the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 24 Mar 2026
LaScape 1.0: An open-source module for three-dimensional thermo-mechanical and landscape evolution modeling
Abstract. The feedback between tectonic events and surface processes fundamentally shapes landscapes and lithospheric deformation. However, the quantitative interaction remains poorly constrained. While numerical simulations offer powerful insights, 3D numerical models that couple thermo-mechanical processes with landscape processes remain uncommon due to the mismatches in temporal and spatial scales. Here, we present a 3D coupling methodology that integrates the thermo-mechanical code LaMEM with the landscape evolution code FastScape. A finite-difference marker-in-cell technique solves the thermo-mechanical processes, and a sticky air layer at the top boundary, combined with an internal mesh, effectively and stably simulates the free surface. Each timestep is synchronized with a finite-difference landscape evolution model. The timesteps of LaMEM are read by FastScape, which then subdivides them into smaller, iterative intervals to simulate surface processes. We demonstrate that the coupled model operates efficiently and stably. We validate our couple model by applying it to three classical tectonic regimes: oceanic subduction, continental collision, and continental extension. These cases converge quickly and align well with geologically realistic results. This approach provides a powerful, quantitative tool for exploring the bidirectional feedback mechanisms between the deep Earth and its surface, offering insights into the genesis of complex geological structures and landscapes.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Geoscientific Model Development.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(24470 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-1170', Anonymous Referee #1, 02 May 2026
-
CC2: 'Reply on RC1', Jianfeng Yang, 06 Jun 2026
We thank the reviewer for the detailed assessment of the manuscript’s suitability as a model-description paper. In our revision, we will present our coupling module/plugin for FastScape in LaMEM more clearly rather than as an entirely independent piece of software. We will add a proper user manual, workflow documentation, model input descriptions, exact code/version information, computational-cost information, and a more detailed explanation of the coupling algorithm. We will also add a controlled 2D/quasi-2D versus fully 3D plume-style comparison to demonstrate why fully 3D coupling is useful, while removing or softening broad geodynamic conclusions that are not supported by the proof-of-concept examples. Such fully 3D coupled modeling, especially open-source codes, is still uncommon and we believe this is a contribution to our community. However, we fully agree with the reviewer that we should be careful during the “benchmark” process and describe the limitations and caveats of the coupling explicitly.
Citation: https://doi.org/10.5194/egusphere-2026-1170-CC2
-
CC2: 'Reply on RC1', Jianfeng Yang, 06 Jun 2026
-
RC2: 'Comment on egusphere-2026-1170', Anonymous Referee #2, 02 Jun 2026
The manuscript by Luo et al. aims at 1) developing a new approach for coupling in 3D a landscape evolution model to a thermo-mechanical model, and 2) using it to investigate the role of surface processes on the dynamics of a convergent and of a divergent setting. If I acknowledge the efforts made by the authors, none of these two goals are well tackled in this paper. The paper fails in explaining with clarity the novelty and efficiency of the coupling approach, which leaves major doubts concerning the validity of the results. In particular, the results are not convincing as the landscape evolution model produces some landscapes which are not realistic, likely highlighting some limits of the coupling approach. The discussion and conclusion are almost non-existent. I therefore recommend rejecting the paper at this stage.
Line comments (including minor or major comments):
First paragraph of the introduction: some important papers, on the interaction or the impact of erosion on tectonic stresses and deformation, are not cited:
- Avouac, J. P., & Burov, E. B. (1996). Erosion as a driving mechanism of intracontinental mountain growth. Journal of Geophysical Research: Solid Earth, 101(B8), 17747-17769.
- Vernant, P., Hivert, F., Chery, J., Steer, P., Cattin, R., & Rigo, A. (2013). Erosion-induced isostatic rebound triggers extension in low convergent mountain ranges. Geology, 41(4).
