General Comment

In their study, Sun and Cao, present a new permafrost model, evaluate its performance at certain locations and apply it to the whole Tibetan plateau. They compare observations with simulations forced with local weather station records and with large scale reanalysis datasets. They discuss what makes the model perform better or worse and presents results on the evolution of Tibetan permafrost since the 80s. The model includes Surface Energy Balance calculation but its calculation and its coupling to the energy budget of the soil column is, from my understanding, either poorly described or problematic in its design (see my Important Comments). It resolves heat conduction in the ground with effective heat capacity, freezing curves but the water content of each cell is static (no infiltration or upward suction via evaporation and matrix potential). At the surface it also includes a snowpack module that does not consider snow mass balance.

The model claims to be flexible and to propose novel parameterizations, but according to me the flexibility is not explained or demonstrated (see my Important Point on L58) and I do not see novel parameterizations. Additionally, I do not see what are the new possibilities that this model offers compared to already existing models. I think the study needs to better present what makes it original and motivated its development. A detailed summary of its strength and weakness compared to other models would help understand the motivations for its development.

Also, in my view, the study might be more aligned with the scope of GMD than TC, given that its primary focus is on presenting a new geoscientific model. Typically, I would expect TC publications to emphasize scientific results or insights directly related to the cryosphere, whereas GMD is intended for studies of this nature. That said, since the manuscript has already passed the initial editorial screening and the editor has decided to proceed with the review process, I take it that its placement here is considered appropriate.

Altogether, for now, I have the feeling that the model in itself is not particularly novel/needed by the community (but I am happy to be proven wrong) and, unless I am mistaken, its description includes important flaws that needs to be addressed. Therefore as it is, I recommend major revisions to address those crucial points. I have not provided detailed comments on the rest of the manuscript at this stage, as I believe the major issues outlined above should be addressed before a more thorough evaluation is meaningful. Overall, I did not identify major issues with the model setup, validation, results, or discussion. However, I find the robustness of the results at the Tibetan scale questionable, given the discrepancies between observations and simulations when using the reanalyses.

Important comments

L30-43

When a new model is published, it becomes part of the existing landscape of models, and it is important to explain how this new model positions itself relative to the existing ones in order to justify its relevance. I think the study should be more thorough in this regard and more exhaustive regarding which permafrost models are used in Tibet to study what. The Introduction should be expanded in consequence (see my comments about L42 and L43).

L58

“The application of FPM with lateral heat is provided in Sun et al. (2023).” This is very confusing to me for two reasons. First, vertical 1D models and cross sectional 2D models are usually very different types of models that implies different formulation of their physical equations, different numerical schemes for their resolution and different types of upper boundary conditions. The present study mentions surface energy balance calculation whereas Sun et al. (2023) forced their model with ground surface temperatures. So I do not understand how a 1D model can become a full 2D model (not a coupling of 1D simulation together).

Second, the given explanation is confusing. Sun et al. (2023) says “A 2D heat conduction model developed by Ling and Zhang (2004a) was used to simulate the permafrost thermal regime.” So if it is the same model in 1D and 2D, is the present study actually presenting a new model or is it presenting improvements brought to a model published initially in 2004? Since the goal of this study is to present a new model, this kind of question need to clarified to consider publication.

L62-104

“A physically-based surface energy balance scheme for different land surface cover types with varying snow regimes and properties was coupled to FPM, and was formulated as:

(1−α)Qsi +Qli +Qle +Qh +Qe +Qc = Qm”

Major problem here. The study intends to describe the ground surface SEB (because we are in part 2.1 of a study presenting a permafrost model, with no mention to the snow scheme yet). Yet it looks like a description of a snow SEB, as evidenced by the expression “varying snow regime” and by the fact that the equation calculates energy for melt, which is inappropriate for a ground surface scheme. For a ground surface scheme, we want the SEB to give us access to the dE/dt of the surface so that it can force the heat diffusion/advection in the soil column. If we do not have that, we do not have the coupling between the climate and the temperature in the ground. This problem persist over the whole section 2.1. These are very important aspect of the model description that need to be carefully addressed so that we can understand what we are talking about.

