the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Modelling the Fate of Mercury Emissions from Artisanal and Small Scale Gold Mining
Abstract. A mercury (Hg) tracer model based on WRF-Chem has been developed to provide a rapid and relatively simple tool to evaluate the local and regional impact of Artisanal and Small Scale Gold Mining activities which use Hg amalgamation.Artisanal gold mining, its use of mercury amalgamation and its potential for human and environmental harm is mentioned specifically in the Foreword by the Secretary-General of the United Nations António Guterres of the Minamata Convention on Mercury: Text and Annexes (2023 edition, url below). Much of this artisanal mining occurs in the Tropics, and often in densely forested regions and the role of vegetation in the global atmospheric mercury cycle has been shown to be significant in recent years. The model employs a simple lifetime approach to Hg0 oxidation based on KPP and an ad hoc deposition scheme which calculates the foliar uptake of Hg0(g) based on the Leaf Area Index, dry deposition of HgII(g) , and the wet deposition of HgII(g) by convective and non-convective precipitation. A number of demonstration simulations are presented using four example domains from South-East Asia and South America, and five from Africa. The results highlight the diversity of the local impacts of ASGM due to land category, geography and meteorology, but also point to the fact that just as there are always local impacts there are also repercussions for the global atmospheric mercury burden (https://minamataconvention.org/sites/default/files/documents/information_document/Minamata-Convention-booklet-Oct2023-EN.pdf).
- Preprint
(24989 KB) - Metadata XML
-
Supplement
(31667 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-861', Anonymous Referee #1, 01 May 2024
General Comments
The study presents a new tool based in WRF-Chem to assess the local and regional deposition of mercury (Hg) emitted by Artisanal and Small-Scale Gold Mining. I appreciate the potential of such a tool for assessing the role of ASGM in the global mercury cycle. However, the manuscript is hindered by its presentation quality and clarity/justification of methodology, as described in the Main Issues below. In its current state, I do not think it is appropriate for publication in GMD.
Main issues
Introduction:
The introduction is written in an unclear fashion, with short paragraphs jumping between diverse topics that are not well connected. For example, it is not clear why vegetation is relevant to the ASGM problem until almost the end of the vegetation section (L46). Secondly, from L50–53 it is unclear whether this mercury model has already been developed by Grell et al. (2005) or this is a novel contribution by this study. Subsequently, L54–55 intersperses with a different topic (rainfall in the tropics) before returning to the model in the study. I believe that the introduction needs to do a better job at introducing what we know about the ASGM issue (supported by data in the literature, e.g., global emissions, distribution of emissions etc.) and previous modelling efforts/capabilities.
Methods:
In their parametrization of vegetation uptake, the authors calculate uptake rates based on dividing global estimates for Hg uptake by vegetation by global average LAI. I find this approach flawed, as both observational (e.g., Zhou et al., 2021; Zhou and Obrist, 2021) and modeling (e.g., Yuan et al., 2023; Feinberg et al., 2022) studies have highlighted that Hg uptake varies strongly spatially, depending not only on LAI but also on forest type and region. The authors are studying Hg dynamics in tropical and subtropical areas (which have both much higher uptake and LAI than the global average), which I do not think will be represented well by the global averages. There are many published parametrizations already available to calculate deposition velocities based on LAI, so I do not understand why the authors chose such a primitive approach.
Presentation of results:
The discussion of results is very difficult to follow. The main paper includes 5 Tables (including many columns) + 9 Figures. This is in addition to 3 appendix tables and 21 supplementary figures. The text does not cover many of these figures, so it is unclear why they are all necessary. The reader is left mainly on their own to decipher the many results in the figures and tables.
Relatively much text (L190-245) is devoted to the sensitivity experiment related to the dry deposition velocity of Hg(II). It’s not clear to me why this uncertainty is the focus of the discussion (compared to other uncertainties, e.g., in oxidation, foliar uptake, etc.) and where these two different settings come from. Are there references to support using 1 cm s-1 vs. 2 cm s-1 for the velocity and 1 or 2 layers for deposition?
Discussion focuses on obvious results:
We already know that with faster oxidation, faster foliar uptake, or faster deposition velocities more mercury will deposit locally. The authors should distill the novel results that we can learn from this study.
