This paper presents the ambitious plan of the Multi-Compartment Hg Modeling and Analysis Project (MCHgMAP) for a multi-model ensemble approach to supporting the Minamata Convention as it seeks to assess the degree to which international efforts are achieving the convention’s  goals of reducing Hg in the environment. Since anthropogenic emissions to the atmosphere that lead to transboundary air pollution are the subject of another international treaty, the Convention on Long-Range Transboundary Air Pollution consideration is given to those policy goals as well.

The scope of the task envisioned by the MCHgMAP is huge. The length and the depth of the paper is commensurate to that task. Despite the paper’s great length, and the complexity of the issues that it brings up, the paper has a clear structure that keeps the many parts of the paper from getting in the way of each other. The text is also commendably well-written, with just a few places where some clarification is needed. The survey of the literature is similarly well-done,  building confidence that the knowledge base going into the modeling exercise draws on much of the relevant literature – even though that literature is developing rapidly – which is a reflection of how the understanding of Hg cycling continues to develop.

I have two major concerns, as well as some smaller points for the authors to consider.

The first of the major concerns regards the value of providing examples of similar efforts to use multi-model ensembles to support policy efforts. As a reader feeling overwhelmed by the scope of what is to be attempted, and trying to get a clearer vision of where the overall process presented in this paper is heading, it would be helpful to be given other examples where MME’s have been applied. (I presume the work of the IPCC has such examples.) So please try to present what is envisaged for supporting the Minamata Convention in relation to what has been achieved in modeling other global scale pollution issues to support the policy process.

My other suggestion concerns Section 8 where I am missing treatment of the Terrestrial-Hydrological modeling. Earlier in the paper, three classes of models are presented (Atmosphere, Ocean and Terrestrial-Hydrological models (Sections 2.1-2.3, and the “Contemporary Cycle” in Figure 1.). These all seem to have a place in the lower part of Fig 1, the “Long-Term Cycle” of “Multi-Media Mass Balance models”. But when one gets to “Section 8 Modeling analysis: process-understanding”, there are two sections, 8.1 on atmospheric models and 8.2 on ocean models. The issues addressed for these two classes of models would have been appropriate for Terrestrial-Hydrological Models as well. More generally, I find the presentation of the Terrestrial-Hydrological modeling as hard to understand fitting into the global scope of the MCHgMAP, since there are so many more potential compartments in the terrestrial-hydrological modeling domain compared to the ocean or atmosphere..

Below are more minor points for the authors to consider:

Line 1615. There will be perturbation modeling and sensitivity modeling. It seemed hard for me to understand the need to differentiate these, since the two seem so closely related.

I have a sense that a major challenge for global Hg modeling is the apparent discrepancy between increasing anthropogenic emission inventories during recent decades, and the regional decline in atmospheric Hg over Europe. If this is indeed the case in the eyes of the authors, it would be nice to see this lifted up as an example of the type of issue that the MCHgMAP will address.

In section 2.3, a number of terrestrial hydrological models are mentioned. I missed one published in Environmental Science and Technology that seems appropriate to include: Eklof, Karin, et al. "Parsimonious model for simulating total mercury and methylmercury in boreal streams based on riparian flow paths and seasonality." Environ. Sci. Technol. 49.13 (2015): 7851-7859.

Lines 1212-1218. Here the authors state that

“It is currently not possible to determine the origin (anthropogenic/geogenic) or period of origin of the primary emission of Hg that are currently being reemitted from the ocean based on traditional measurements of Hg concentrations (for an explanation on how the Hg(0) evasion flux is determined see Sect. 4.4 and Appendix B). However, both box and 3D models can be used to gain insights into the extent of the Hg(0) evasion flux from the ocean, the fraction that different sources or historic periods contribute, and how concentrations will change in response to changes in Hg ocean inputs as a result of measures under the Hg MEAs.”.



This would seem like great place to mention the use of Hg isotopes to resolve these issues as Jiskra et al. did in their 2021 Nature publication. That article is cited in the first paragraph of the papers introduction (line 75), and one other place, but not in relation to where it is most relevant, specifically in this section at line 1212-1218. I am also concerned that the power of isotopes is not going to be brought to bear in the MCHgMAP as fully as they could be. But I expect that is something that will have to be wrestled with when looking at the results of the MCHgMAP.

In the reference list there are three references where the first author is Li, C. But there are actually two different people involved. So good that their names are distinguished…

Line 1506 begins a new paragraph. It was difficult to understand that this new paragraph had started taking about lake sediments. (The previous paragraph was about tree-rings).

This reviewer commends the authors’ endeavor to have produced such an ambitious proposal on multi-medium modeling and analysis to support the Minamata Convention and CLRTAP. The document is well structured for the most part, and the literature is fairly up to date. Below are a couple of main comments.

1. This document is essentially a proposal, which should be clearly reflected in the title, and in the same vein, the approach or the set of modeling experiments proposed were not exactly “developed”, which implied evaluation or demonstration of validity and effectiveness.
2. The goal, objectives and the time periods for the present day and the future scenario should be presented upfront, rather than being buried two-thirds of the way through the document.
3. Did I miss whether the physical input data for all models were to be harmonized? For example, interannual variability in meteorology can significantly impact transport, deposition, chemistry, and emissions/releases. To quantify the impact of changing emissions, the influence of physical parameters needs to be assessed.
4. The difference between “perturbation” and “sensitivity” runs appeared to be rather subtle (L1615-1621). The latter seemed to encompass the former. They might need to be better defined.
5. Sections 5 and 7 included not only atmospheric and ocean models but also other environmental mass balance models. Sections 6 and 8 appeared to include atmospheric and ocean models only.
6. Provide a glossary of terminology used in the document.