L78 and 86

I have 2 problems with this quantification of the turbulent fluxes. First, if you take Preistley and Taylor for Q_e, then you have to take Q_h as the residual of the available energy for turbulent fluxes, which is:

Q_h=(1-α ∆/(∆+γ))×(Q_n-Q_g )

Otherwise your SEB is not energy conservative. Preistley and Taylor considers that the energy available for both turbulent fluxes is Q_n - Q_g and provide a formula to find what fraction of that available energy goes to the latent flux. So, in order to have a consistent surface energy balance budget, Q_h has to be the complementary fraction of Q_n - Q_g, not another formula based on another theory.

Second, if you use Preistley and Taylor you will assume that the sum of the terms of your SEB equals 0 and not the energy variation of the soil surface. Otherwise you’d have to work with a modified version that would look like:

Q_e=α ∆/(∆+γ) (Q_n-Q_g-(∂E_surf)/∂t)

But then you would not be able to calculate together the turbulent flux and the energy variation of the surface (one too many unknown). Yet, it is through the energy variation of the surface that you can couple the SEB and the energy budget at depth in the ground (because the energy budget of the surface will drive the one of the subsurface). Therefore, this whole SEB description is very confusing to me and absolutely need to be fixed. For now I cannot understand the coupling with the subsurface.

Specific comments

L14:    “shallow soil columns” indicate typical depth

L36:    Wrong reference with Lan et al. 2025 here ? It is a review of the strength and weakness of a climate reanalysis datasets, it does not present a permafrost model.

L43:     I am surprised the authors do not mention other models used to study Tibetan permafrost like the GBEHM model that is greatly used within the Chinese community (Fang et al., 2025; Gao et al., 2018; Qin et al., 2017; Shi et al., 2020; Wang et al., 2023, 2018; Wang and Gao, 2022, 2025; Yang et al., 2023b, a), or the recent DHTC model (Linmao et al., 2024) and FLEXTopo-FS model (Gao et al., 2022).

Eq13:    I would avoid writing the equation, it gives the impression that your are describing how the model works during simulations whereas you just used the equation once for the evaluation of static parameters. I would rather just state the values of k and C in the text and say that they were calculated based on the ref you mention.

Fig.2:    “Comparison of simulated active layer soil temperature with time series at the synthesis sites.” What is the methodology here? The active layer can be pretty deep. You averaged the temperature over the whole the active layer at a daily time step? Also what do the red and blue numbers with and without parenthesis correspond to? I assume from the text that it is the bias ? It should be written in the caption for more clarity.

Fig.3:    The methodology described in the caption is hard to understand, please elaborate more.

Phrasing and typos

L214:    Align

L218:    “via the processes of latent heat and soil moisture” the reader understands, but these are not processes, please rephrase.

L221:    “The remote-sensing datasets are different in temporal coverage, so we use the climatology to represent the long-term conditions.” I think the phrasing “use the climatology” can be improved.

Fig.2:    “The daily soil temperature present are averaged”, syntax problem.

 

References

 

Fang, P., Wang, T., Yang, D., Yang, J., and Tang, L.: Permafrost Degradation and Concomitant Hydrological Changes Dominated by Anthropogenic Greenhouse Gas Emissions in the Northeastern Tibetan Plateau, Geophysical Research Letters, 52, e2024GL113679, https://doi.org/10.1029/2024GL113679, 2025.

Gao, B., Yang, D., Qin, Y., Wang, Y., Li, H., Zhang, Y., and Zhang, T.: Change in frozen soils and its effect on regional hydrology, upper Heihe basin, northeastern Qinghai-Tibetan Plateau, Cryosphere, 12, 657–673, https://doi.org/10.5194/tc-12-657-2018, 2018.

Gao, H., Han, C., Chen, R., Feng, Z., Wang, K., Fenicia, F., and Savenije, H.: Frozen soil hydrological modeling for a mountainous catchment northeast of the Qinghai–Tibet Plateau, Hydrology and Earth System Sciences, 26, 4187–4208, https://doi.org/10.5194/hess-26-4187-2022, 2022.

Linmao, G., Genxu, W., Chunlin, S., Shouqin, S., Kai, L., Jinlong, L., Yang, L., Biying, Z., Jiapei, M., and Peng, H.: Development of a modular distributed hydro-thermal coupled hydrological model for cold regions, Journal of Hydrology, 644, 132099, https://doi.org/10.1016/j.jhydrol.2024.132099, 2024.