Specific comments
Faster than a CTM with full chemistry (L62) - this statement should be supported quantitatively — what is the difference in computational time between WRF-Chem and a traditional CTM?
L71 - Hg tracers with multiple lifetimes (180, 270, and 360 days) - what do these multiple lifetimes refer to? Are these for sensitivity studies? From different sources? Do these represent Hg(0) or different Hg species?
L87 - “presence of ASGM activity” - what was the data source used to define that ASGM activity occurs in the yellow boxes of Figure 1? If it was EDGAR, was some threshold (in emission magnitude) used to define which areas should be studied?
L107–L130 - this type of information belongs in the introduction, not a methods section
What is the resolution of the LAI data used by WRF?
L142- what is the justification for using 1.4 ng m-3 for Hg(0)? Previous literature has reported lower concentrations for tropics/subtropics, especially in forested areas.
L170 - I don’t understand “Largest modelling domain” - please explain if you mean the yellow or red boxes
L173, Figure 3-5 - the region names (e.g., Kalimantan) should be defined somewhere relative to the yellow boxes on Figure 1.
L177 - which table?
Technical Comments
-L7 - Abbreviations in the abstract (KPP, Hg0g, HgIIg, )
-L13 - url in abstract is not necessary, should add as a listed reference
-L24 - define ASGM acronym
-L141- should read 1.78e-4 ng m-2 s-1
Figure 2- this is not South America
L258 - this sentence is not finished
L268 - 1000’s
Citation: https://doi.org/10.5194/egusphere-2024-861-RC1 -
AC1: 'Reply on RC1', Ian M. Hedgecock, 08 Jul 2024
Reply to Referee 1
We would like to thank the referee for their thorough and helpful review.
Concerning the Main Issues raised by the Referee.
We added a new paragraph to the Introduction, with numerous references, which we think makes clearer the reasoning behind the development of this modelling tool, and which provides a broader introduction to artisanal gold mining, its economic, social and environmental relevance. We hope that our intention to create an aid/tool to provide a guide for interested and concerned parties to assess the local implications of mercury use on the local environment is now made immediately obvious.
Concerning the role of vegetation, the last sentence of Section 1.1 states that much ASGM takes place in densely vegetated regions, and section 1.2 immediately follows to discuss the role vegetation plays in the Hg cycle.
We have rewritten the section describing the model to make it clearer that the mercury chemistry/dry and wet deposition module was written and added to WRF-Chem by the authors. We have moved the comment referring to rainfall in the Tropics.
As we said above a new paragraph has been added to the Introduction to address the referrers concerns about a fuller description of the ASGM issue. We have also moved the comment on rainfall characteristics in the Tropics because it was out of place as the Referee pointed out.
In the Methods section, perhaps we were not entirely clear. We use the global average LAI, with the global estimated vegetation uptake, to obtain a LAI dependent uptake flux. That LAI dependent flux is then used with the grid cell and time dependent LAI from the WRF-Chem input files to calculate the Hg uptake flux at each timestep. Thus the model takes into account the different types of vegetation present in the modelling domain and their ‘leafiness’ over the seasons.
The approach does not therefore use an average in the uptake flux calculation and nor is the approach primitive, it is rather, parametrised.
In the Results section we have simplified the tables 2-5, and have moved the rather more complex versions present in our original draft to the Supplementary Information. We have also moved and reduced the section discussing the results with different deposition velocities for Hg(II). This way the section does not interrupt the flow of the discussion from the general within domain deposition results at the beginning, and the section which looks more closely at the Deposition vs. Distance relationships for the various deposition pathways.
We have also added some more detail in section 2.4 of the model description concerning dry deposition, with further references, to explain the choice of values and levels.
We agree that it is not surprising that faster oxidation and higher fluxes to vegetation leads to more local deposition. What is interesting is that there are really quite large differences between ASGM sites, and therefore that it is not possible to generalise on the fate of emissions from ASGM. There are two important points which we think are worth making concerning ASGM, one from a local point of view, that each site needs to be treated separately and that it is quite possible that in some cases there is quite substantial load contamination, while in others there may not be. The second is that as well as being a local problem ASGM is a global problem which ideally requires a coordinated international effort to confront.