- Steer, P., Simoes, M., Cattin, R., & Shyu, J. B. H. (2014). Erosion influences the seismicity of active thrust faults. Nature communications, 5(1), 5564.
- Jeandet Ribes, L., Cubas, N., Bhat, H. S., & Steer, P. (2020). The impact of large erosional events and transient normal stress changes on the seismicity of faults. Geophysical Research Letters, 47(22), e2020GL087631.
Line 52: “On the spatial scale, most landscape evolution models are designed for length scales ranging from a few meters to several kilometers and require high resolution” – Most LEM actually uses resolution of 10s of m at a spatial scale of about 10s to 100s of km.
Line 55: “This characteristic necessitates the inclusion of horizontal advection in landscape models and may require the use of interpolation schemes.” – This is true (advection) but has nothing to do explicitly with the question of resolution. It is rather an issue related to the tectonic processes that one wish to include in its models. Moreover, interpolation is not a fatality, as one could use lagrangian thermo-mechanical model and a lagrangian LEM (even though this has never been done to my knowledge)
Line 59: “The evolution of river systems is significantly influenced by variations in lateral tectonic structure and events, which 2D models cannot accurately capture.” – you need to support this statement by references or else. It is well known that rivers respond to longitudinal changes in uplift, lithology, climate, etc., but the lateral impact of tectonic structures and events (?) is quite an open question for me.
Line 71 - “This module achieves a more realistic interaction between surface processes and tectonic events.” – compared to what? This sentence is unclear.
Line 87 – “𝜆̇ is the magnitude of the plastic strain rate, and 𝑄 is the plastic flow potential.” – Could you please briefly explain how these two terms are computed?
Line 110- “Based on the special node ordering method and implicit solve structure” – Please be explicit
Section 2.2 – Very little info is given on the numerical method used to solve equation 11
Section 2.3 – This section, a pivotal one, is really poorly described, resulting in a clear lack of clarity:
- Are the two models you couple Eulerian or Lagrangian?
- For instance, some sentences are not sufficiently clear: “We include horizontal advection in surface processes to simulate plate movement.” _ I do not understand what this means, If you want to say that horizontal advection is included in Fastscape, says that rather than surface processes. “A bilinear interpolation algorithm is used to produce more refined surface processes.” - I do not understand why a bilinear interpolation produces more refined surface processes (this sentence makes no sense).
- It is also unclear how the two topographic grid correspond with each other : “The landscape evolution model receives topography from the thermo-mechanical model at the first timestep and uses the previous topography field stored in the FastScape grid for iteration without re-receiving the topography data from LaMEM at every timestep during the landscape evolution” – When you say first time step, you mean the first time step of the LEM or of the thermo-mechanical model? What happens at the second time step of the thermo-mechanical model (this is not covered by the explanation given here)? You reuse the topography of the last iteration of the previous LEM run, or do you restart a new Fastscape grid based on the LaMeM grid?
- “After the calculation of landscape evolution, the updated topography field will
be passed to LaMEM.” – you need to explain how!
- I do not understand at which stage the bilinear interpolation is used – I guess it occurs during the transfer of information from Fastscape to LaMeM, but saying it explicitly would help.
- “As a result, the grid resolution is increased by the refined factor n in both the x and y directions.” – Unclear how this operates. If this operates at each time step of the LaMeM model? If yes, this will generate rapidly an explosion of the number of node elements?
- Making a simple illustration (better than panel C of Fig 1 which is not informative) where you show the grids of the two models and how they correspond to each other through time would help to clarify this part. You can also refer to Thieulot et al. (2014) for inspiration, as they did a much better job at explaining the coupling between the grids of their LEM and themo-mechanical model.
- Fastscape uses a regular grid, but you interpolate it horizontally – how do you manage this? If this is through the bilinear interpolation scheme you mention, I do not believe it satisfies to my following point (see next bullet).
