior appears to be a feature of the model and I believe you simply can't compare your results with Vernant et al. (2013). 

+ line 483-487: I admit am not very familiar with the geological history of the Grand Canyon, but from what I recall there is a long protracted evolution of the Early Proterozoic basement predating the emplacement of the Colorado plateau (e.g the Vishnu schists). It seems evident that the lower crust material eroded during the plateau incision were already near the surface before Canyon incision took place. 

+ line 517: "performed a series of thermo-mechanical AND SURFACE PROCESSES numerical models"

Appendices

+ line 542: add \dot\varepsilon_{ij}^{diff} term for the viscous strain as you have this process for mantle rocks, along with Peierls.

+ line 549-550: replace "second tensor invariant" with "second invariant of the strain rate tensor".

+ line 551: I believe you meant uniaxial rather than axial.

+ line 567: misspelled "deviator stress. Remove final "e".

+ Table B1: replace Mpa with MPa in the first column.

 

Guillaume Duclaux, Nice 26/06/2025