The model is however designed not so much for research as such but as an aid should work more than adequately to give an idea of the quantities of Hg impacting regions where ASGM occurs, so that with this information organisations and authorities can target their efforts in the most at risk areas.
Specific comments
Faster than a CTM with full chemistry (L62) - this statement should be supported quantitatively — what is the difference in computational time between WRF-Chem and a traditional CTM?
The KPP version of WRF-Chem we used in the past with full Hg chemistry (added to the RADM2 mechanism) is 10 to 15 times slower than this tracer version. We have added a comment to this effect.
L71 - Hg tracers with multiple lifetimes (180, 270, and 360 days) - what do these multiple lifetimes refer to? Are these for sensitivity studies? From different sources? Do these represent Hg(0) or different Hg species?
This is to cover the current uncertainty in the atmospheric oxidation processes which turn Hg(0) to Hg(II), and yes in a sense they are sensitivity runs to provide limits on our ballpark estimation of local, regional and deposition from ASGM and its contribution to the global atmospheric Hg pool. We have reordered section 2.1 so that the comments on atmospheric Hg oxidation follow immediately after the statement concerning the lifetimes, making it clear where these values come from
L87 - “presence of ASGM activity” - what was the data source used to define that ASGM activity occurs in the yellow boxes of Figure 1? If it was EDGAR, was some threshold (in emission magnitude) used to define which areas should be studied?
The data source was EDGAR and the sites were chosen looking at the global map of ASGM to find examples with significant emissions but with different geographical, ecosystem and meteorological characteristics.
L107–L130 - this type of information belongs in the introduction, not a methods section
The referee is quite right, we have moved and rearranged the introductory section which refers to vegetation uptake (1.2)
What is the resolution of the LAI data used by WRF?
30 arc seconds, there is a choice in the WRF Preprocessor, we opted for the highest resolution. A note to this effect has been added to the text.
L142- what is the justification for using 1.4 ng m-3 for Hg(0)? Previous literature has reported lower concentrations for tropics/subtropics, especially in forested areas.
It’s a global value, it is being used to calculate the LAI dependent ‘deposition velocity’ due to foliar uptake. As the studies by Jiskra et al., Zhou et al., etc look at global uptake, we used an estimate of the global ground level concentration for the calculation.
L170 - I don’t understand “Largest modelling domain” - please explain if you mean the yellow or red boxes
The whole map is the largest modelling domain, the area with the yellow border is the first nested domain (at 1/3 of the resolution of the largest domain), and the red ones the smallest domains (1/9 the resolution the largest domain). We have added a note in the text to explain that the mapped area corresponds to the largest modelling domain.
L173, Figure 3-5 - the region names (e.g., Kalimantan) should be defined somewhere relative to the yellow boxes on Figure 1.
We have added the domain names to the figure captions.
L177 - which table?
We have added table numbers every time the word ‘table’ occurs to avoid ambiguity.
Technical Comments
-L7 - Abbreviations in the abstract (KPP, Hg0g, HgIIg, )
We have replaced these.
-L13 - url in abstract is not necessary, should add as a listed reference
We have removed the URL and added the reference.
-L24 - define ASGM acronym
We have defined ASGM in the first sentence of the Introduction in the revised manuscript
-L141- should read 1.78e-4 ng m-2 s-1
Amended
Figure 2- this is not South America
Our apologies the file was misnamed in the figures folder, it has been corrected
L258 - this sentence is not finished
We have corrected this
L268 - 1000’s
We have used the word ‘thousands’ instead, to be unambiguous.