Next are specific comments:

Abstract: L66, L67: Instead of "is developed", it might be more appropriate to say "is proposed".

L156: What is the difference between “emission” and “release”?

L201-202: For regional models, BCs can be the most dominant factors in determining GEM concentrations.

L202: There is a "constant" ring to "prescribe", but that’s not exactly how emission fluxes are handled in some models.

L224-226: A most recent version of CMAQ-newHg-Br v2 included a newly developed gas-particle partitioning scheme with most up-to-date gas-phase redox reactions (Wu et al., 2024, JAMES, https://doi.org/10.1029/2023MS003823), which is probably one of the latest breakthroughs in atmospheric Hg modeling.  

L394: What does the term “secondary Hg emission” refer to? Legacy emission? Release? Or chemically formed?

L472-477: These definitions should be presented upfront, instead of being buried here after having already used such terms in the preceding text. Or provide a glossary as suggested in one of my main comments.

L779-L784: This part is confusing. It was said that 2010 should be used as the baseline year. Then later, in scenario design, the period of 2010-2020 was said to be the present-day period.

L846: What is the “mass-flow” approach? Add a reference.

L944: Under “Scenario design”, there’s no scenario design. It’s unclear what was meant to be shown there.

L1124-1127: On what was this statement based? This statement is valid only if those parameterizations were correct, but are they? It seems unclear which causes larger uncertainty: the air-surface exchange models or scaling up measured fluxes, unless work is done to compare the two.

L1298-1299: It's such a popular result. The dry deposition flux of Hg° may exceed that of HgII, but it doesn't make the former more "important" than the latter. The question is how much of the deposited Hg° gets oxidized.

L1584: Present: 2010-2020; Future: Spell out the simulation period. It was said to be “up to 2050” - unclear.

L1605-1613: Simulation objectives: Came too late and buried.

L1625: What is “an analogous strategy”?

L1807: Why interpolate twice?

L1948-1949: “for different years of the period 2010-2020 to reveal their temporal changes and causes” -  It should probably just be said as  “interannual variability and its driving factors”.

L1950: “On the annual basis and for different seasons” – Awkward wording.

L2271: “Coupled offline”: If it’s coupled, then it’s online. If it’s offline, then it’s decoupled.

This manuscript provides a very ambitious and comprehensive description of the current status of global multi-compartment modelling of mercury and the key anthropogenic and environmental processes that are necessary to describe in the modelling. The initiative and the ambitious approach is driven by a strong policy need and thus represents an important step to provide a scientifically sound approach to answer policy-relevant questions and to provide support for developing common and realistic expectations of effects of emission reductions as well as for setting common and realistic policy targets. Descriptions of the relevant policy frameworks (Minamata Convention, LRTAP convention) are included as are the direct policy related questions posed by these frameworks. 

The manuscript provides and excellent state-of-the-art review of mercury (occurrence, transformations, trends etc.) in the atmosphere and oceans, and to some extent for terrestrial systems, and provides a basis for designing the multi-compartment modelling which is described in the following sections. Uncertainties and lack of knowledge are also included. 

A main part of the manuscript contains a description of the plans for modelling activities and how the results will be evaluated. This is quite lengthy but necessary to provide the scientific community to engage, comment/criticize and potentially contribute to the planned activities.

In summary, the opinion of this reviewer is that the publication of this manuscript will make an important contribution to the scientific understanding and development of models for global mercury cycling, and to supporting policy efforts to reduce mercury emissions and impacts globally.  

Some specific comments/suggestions are:

- An important factor not described in detail is the importance of "permanent" sinks for mercury (oceanic sediments, deep soil layers etc), i.e. what is the current status of knowledge of these sinks, are they permanent and how can they be described in the models? From a policy perspective, a common understanding of the global permanent sinks and how important they are for setting the pace of reducing environmental mercury burdens in a scenario of emission reductions would be beneficial.

- In general, the current understanding of terrestrial mercury cycling/storage and factors that influence the mobility/releases of mercury to the aquatic compartments are very limited in comparison to the atmospheric and (to some extent) oceanic cycling and, consequently, the possibilities for modelling are limited. This is also reflected in the initial state-of-the-art sections. Nevertheless, a description of how mercury in the terrestrial compartment will be handled in the modelling activities would be beneficial. Soil-air, and vegetation-air exchange is covered but releases of soil-bound mercury to surface waters are not. For instance, land-use change (e.g. forestry) and climate (increased precipitation) are factors that can influence mercury (and methylmercury) burdens in freshwater and coastal systems and thus cause increased mercury concentrations in fish and seafood, the primary exposure route for most population groups.

- A large part of the manuscript concerns the ambitious plans for modelling activities and evaluation of results. As mentioned above, this is very positive  but it would also be of interest for the reader (and for potential contributors) to understand if these plans are fixed and funded or if they are more of a long-term ambition, e.g. a brief description of the MCHgMAP initiative.