- A very important point: This is absolutely crucial that you do not change the grid of the Fastscape model (except for the tectonic advection of the grid nodes following the LaMEM surface advection field), as the organization in valleys, hillslopes and ridges is a non-instantaneous process, associated to a response time that can span up to a few Myr. Any change to the topography which would disrupt this organization (e.g., a river node that disappears), would lead to a transient response that would completely and artificially disrupt the expected surface evolution. I was not able, during the review of this paper, to make sure of this due to a lack of clarity of this part.
Results: Subduction/convergence and Figure 3 and 4: These two figures by themselves could justify rejecting the paper in my opinion. In particular, Figure 4b, showing the topography resulting from the model with surface processes, illustrates many issues related to either the landscape evolution model used or more likely, to the way it is coupled to LaMeM. Indeed, rivers are perfectly aligned, too densely spaced, non-continuous from upstream to downstream. Even if lateral advection is expected to elongate the catchments, the produced results are not realistic nor do they make physical sense. And they illustrate that the coupling approach is not efficient, due to loss of river continuity from upstream to downstream. Moreover, concluding that “While rivers play a role in tectonically stable zones, they do not significantly shape the orogenic wedge due to a balance between erosion processes and bedrock uplift.”(Line 210) based on only two models, that moreover are not convincing, while contradicting a large literature, is extremely bold!
Results: Continental extension and Figure 7 and 8 – same as for previous section. The landscape evolution is really not convincing.
Discussion and Conclusion: The discussion and conclusion are not sufficiently developed. There is no attempt to position the outcomes of this paper in a state-of-the-art. In turn, in it is unclear what is the novelty of this paper or what it brings to our understanding of the interaction between tectonics and surface processes. It is also unclear how the developed methodology compares to past modelling work. The conclusions are vague and not direct enough to push for publication of this manuscript.
Citation: https://doi.org/10.5194/egusphere-2026-1170-RC2 -
CC3: 'Reply on RC2', Jianfeng Yang, 06 Jun 2026
We appreciate the reviewer’s comments on the unclear technical description and the unrealistic river geometries. We will substantially expand the coupling-method section, including how LaMEM surface velocity and topography are transferred to the uniform FastScape mesh using bilinear interpolation, why interpolation is necessary for different resolutions and non-uniform LaMEM grids, and how information is passed back to LaMEM. We will also clarify the time-stepping and grid treatment through the coupled run. The current river geometry is mostly caused by deactivating random noise/seeding for controlled comparison; we will address this by adding seeded runs and replacing the problematic figures where needed, and by discussing remaining limitations and caveats of the coupling explicitly.
Citation: https://doi.org/10.5194/egusphere-2026-1170-CC3
-
EC1: 'Comment on egusphere-2026-1170', Stefan Hergarten, 04 Jun 2026
Dear Authors,
sorry that it took a bit longer than expected to get two reviews for your manuscript! The reason was that I wanted to have one expert on the convection part and one expert on landform evolution modeling.
You probably have noticed that both reviews are quite critical concerning the scientific/technical part as well as concerning the focus as a model description paper. One point is that there have already been approaches to couple convection with landform evolution and it is not sufficiently clear enough what is new here. The first reviewer even sees LaScape is rather as a small plugin for LaMEM and finds that too little weight is on the description and potential challenges of the coupling module compared to the two coupled models.
Among several other points, the second reviewer emphasizes that there are apparently severe problems with the topography. Another scientist who unfortunately declined to review the paper replied to my request "... This would be a lot of work to review, because it is incomplete and contains spurious results (cf advection problem at left boundary in Fig. 7, straight rivers without proper "seeding", etc)." I would agree that there are problems with the topography. From my own experience in this field, I guess that there is a fundamental problem with the Eulerian description of horizontal movement in the landform evolution model.
Although it does not look very positive at the moment, I would be happy to get your opinion about the (apparent) issues with the topography and also about the other major points raised by the reviewers.