Citation: https://doi.org/10.5194/egusphere-2024-861-AC1
-
AC1: 'Reply on RC1', Ian M. Hedgecock, 08 Jul 2024
-
RC2: 'Comment on egusphere-2024-861', Anonymous Referee #2, 01 May 2024
The authors developed a highly simplified approach to obtain ballpark estimates of the local and regional Hg deposition fluxes resulting from ASGM activities. It can be a useful tool for a policymaker, with minimal modeling experience, to have a first look at the effect of a certain ASGM facility. In this context, scientific rigor was conceivably not a priority, and thus the model was understandably not evaluated against observations. The results were more qualitative than quantitative. The highly parameterized relationships for nearly all processes inevitably would most likely induce great uncertainties in estimates of the spatiotemporal extent of the impact of ASGM emission sources, and thus it would be nearly impossible to guide mitigation strategies that need quantitative recommendations. In other words, the results from this approach would be good for a first look at the impact of ASGM activities, but to develop policies and regulations, sound quantitative measures with less uncertainties would be needed. Therefore, the modeling approach seems better targeted for policymakers than the scientific community. The authors might want to consider working “policy” into the title of the paper.
Some specific comments are as follows:
- I didn’t see the point of separating convective and non-convective deposition flux, since the authors aimed to estimate foliage uptake, dry deposition, and wet deposition flux. What extra information can such a separation add?
- Section 2.1.2: What initial and boundary conditions were used for their WRF-Chem simulations over the selected regions? How did they decide on the 1-month spin-up?
- L137-139: Wouldn’t the maximums of LAI and uptake rate give the upper end and the minimums of both give the lower end of the range of foliage uptake? The way the authors did there seemed to narrow the range.
- L140: “Three rates” of foliage uptake?
- Section 2.3, L149: “Scavenging ratio” was brought up but didn’t seem to be used later in the text. In reading the section, to calculate the scavenging ratio, for the HgII air concentration, the authors seemed to use the HgII concentration averaged from the surface to 4km for non-convective and to 12 km for convective precipitation. Did WRF define 4km to be the cloud top for non-convective clouds and 12km to be the cloud top for convective clouds? L151: the units for the ratio should be m3air/m3rain if the ratio was defined as rain concentration to air concentration.
- Section 2.4 is confusing. Dry deposition velocity should be applied to the surface layer. For layers above, what does the application of dry deposition velocities mean? Was the dry depositional loss in the layers above the surface layer included in the total dry deposition flux? It doesn’t make sense. How did they come up with 1 cm/s and 2 cm/s values for dry deposition velocity?
- Tables 2-4 were tedious to look at. I doubt much information would be retained by a reader. I suggest visualizing them as much as possible. Also, nothing in those tables was highlighted in red or blue as referred to in their discussion (L175-179).
- L256-257: Incomplete sentence.
- Figure 2 caption: It says emissions and relief from the South American domain, but the names on the top of each panel are Southeast Asian regions.
Citation: https://doi.org/10.5194/egusphere-2024-861-RC2 -
AC2: 'Reply on RC2', Ian M. Hedgecock, 08 Jul 2024
Reply to Referee 2
We would like to thank the referee for their thorough and helpful review.
To reply to the Referee’s general comments:
The Referee is quite right, that scientific rigour was not our primary goal in this study. Certainly without a full set of emissions a comparison with measurements is not feasible. Equally true is that in essence the study is qualitative, which is why we chose to express deposition in terms of percentage of emissions. However, with local knowledge of the Hg trade estimates of the amount of Hg entering local ecosystems from ASGM could be estimated. Potentially permitting to intervene where the situation is most serious.
I didn’t see the point of separating convective and non-convective deposition flux, since the authors aimed to estimate foliage uptake, dry deposition, and wet deposition flux. What extra information can such a separation add?
In part it was out of curiosity, the two are separate in the WRF code, and because their scavenging characteristics (in terms of height) are different it is necessary to treat them separately in the wet deposition scheme. However as described later we have simplified the tables in the main text, and the two are separate only in the tables in the Supplementary Information.
Section 2.1.2: What initial and boundary conditions were used for their WRF-Chem simulations over the selected regions?
The simulations were initiated with an empty atmosphere. We’ve added a note to this effect in the text.
How did they decide on the 1-month spin-up?
To overcome potential instability in the integrator at very low concentrations, and in some cases the choice of simulation domains had caused numerical instability in the model. One month was probably over cautious as we were starting from an empty atmosphere.
L137-139: Wouldn’t the maximums of LAI and uptake rate give the upper end and the minimums of both give the lower end of the range of foliage uptake? The way the authors did there seemed to narrow the range.
???
L140: “Three rates” of foliage uptake?
This has been corrected to fluxes.
