the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Introducing Volatile Organic Compound Model Intercomparison Project (VOCMIP)
Abstract. Volatile organic compounds (VOCs) play an important role in atmospheric chemistry, influencing the cycling of peroxy and hydroxyl radicals, the formation of tropospheric ozone, hydrogen, secondary organic aerosol, and the lifetime of methane and other greenhouse gases. Their interactions shape overall atmospheric composition and air quality, with implications for both climate and human health. Given their significance, accurate representation of VOCs in global atmospheric chemistry models is crucial. In this context, we introduce the Volatile Organic Compound Model Intercomparison Project (VOCMIP) and invite atmospheric chemistry modelling groups to participate in this collaborative effort. VOCMIP aims to identify model consistencies and discrepancies, enhance the formulation of chemical mechanisms, and advance our understanding of VOC-related processes in the atmosphere. Global atmospheric chemistry model output will be compared to in situ measurements from surface stations and aircraft campaigns, plus satellite data for key VOCs. Special emphasis will be placed on formaldehyde (HCHO), examining its chemical sources and sinks given its central role as a radical source and as an intermediate in the photochemical destruction of VOCs.
- Preprint
(1029 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 05 Nov 2025)
- RC1: 'Comment on egusphere-2025-3057', Anonymous Referee #1, 23 Sep 2025 reply
-
RC2: 'Comment on egusphere-2025-3057', Anonymous Referee #2, 05 Oct 2025
reply
The manuscript introduces the Volatile Organic Compound Model Intercomparison Project (VOCMIP), proposing simulations for 2015 and 2019 using eleven global atmospheric chemistry models with common anthropogenic emissions (CEDS, GFED) and meteorological forcing from reanalysis data. The authors plan to compare model outputs with satellite retrievals (IASI, CrIS, TROPOMI), surface station measurements (EBAS database), and aircraft campaign data (KORUS-AQ, ATom) for key VOCs including formaldehyde, methanol, acetone, isoprene, and various alkanes. Participating models include CESM2 CAM-Chem, EC-Earth, EMAC, FRSGC/UCI CTM, GISS, GFDL, LMDZ-INCA, NorESM2-LM, OsloCTM3, UCICTM, and UKCA. The paper presents requested model output (3-hourly 3D fields in Table 1, monthly budget terms in Table 2), describes available observational datasets (Table 4, Figure 3), and shows preliminary burden comparisons between two models (Figure 2).
The VOCMIP concept has scientific value, but this manuscript represents preliminary project planning rather than a completed research contribution or even a rigorous protocol document suitable for publication. Therefore, I cannot recommend this manuscript for publication in GMD in its current form. The paper has two fundamental problems: (1) it is written in a grant proposal form, and it does not contain model development and therefore does not fit GMD's scope, and (2) it lacks clearly articulated scientific objectives and a coherent experimental design to achieve them.
Specifically, journal scope mismatch: GMD publishes papers describing model development - new models (with model version numbers), significant model updates, or novel parameterizations and methodologies. This manuscript describes a protocol for running existing atmospheric chemistry models with specified inputs and comparing their outputs to observations. There is no new model code, chemical mechanisms, parameterizations, diagnostic tools, or software development. I recommend the authors consult with the editor about appropriate journal placement.
More importantly, the absence of scientific questions, which I think is the greatest problem with this paper: the introduction (lines 42-107) provides valuable background on VOC atmospheric chemistry, but this does not translate into specific, testable hypotheses. The stated aim to "identify model consistencies and discrepancies, enhance the formulation of chemical mechanisms, and advance our understanding of VOC-related processes" (lines 35-37) is too vague. What specific aspects of VOC chemistry are uncertain? What processes drive model differences in formaldehyde or methanol budgets? Can we constrain emissions using model-satellite disagreement? Without 2-4 concrete scientific questions, the project risks producing model comparisons without scientific insight.
Also, the inadequate experimental design: the proposed experimental setup (Section 2) will produce model differences that cannot be attributed to specific causes. Lines 132-133 specify common anthropogenic emissions (CEDS, GFED) but allow each model to use different natural emissions "as defined for each participating model." This ensures models will simultaneously vary in: natural emissions inventories, VOC speciation schemes, chemical mechanisms, deposition parameterizations, and photolysis calculations. When Figure 2 shows models differing by factors of 2-3 for acetone, ethane, and methanol, what causes these differences? Emissions? Chemistry? Deposition? Transport? Without sensitivity experiments that isolate individual processes, these questions cannot be answered.