Best regards,
Stefan HergartenCitation: https://doi.org/10.5194/egusphere-2026-1170-EC1 -
CC1: 'Reply on EC1', Jianfeng Yang, 06 Jun 2026
Thank you for handling our manuscript and for summarizing the main concerns. We agree that LaScape is best described as a FastScape coupling plugin/module for LaMEM rather than as an entirely independent model, and we will revise the manuscript title accordingly. Note that the editorial staff of GMD urged us to put a code name in the title, which we did not have initially. We now understand that it should in principle also be possible to mention both LaMEM and FastScape in the title. Although the plugin is relatively compact in terms of code size, we believe it provides an important capability. It will enable LaMEM users to couple three-dimensional thermo-mechanical models with FastScape surface-process calculations. Such fully 3D couplings remain uncommon (especially for open-source codes), and this functionality opens the possibility of addressing more complex tectonic and geomorphic problems within a well-established community geodynamic code.
We also accept the editor’s and reviewer’s point that the present manuscript gives too much space to the two component models and not enough to the coupling module itself. In the revised manuscript, we will shift the focus toward the coupling design, data exchange, grid interpolation, time stepping, workflow, computational cost, limitations, and potential failure modes. We will also improve the documentation and examples so that the contribution is presented more clearly as a useful, reproducible LaMEM plugin for the community.
We understand that the present version raises serious questions about the topographic and river-network results. The straight and overly regular river patterns mainly result from our choice to turn off random topographic noise/seeding so that models with and without surface processes used exactly the same initial setup. We agree that this choice makes the resulting landscapes look artificial and weakens the manuscript. In the revision, we will add appropriate seeding/noise, rerun or replace the affected examples, and explicitly diagnose whether any remaining artifacts arise from setup choices, boundary conditions, interpolation, or the Eulerian horizontal-advection treatment. We will also refocus the paper as a GMD model-description contribution rather than a paper making broad scientific claims.
Citation: https://doi.org/10.5194/egusphere-2026-1170-CC1
-
CC1: 'Reply on EC1', Jianfeng Yang, 06 Jun 2026
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 1,160 | 448 | 59 | 1,667 | 40 | 62 |
- HTML: 1,160
- PDF: 448
- XML: 59
- Total: 1,667
- BibTeX: 40
- EndNote: 62
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
In this paper the authors coupled the geodynamics model LaMEM with the landscape evolution model FastScape. They present in this paper this coupling as a separate code coined LaScape1.0. In the first part of the paper they briefly explain how LaMEM and FastScape work. Then they explain how LaScape1.0 achieves the desired coupling between the two models. Finally the authors show two examples of the coupling, one subduction models, which later goes into continental collision and one continental extension model. The authors claim that their works shows the importance of fully 3D coupled models and that their models show that the impact of surface processes on geodynamic processes is limited to relatively small spatial scales.
Since the article type is a "Model description paper", I will use the description from https://www.geoscientific-model-development.net/about/manuscript_types.html as a guideline.
- "The publication should consist of three parts: the main paper, a user manual, and the source code, ideally supported by some summary outputs from test case simulations."
Besides a small readme in the Zenodo repository, I could not find a user manual. The source code is available in the Zenodo repository. It contains LaMEM, the code for the coupling FastScape to LaMEM, and the input files to run the models shown in the paper. So, besides the missing manual, the submission seems compliant. It is very clear tough that this is an extension of/plugin for LaMEM which allows the use of FastScape, and not a separate code or model.
- "The main paper should describe both the underlying scientific basis and purpose of the model and overview the numerical solutions employed. The scientific goal is reproducibility: ideally, the description should be sufficiently detailed to in principle allow for the re-implementation of the model by others, so all technical details which could substantially affect the numerical output should be described. Any non-peer-reviewed literature on which the publication rests should be either made available on a persistent public archive, with a unique identifier, or uploaded as supplementary information."