Section 2.3, L149: “Scavenging ratio” was brought up but didn’t seem to be used later in the text. In reading the section, to calculate the scavenging ratio, for the HgII air concentration, the authors seemed to use the HgII concentration averaged from the surface to 4km for non-convective and to 12 km for convective precipitation. Did WRF define 4km to be the cloud top for non-convective clouds and 12km to be the cloud top for convective clouds?
The scavenging ratio was used to calculate the, we have added a comment in section 2.3 to ensure that this is clear. WRF does not prescribe cloud heights, we have used these as upper limits the model loops over the grid cells in the column above the cell where rain falls, and averages the Hg(II) concentrations in the column to calculate the wet deposition flux for the time step via the scavenging ratio.
L151: the units for the ratio should be m3air/m3rain if the ratio was defined as rain concentration to air concentration.
We have corrected this.
Section 2.4 is confusing. Dry deposition velocity should be applied to the surface layer. For layers above, what does the application of dry deposition velocities mean? Was the dry depositional loss in the layers above the surface layer included in the total dry deposition flux? It doesn’t make sense. How did they come up with 1 cm/s and 2 cm/s values for dry deposition velocity?
We have expanded the section on dry deposition (2.4), and added some more references to explain the reason for choosing the values we did. The question of layers was effectively a work around to avoid running the simulations again with a different vertical distribution of layers.
Tables 2-4 were tedious to look at. I doubt much information would be retained by a reader. I suggest visualizing them as much as possible. Also, nothing in those tables was highlighted in red or blue as referred to in their discussion (L175-179).
We have reduced the information in the tables, they are less cumbersome now. We have however moved the original tables to the Supplementary Information. Originally we had used blue and red to highlight particular values in the tables, but this is a problem for the xml version of the article and so these were replaced with daggers and double daggers, unfortunately we overlooked some of the time we referred to these values in the text. This has now been amended.
L256-257: Incomplete sentence.
We have corrected this.
Figure 2 caption: It says emissions and relief from the South American domain, but the names on the top of each panel are Southeast Asian regions.
Thank you, this was caused by a mislabelling of one of the figures in the folder containing the images, and has been corrected.
Citation: https://doi.org/10.5194/egusphere-2024-861-AC2
-
CEC1: 'Comment on egusphere-2024-861', Astrid Kerkweg, 02 May 2024
Dear authors,
in my role as Executive editor of GMD, I would like to bring to your attention our Editorial version 1.2: https://www.geosci-model-dev.net/12/2215/2019/
This highlights some requirements of papers published in GMD, which is also available on the GMD website in the ‘Manuscript Types’ section: http://www.geoscientific-model-development.net/submission/manuscript_types.html
In particular, please note that for your paper, the following requirements have not been met in the Discussions paper:
- "The main paper must give the model name and version number (or other unique identifier) in the title."
- “If the model development relates to a single model then the model name and the version number must be included in the title of the paper. If the main intention of an article is to make a general (i.e. model independent) statement about the usefulness of a new development, but the usefulness is shown with the help of one specific model, the model name and version number must be stated in the title. The title could have a form such as, “Title outlining amazing generic advance: a case study with Model XXX (version Y)”.''
As you used WRF-Chem, please add something like “a case study using WRF-Chem version x.y” to the title of your manuscript.
Additionally, a remark how to access the used version of the WRF-Chem code is required in the Code availability section.
Yours, Astrid Kerkweg
Citation: https://doi.org/10.5194/egusphere-2024-861-CEC1 -
AC3: 'Reply on CEC1', Ian M. Hedgecock, 08 Jul 2024
Dear Professor Kerkweg,
We have added the link to the WRF-Chem v4.3 code, and we have changed the title to, “A Policy Aid based on WRF-Chem v4.3 to estimate the local and regional Fate of Mercury Emissions from Artisanal and Small Scale Gold Mining”, we hope this is acceptable.
Citation: https://doi.org/10.5194/egusphere-2024-861-AC3
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
291 | 63 | 31 | 385 | 47 | 18 | 17 |
- HTML: 291
- PDF: 63
- XML: 31
- Total: 385
- Supplement: 47
- BibTeX: 18
- EndNote: 17
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1