Lastly, there are also structural issues: the manuscript describes plans rather than results. It has no Results or Discussion sections. This reads as a project proposal or protocol document, not a completed research paper suitable for publication. The summary is only 3 sentences that did not actually summarize the work, and it ended on a hanging note.
Specific Comments:
1. Lack of process isolation in experimental design
The core flaw is that all potential sources of model uncertainty vary simultaneously. For meaningful intercomparison, you need tiered experiments. Why not test sensitivity to key processes individually? You could design experiments that:
- Prescribe identical OH, O₃, NO₃ fields to isolate VOC oxidation rates from oxidant chemistry
- Use identical deposition velocities to isolate chemical vs. depositional sinks
- Vary only emissions inventories while fixing chemistry to isolate emission uncertainty
- Compare different chemical schemes within the same model structure (EC-Earth and EMAC offer this capability per your Table 3)
Without such experiments, Figure 2's model differences become a catalog of disagreement without physical understanding.
2. Methane treatment is contradictory (lines 114-116)
The text states methane is "the most abundant VOC in the atmosphere and the dominant source of key VOCs like HCHO and methanol" then immediately excludes it from detailed analysis. This creates logical problems:
- Figure 1 shows CH₄ as a formaldehyde source, yet CH₄ chemistry won't be diagnosed
- Table 1 requests CH₄ mixing ratios, contradicting "not a major focus"
- Prescribing surface CH₄ doesn't eliminate inter-model differences in vertical profiles, oxidation rates, or HCHO yields
- Methane's relatively simple chemistry (compared to isoprene or aromatics) makes it an ideal test case for understanding model differences
Either include methane with full budget diagnostics (Table 2), or explain scientifically why CH₄ chemistry differences are irrelevant to your objectives. The current justification is inadequate.
3. Emission specifications are insufficient (lines 132-133)
"Supplemented with natural emissions (e.g., MEGAN) as defined for each participating model" provides no standardization. This means models may use:
- Different MEGAN versions (2.0, 2.1, 3.0 have substantial differences)
- Different emission factors and environmental response algorithms
- Different land cover datasets
- Different isoprene emission estimates (you note "significant uncertainties" at line 62-64)
How will you ensure comparable VOC emission totals across models? Lines 89-95 discuss conserving mass, carbon, moles, or reactivity when lumping species - which approach should models use? This must be specified or explicitly diagnosed. For a protocol paper, emission processing must be precisely defined or you need sensitivity experiments varying emissions systematically.
4. Diagnostic definitions need clarification (Table 2)
Table 2 requests budget terms but definitions are ambiguous: "Chemical destruction" - Does this include photolysis (listed separately) or not? Should this be total loss rate or separated by oxidant (OH vs. O₃ vs. NO₃)? For formaldehyde specifically, understanding whether OH oxidation or photolysis dominates requires separate diagnostics. "Chemical production due to photolysis" - Does this mean HCHO produced when other species photolyze? Or HCHO loss via its own photolysis? The naming is unclear.
Missing speciation: For species like formaldehyde with multiple production pathways (methane oxidation, isoprene oxidation, direct emission, etc.), requesting only "total production" limits process understanding. Consider requesting production separated by precursor VOC. Every diagnostic must have unambiguous definition for consistent model output.
5.Strategy for comparing models with different chemical complexity (line 192, Figure 2)
Line 192 and Figure 2 reference "BIGALK" and other lumped species without explaining the comparison strategy. How will you compare:
- Models using BIGALK (lumped C4+ alkanes) vs. models with explicit butane, pentane, hexane, etc.?
- Different lumping approaches (by reactivity vs. by carbon number vs. by structure)?
- Oxidation products from lumped vs. explicit schemes?
This is fundamental to your stated goal of understanding "impacts of these approaches" (line 97). You need a clear methodology for cross-walking between chemical mechanisms of different complexity.
6. Observational comparison methodology is underdeveloped
Section 3 lists datasets but lacks comparison protocols:
e.g. satellite data (Table 4):
- Will you apply averaging kernels to account for retrieval vertical sensitivity?
- How will you handle sampling biases (clear-sky only, specific overpass times)?
- What retrieval uncertainties will be used for model evaluation?
Surface and aircraft data (lines 191-204):
- Which EBAS stations will be selected and why?