Qin, Y., Wu, T., Zhao, L., Wu, X., Li, R., Xie, C., Pang, Q., Hu, G., Qiao, Y., Zhao, G., Liu, G., Zhu, X., and Hao, J.: Numerical Modeling of the Active Layer Thickness and Permafrost Thermal State Across Qinghai‐Tibetan Plateau, JGR Atmospheres, 122, https://doi.org/10.1002/2017JD026858, 2017.

Shi, R., Yang, H., and Yang, D.: Spatiotemporal variations in frozen ground and their impacts on hydrological components in the source region of the Yangtze River, Journal of Hydrology, 590, 125237, https://doi.org/10.1016/j.jhydrol.2020.125237, 2020.

Wang, T., Yang, D., Yang, Y., Zheng, G., Jin, H., Li, X., Yao, T., and Cheng, G.: Pervasive Permafrost Thaw Exacerbates Future Risk of Water Shortage Across the Tibetan Plateau, Earth’s Future, 11, e2022EF003463, https://doi.org/10.1029/2022EF003463, 2023.

Wang, X. and Gao, B.: Frozen soil change and its impact on hydrological processes in the Qinghai Lake Basin, the Qinghai-Tibetan Plateau, China, Journal of Hydrology: Regional Studies, 39, 100993, https://doi.org/10.1016/j.ejrh.2022.100993, 2022.

Wang, X. and Gao, B.: Future hydrological changes of the upstream basin and their effects on the water level of the Qinghai Lake, Journal of Hydrology: Regional Studies, 59, 102425, https://doi.org/10.1016/j.ejrh.2025.102425, 2025.

Wang, Y., Yang, H., Gao, B., Wang, T., Qin, Y., and Yang, D.: Frozen ground degradation may reduce future runoff in the headwaters of an inland river on the northeastern Tibetan Plateau, Journal of Hydrology, 564, 1153–1164, https://doi.org/10.1016/j.jhydrol.2018.07.078, 2018.

Yang, J., Wang, T., and Yang, D.: Divergent responses of permafrost degradation to precipitation increases at different seasons on the eastern Qinghai–Tibet Plateau based on modeling approach, Environ. Res. Lett., 18, 094038, https://doi.org/10.1088/1748-9326/acf05c, 2023a.

Yang, J., Wang, T., Yang, D., and Yang, Y.: Insights into runoff changes in the source region of Yellow River under frozen ground degradation, Journal of Hydrology, 617, 128892, https://doi.org/10.1016/j.jhydrol.2022.128892, 2023b.

The manuscript presents a new simulation framework for large scale numerical simulation of permafrost dynamics, and apply it to the quantification of permafrost metrics (ALT, MAGT, extent) of the permafrost cover over the Tibetan Plateau from 1950 to 2023. The presented simulations are suffering of strong assumptions regarding soil water content and snow cover (both ‘static’), which in my opinion hampers the possibility of temporal evolution analysis. The bibliography of the permafrost modelling landscape is also incomplete.

I do not think that the manuscript may be published in TC in its present form. A significant work for better discussing the limitations of the simulations and put them in the context of permafrost modelling across scales is needed. Thus I recommend a major revision prior to reconsider whether or not it may be published in TC.

L1 : “Permafrost remains a largely subsurface phenomenon” Clumsy. Permafrost is a subsurface phenomenon. I guess the authors want to point out the difficulty of direct observation of this subsurface phenomena as the reason why its understanding largely relies on numerical simulations. First sentence to be rephrased.

L 31-43 : An important part of the permafrost modelling landscape is overlooked in the bibliographical survey given in the introduction : the cryohydrogeological simulators (e.g.: Grenier et al., 2018, Hu et al., 2023). These mechanistic models are based on the numerical resolution of the equations on the continuum mechanics, and thus have a much bigger predictive potential than conceptual, calibrated models. I think that, for the sake of completeness, this type of model should also be included in the survey.

Grenier et al., 2018 “Groundwater flow and heat transport for systems undergoing freeze-thaw: Intercomparison of numerical simulators for 2D test cases”, https://doi.org/10.1016/j.advwatres.2018.02.001

Hu et al., 2023. “ Water and heat coupling processes and its simulation in frozen soils: Current status and future research directions”, https://doi.org/10.1016/j.catena.2022.106844

L 36 : Lan et al., 2025 seems to be a reference related to a reanalysis, not to a model. Reanalysis are built using models, but they are not models.

