This is a very needed, detailed, and well-documented paper developing new Hg scenarios (and an underlying model) that is sorely needed in the community. The underlying work is of high-quality and ACP is an appropriate journal for it.

Before publication, however, there are two substantial issues that should be addressed. 

1) There are several substantive errors and misstatements on the requirements of the Minamata Convention, some of which are just in the writing, but some of which also percolate into the analysis and the technical analysis. These are listed below in detailed comments. The most significant conceptual one is that scenarios are defined as BAT/BEP scenarios, but as far as I can tell this doesn't actually follow the Minamata definition (which depends on economic considerations and qualitative feasibility concerns which aren't taken into account in the modeling done here). This should be clarified, especially where BAT/BEP is used to imply mercury-specific controls (where in many regions, countries may define BAT/BEP as co-benefits in practice, especially in the near term). This makes the discussion of implications a bit more difficult, but is fixable I think with some careful attention to language and examination of the degree to which this might affect the underlying analysis. 

2) The climate scenarios included are (somewhat infeasibly) optimistic. This is fine in principle, but uncommented on (at least the paper should note that these are very low C scenarios); plus, it might be more useful for policy-making for a more realistic set of scenarios to be available. 

I don't think either of these changes require a lot of additional work (although I'd argue that showing a climate policy scenario that's somewhat less infeasible would really add value). But they are substantive, and the paper as it stands really needs to be corrected. Finally, I would recommend that some edits to be made to the discussion in section 5 to make the take-home messages clearer and more apparent to the reader.

Specific comments follow:

Line 8: this is incorrect, the Minamata Convention does not ban mercury trade. Please refer to Article 3 -- there is a Prior Informed Consent procedure. Relatedly, it would be more accurate to say that the Convention bans certain mercury uses, or specific mercury uses (here, the language could be interpreted to apply to all Hg uses)

Lines 8-10: these "co-benefits" are specifically set out for in the Minamata Convention, please see Article 8 (5d) and related BAT/BEP guidance

Line 21: "lie in the in" -- typos here

Line 60-61: the language here implies that it is both the volatility and reactivity that facilitate long-range atmospheric transport, whereas in actuality the reactivity decreases propensity to transport. This sentence could be edited to ensure accuracy.

Line 63-65: Long-range Hg damages were known before 2002 -- in particular, Hg was considered by CLRTAP in the mid-1980s. Again, here, rephrase to ensure accuracy. 

Line 74: the Minamata Convention defines releases as those to water and land, and emissions as those to air. Thus, this sentence is not fully correct -- I would recommend using these terms consistently with how the Convention defines them. 

Line 78-80: addressing releases does not require trade bans. The citation here to Giang et al. 2015 seems incorrectly placed.

Line 100: it might be useful here to note how incredibly old the SRES scenarios are (this could be accomplished by citing them to their original source)

Lines 97-112: beyond describing the existing scenarios it would be useful here to describe, at least qualitatively, what they project for emissions in terms of range

Line 165: it would be useful to describe the iPOG tool here for the unfamiliar reader (or at least expand the acronym)

Line 213-217: these are all very optimistic climate scenarios in terms of energy use. How does this compare with the current estimates of CO2 emissions from the Paris Agreement?

Line 237: this sentence implies that VCM phase-outs are mandated by the Minamata Convention, but that is incorrect. 

Line 424-425: it's interesting that the Hg reductions occur at a lower rate than CO2. Is this because of other sources? It's hard from the graph to determine the extent to which there is a proportional association for combustion sources with CO2 decreases? If not other sources, what is the intuition behind this? Perhaps there could be a plot of the ratio with time (maybe in the SI?)

Line 509-512: it would be interesting to compare this estimate quantiatively with previous estimates of emissions not covered by the Minamata Convention -- in the 2017 ICMGP synthesis in Ambio, the policy paper estimated this quantity at 97 Mg. This seems comparable, but maybe this estimate is a bit bigger? Is there a key sector missing in the previous study, or is this just a function of a different base year emissions inventory? 

Public access: please check to make sure the code is accessible upon official publication. Currently, the link refers to a password protected page that is only accessible to collaborators. 

This study represents a milestone in the literature. The mercury community, both from the scientific and policy sides, waited for a similar paper for more than a decade. Of course, other similar studies were published, although only at a limited regional or sectoral scope. There is plenty of very useful information for any researcher in the area. On the flip side of the coin, it is difficult to thoroughly review all of the technical aspects covered by the paper. However, the methodology and the tool used have been very consolidated through the years and do not require a detailed review in this regard. Therefore, I will focus on the aspects I have personally faced in the recent past in my research activities.  

In this regard, the only critical issue I can see is not methodological and does not affect the quality or validity of the paper, but in my opinion requires it to be addressed by the authors before publication. It is strictly linked to ASGM, their Hg emission estimates, and, in particular, the associated uncertainty. The aspects of uncertainty associated with ASGM Hg emissions are covered through the text in many sections; however, this is not the case for the associated implications. The most critical factor is that, considering the relative weight of ASGM Hg emissions, the differences between the scenarios designed and depicted in Figure 3 (upper panel), Figure 4 (regional panel where ASGM apply), and Figure 5 are likely to be completely covered by the associated uncertainty (of ASGM Hg emissions). At first sight, this makes de facto indistinguishable most of the scenarios (except for 12_CLIM2_HgMFR, which considers the utopian ban of ASGM impossible for many reasons). One would argue that this makes the study somewhat inconclusive. Conversely, I believe this is a very important outcome to be considered by policymakers.

As a reader, I would expect error bars to be present in the indicated figures; however, as a researcher, I believe I understand their absence. I could suggest authors realise a version of the figures (at least figure 5) that excludes ASGM, allowing a detailed evaluation of the scenarios for the other sectors.

ASGM is a complex social phenomenon rather than an industrial trend; linking it to large-scale gold makes sense for many aspects and countries but can be irrelevant for others. Of course, the authors have to make a choice, and this does not affect the quality of the paper or its publication. However, I suggest the authors underscore all these aspects in the pertinent sections and in the conclusion and expand par. 4  of Section 5.7 (line 547).