This is a well-written and timely paper that provides a comprehensive overview of two decades of progress in ParFlow–land surface model coupling. It offers clear insights and a thoughtful vision for future sustainable model integration. I’ve included a few specific suggestions and questions to help improve clarity and flow in certain sections, but overall, I find the paper to be of high quality and fully supportive of its publication after minor revisions.

Abstract

• Lines 17–24: The first two sentences are quite long. Consider splitting after “Earth system models (ESMs)” to improve readability.
  
  Suggestion: “Groundwater plays a vital role in terrestrial water and energy cycles. Yet, it remains oversimplified in most Earth system models (ESMs)...”

Introduction

• Lines 44–52: The sentence from “It introduced the coupling…” to “…incorporating physically based groundwater dynamics…” is long and dense.
• Line 66: The phrase “groundwater models are often run offline with limited interaction…” could be clearer if you add “thereby missing dynamic feedbacks with the land–atmosphere system.”
• Lines 77–87: The transition from describing model evolution to the need for a “sustainable framework” is abrupt.
  
  
  
  

Section 2 – Review of Coupled Modeling

• Line 109: “Three key areas” — consider making these bold or in numbered format (1), (2), (3) for visual clarity.
• Lines 116–124: The paragraph is rich in technical context but could use a short topic sentence like “Groundwater strongly modulates the surface energy balance through its effect on soil moisture distribution.”
• Lines 147–153: Consider simplifying the long sentence beginning with “Simulation results showed…” into 2–3 shorter sentences.
• Lines 213–224: The description of rooting depth could be condensed or summarized for focus; much of the detail could be cited rather than explained.

Section 3 – Foundational Recoupling

• Lines 319–324: The transition from “domain description” to “vertical discretization” is abrupt.
• Lines 378–386: The sentence “The overall advancement in model performance…” is strong; consider adding a quantitative statement (e.g., RMSE reduction or bias improvement) if available.

Section 4 – Sustainable Framework

• Lines 470–483: Clarify how “logging system” differs from a typical software version control or provenance tracker — perhaps add a brief example.

Section 5 – PLCMIP

• Lines 502–517: The description of existing models could benefit from a small schematic figure showing coupling hierarchies.
• Lines 529–537: Replace “To quantify… To compare… To identify…” with bullet points for easier reading.

Section 6 – Summary

• Lines 547–558: Strong conclusion, but it repeats parts of the abstract.
  
  Suggestion: Focus this section on future directions — e.g., “Key next steps include expanding the recoupling to other LSMs, refining coupler standards, and launching PLCMIP.”
• Table 1. I would also include this paper published recently on ParFlow-NASA/LIS-Noah-MP. Abbaszadeh, P., Zaouna Maina, F., Yang, C., Rosen, D., Kumar, S., Rodell, M., and Maxwell, R.: Coupling the ParFlow Integrated Hydrology Model within the NASA Land Information System: a case study over the Upper Colorado River Basin, Hydrol. Earth Syst. Sci., 29, 5429–5452, https://doi.org/10.5194/hess-29-5429-2025, 2025.

Lines 299–306: When you mention “basic water and energy modules of CoLM,” which specific subroutines were included or excluded (e.g., snow, plant hydraulics)?

Lines 343–356: Why was 2018 chosen as the baseline year? Does this year represent typical hydrometeorological conditions for the domain?

Lines 379–381: You note timing discrepancies in SWE accumulation. Have you explored adjusting the snow albedo or canopy interception parameters to resolve this?

Lines 395–398: Could the deeper water table in CoLM/PF be partly influenced by changes in infiltration or soil parameterizations, rather than transpiration alone?

Lines 531–537: Will PLCMIP include standardized benchmark datasets or synthetic cases (similar to IH-MIP) to ensure comparability?

Thank you for the opportunity to review the manuscript titled “Twenty Years of Trials and Insights: Bridging Legacy and Next Generation in ParFlow and Land Surface Model Coupling” by Yang et al. In this manuscript, the authors first reviewed milestones in coupling a groundwater model, ParFlow, to land surface models and atmosphere models. Then the authors presented their preliminary results in updating ParFlow-CLM to the most recent version of ParFlow-CoLM. They ended by reviewing current models that have incorporated ParFlow, and providing high-level recommendations for future model coupling efforts. My major and minor comments are detailed below.

1. I struggle to decipher the type of this article. It was not clear whether this is intended to be a review, insight, or research paper. The current presentation has a feel of mixture. Section 2 is a review of previous ParFlow-CLM and ParFlow-WRF efforts. Section 3 seems to reporting the preliminary ParFlow-CoLM efforts. Section 4 outlines insights moving forward, although a brief table reviewing other ParFlow efforts were also included. Readers would prefer the authors having a clear target where they want this manuscript to be and do a good job at that goal, instead of squeezing 3 very different goals in one paper but only scratching the surface. Personally, I would be very excited about the insights component. I wish the authors could detail their implementation of the sustainable coupling, which would be of great benefit to the community as ParFlow or the LSMs will upgrade inevitably.
2. The scope of this work seems very narrow. The authors focused on parflow-CLM and parflow-WRF, both of which are efforts led by the author team themselves. Some other modeling efforts were also mentioned, but to an extremely brief extent (one sentence, or one line in a table). To make it appealing to the broader community, I suggest the authors looked into work beyond their own. For instance, in addition to CLM, several other LSMs have also be coupled to ParFlow. In addition to ParFlow serving as the subsurface boundary, other groundwater models are also being coupled to LSMs to study similar processes – influence of groundwater on surface fluxes.
3. The section on the preliminary coupling of CoLM with ParFlow is weak in several aspects. 1) It was not clear in this coupling if the authors had implemented a coupler to make future coupling efforts easier. If they only replaced CLM with the updated CoLM, the model development itself is incremental. 2) The test case was also poorly designed. It was not clear why the authors decided to use the model setup in China but the met forcing in Colorado. It was not clear which region did the authors use for model validation. Does that indicate the subsurface setup or the met forcing does not matter that much? 3) A significant amount of results from this section is essentially about the difference between CLM and CoLM, which is not really new results from this study. Instead, papers from the CoLM development should be appropriately cited. 4) since ParFlow-CLM in 2005, ParFlow has also went through several updates as well. It would be helpful to summarize parflow updates and report the influence of model simulations as well.
4. The 4^th section on sustainable recoupling framework was too high level and not very specific. I think the community would be more interested in how a sustainable framework has been or can be implemented to ease the process of model upgrades.
5. regarding the section on the model inter-comparison project, I have some thoughts (not as criticism). I agree the model comparison would be useful exercise. But I think before making model inter-comparison, it would be important to assess model parameter sensitivity. Model development is much more expensive, compared to parameter change. Before we can make any attribution that model development/structure is causing the difference, we need to ensure that the difference cannot be resolved by model parameterization.

Specific comments.

Figure 1. I did not find this figure from the cited reference paper. Please check.

Line 377 why evaluating against model reanalyzed data. why not using a EC tower? where is the validation data located? 

Line 399 these are new changes in CLM, and should be properly cited,

Line 418 main changes in what? CLM or ParFlow.