The introduction is generally fine, but even though the authors are aware, and used the ASPECT-FastScape coupling, as stated in their Author contributions, they do not include it in the discussion or cite the paper which implemented that coupling (https://doi.org/10.1130/G49351.1), which would seem appropriate for this kind of paper. Especially since they state the fully 3D coupled models are uncommon, with the ASPECT-FastScape coupling being a fully 3D coupling (and 2D). The original paper was actually in 3D, but was only very thin, but it was for example used in https://doi.org/10.55575/tektonika2024.2.2.75 as a real fully 3D coupling.
The paper does a decent job of describing the methods they used to implement the coupling, although I do think there could be more focus on the details of the coupling itself compared to the LaMEM and Fascape descriptions. For example, based on the text and figure 1, it is not clear to me whether all data is send to rank0 and then the surface is filtered out, or only the surface node data is send to rank0. Nor is it explained how the velocity and topography parameters are exchanged without any loss of information (no interpolation in the process at all?). Other information which might be useful to potential users might be how much extra compute time does it cost to have this coupling enabled, and how much of that time is spend in the coupling part, compared to LaMEM and FastScape? In my view the approach for the coupling is straight forward and has been done several times before. So although I am sure this is a nice addition to LaMEM, method wise there is as far as I can see nothing new here. What the paper doesn't discuss well what the limitations and caveats of the coupling are. What kind of models can and can't be run, under what conditions does the solution become inaccurate, etc.
It is not fully clear though what versions of LaMEM and Fascape they used. This should be documented more clearly in the paper. Especially since there are now multiple very different versions of FastScape. I think that they used the Fortran version of FastScape, and not the newer C++ version, but this should be documented well in the paper, which exact version numbers, to make it reproducible.
- "The model description should be contextualised appropriately. For example, the inclusion of discussion of the scope of applicability and limitations of the approach adopted is expected."
see point above.
- "Examples of model output should be provided, with evaluation against standard benchmarks, observations, and/or other model output included as appropriate. In this respect, authors are expected to distinguish between verification (checking that the chosen equations are solved correctly) and evaluation (assessing whether the model is a good representation of the real system). ..."
The two codes that have been coupled are standard community codes, and as far as I am aware, there are no standard benchmarks for coupling of these kind of models. Besides that the output seems, in the eyeball norm, alright, there is no verification as defined above. Technically, since the code is available, anyone could check whether the implementation is done correctly, but I don't think that is what is meant with this point.
- "Code must be published on a persistent public archive"
The coupling code together with the used version of LaMEM is in a persistent public archive. The user has to obtain PETSC and FastScape by themselves.
General points:
In my view this is a plugin for LaMEM, which couples it to FastScape, not a separate model. This makes me doubt whether it make it worth it to publish in this journal. There is value to having a detailed description though to how these kind of couplings work, but this paper also seems to lack in that regard.
The paper tries to also present some new scientific findings with the coupling ("Model results indicate that the impact of surface processes on geodynamic processes is limited to relatively small spatial scales"), which this is definitely not the type of article for. In my opinion, besides not being the place to make these kind of statements, I do not think these kind of conclusions can be drawn from these kind of simple, low revolution proof of concept models. I would like to see significant more evidence for such claims to be made.
To add to this, the parameter study the authors do seems to be focused on investigating the effects of varying the parameters on the resulting model, not to show that the coupling works well under various types of conditions.
I would like to add that, although I generally agree with their statement that fully 3D couplings can be useful and important ("further demonstrate the advantages of using fully coupled 3D models"), if you would make such a statement, it would be great to actually add a 1 to 1 comparisons between a T-coupled model (2D LaMEM with 2D fastscape) and a fully 3D coupled model to actually compare the differences. Especially since the examples they show are basically laterally extended 2D models, and do not (initially) contain any 3D complexity.
Conclusion:
I will leave it up to the editor whether this manuscript is fitting for this journal, but even if it is, I think it would require major revisions before it can be published.