L39 : “and influences” I am not sure about what is meant here. To delete, or to be rephrased.

L 47 : “Specially” : Specifically

L 64 : given the large number of symbols used, I recommend to put the table of symbols with full names and unist in the beginning of the manuscript, or at least in the beginning of section 2, rather than in Appendix.

L 135 – 152 : According to equations (16) to (19), soil water content does impacts heat transfers. Meanwhile, no information is given on how is handled hydrology in FPM. This should be discussed here.

L 155-156 : Why these numbers of layers and these thicknesses of grid cells? Any convergence study for justifying these choices?

L 162 : “the static soil moisture is used”. This is an extremely strong assumption, eliminating seasonal dynamics (e.g.: wet season vs dry season) and inter-annual variability (e.g.: dry years vs wet years). Given the importance of soil water content and state for heat transfer properties, this is likely to generate important errors and biases in the result of permafrost dynamics. See for instance Clayton et al., 2021 for the impact of soil moisture distribution on active layer thickness. See also de Vrese et al., 2023 for a study of hydrology – related biases in large scale permafrost modelling. Anyway the manuscript does not give enough information for clearly understanding what is assumed here. According to table 1, only the 2015-2022 period has a complete set of five remote sensing products. Then what is done exactly ? Is the soil moisture in a given pixel considered to be constant equal to the mean of the five 2015-2022 multi-annual averages of each product ?

Leah K Clayton et al 2021, Active layer thickness as a function of soil water content, Environ. Res. Lett. 16 055028 DOI 10.1088/1748-9326/abfa4c

de Vrese, P., Georgievski, G., Gonzalez Rouco, J. F., Notz, D., Stacke, T., Steinert, N. J., Wilkenskjeld, S., and Brovkin, V.: Representation of soil hydrology in permafrost regions may explain large part of inter-model spread in simulated Arctic and subarctic climate, The Cryosphere, 17, 2095–2118, https://doi.org/10.5194/tc-17-2095-2023, 2023.

l 255-256 : “The simulated soil temperature was significantly improved by 2.1 ◦ C, indicating FPM could be improved with more reliable climate forcing and soil profile (Fig. 2).” Interesting. I think that this is a direction to follow to improve the manuscript : study on how to improve LSM permafrost simulations?

l 274-297 : Sections 5.2 to 5.4. Here temporal evolution of ALT, MAGT and permafrost extent are proposed for the period 1950-2023, with two contrasted periods, 1950-1980 and 1980-2023. I have strong concerns over the validity of any temporal evolution analysis while keeping the soil water content constant equal to an estimate based on 2015-2022 products (see my comment on l 162). At least, the way precipitation and evapotranspiration (and thus the overall water balance) have evolved during the whole considered period should be presented. Then the impact of assuming a time constant soil moisture profile should be discussed at the light of this information.

L 301-304 : “In fact, permafrost simulations are hampered by reduced reanalyses quality in cold regions primarily due to inherent challenges in representing nonlinear processes involving ice, or its phase change near 0 ◦ C (Cao and Gruber, 2025). The poorly described soil column, especially the soil organic matter, put additional uncertainty for permafrost simulations.” I insist here on the key role of hydrology, and the especially of the water transfers within the soil colum.

L 309-310 : “The static snow density was used to represent the overall conditions during the snow-covered period.” Most likely concerns analogous to those I rose about static soil water content could be raised about considering a static snow cover. I recommend to make also a study of the evolution of properties of the snow cover over the study period, and to discuss the impacts of assuming a static snow cover on the basis of this information.

L 358 : “Our simulations indicate that current land surface models employing shallow soil columns are inadequate for permafrost research on the Tibetan Plateau, since they have generally underestimated permafrost extent while overestimating degradation rates. Such inadequacy may also pose challenges in other regions characterized by deep active layers (i.e., > 3m);” I don’t think that I saw any data or figure that give a quantitative basis for this statement, such as a comparison between good modelling results with thick soil column vs bad modelling results with shallow soil column. I do not want to say that the statement is not correct, just that it is not clearly established in the manuscript.

L 614-615 : a manuscript cannot cite itself.