- How will you spatiotemporally match aircraft measurements (~80 m × 2 km, instantaneous) to model grid boxes (~100-200 km, 3-hourly)?
- Lines 200-204 note observational gaps for several species in Table 1 - how will models be evaluated without observations?
A protocol paper needs explicit procedures for model-observation comparison, not just a dataset catalog.
7. Missing analysis framework
The manuscript describes data collection (Tables 1-2) without explaining analysis methodology:
- How will you quantify contributions of emissions, chemistry, deposition, and transport to model spread?
- What statistical metrics will assess model skill?
- How will observational uncertainty inform model evaluation?
- What constitutes model success or failure?
Lines 137-139 request extensive budget terms but don't explain their use. Will you perform process attribution? Sensitivity analysis? Budget closure checks? The paper needs an analysis strategy.
Technical Corrections
Line 42: Define VOCs explicitly (vapor pressure threshold or other criteria) rather than assuming reader knowledge.
Lines 45-47: Move the compound type descriptions (hydrocarbons, oxygenated species, aromatics, etc.) immediately after defining VOCs for better organization.
Line 77: Add reference for "oxidation products...nucleate new particles"
Line 82: Replace "Despite" with "Given" or "Because of" for clarity
Line 98: Remove VOCMIP reference before the project is introduced (line 110)
Line 114: "aboundant" → "abundant"
Line 195: Specify which projects contribute data instead of "within various projects"
Table 1 footnote: Move grid specification requirements to table body with exact details (cell area, layer thickness, pressure levels)
Figure 2 caption (lines 180-185): Move simulation details to methods text; caption should describe what the figure shows. Also the colors are not visually friendly, the fonts are squished.
Table 2: Clarify which species require photolysis budget terms
Table 3: Add DOIs or URLs for model references where available
Citation: https://doi.org/10.5194/egusphere-2025-3057-RC2
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
1,295 | 45 | 13 | 1,353 | 13 | 13 |
- HTML: 1,295
- PDF: 45
- XML: 13
- Total: 1,353
- BibTeX: 13
- EndNote: 13
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1
The manuscript egusphere-2025-3057 introduces a new model intercomparison project, named VOCMIP. After providing a comprehensive introduction and clear motivation, the authors describe the experimental design and the observational data used in the study. While I believe this is an excellent initiative and fully support the planned work, I must admit that I have mixed feelings about the manuscript itself.
I began reading the manuscript with great enthusiasm and found the introduction to be a well-crafted and engaging overview. Unfortunately, the experimental design does not live up to the same standard, and, more importantly, the manuscript lacks a clear scientific focus for the project. While some aspects of this focus are touched upon in the introduction, they are never fully articulated or clearly defined.
Hence, although the concept is promising, the manuscript lacks certain essential details and a clear scientific plan that are necessary for a more robust understanding and evaluation of the project. I therefore recommend that the authors revisit the scientific goals of the project and provide a more detailed and precise description of how the intercomparison project will address these goals.
General comments:
As I mentioned earlier, I am unclear about the primary focus of this model intercomparison project. At the moment, it appears to center on "checking the differences in concentrations against observations," but the specific scientific goals are not clearly articulated. It would be very helpful to have these goals explicitly outlined.
If the primary aim is to investigate the chemical mechanisms, why not start by using box models to isolate the chemical processes from the noise introduced by varying atmospheric forcings? For certain models listed (e.g., EC-Earth, EMAC), different chemical decomposition schemes are available. Why not compare the results from the same model but using different decomposition schemes? This could simplify the analysis and help in pinpointing the sources of variability.
If the goal is to study the lifetime of VOCs within each chemical mechanism, why not use the same OH field across models? Since the models are capable of forcing methane at the surface (as mentioned in line 115), they should also be able to apply a consistent oxidative power across the atmosphere.
Alternatively, if the primary objective is to investigate the different parameterizations of deposition, why not focus on intercomparisons of the deposition algorithms? If photolysis rates are the main area of interest, then why not perform an intercomparison of those as well?
While it is certainly possible to pursue all of these objectives, each would require separate simulations. A thorough and meaningful comparison would necessitate a well-thought-out experimental design to ensure the results are attributable to specific causes. Otherwise, we risk obtaining a range of results without understanding the underlying reasons for the differences.
I am concerned that this exercise might turn into a "race" to see which model best matches observations. While this is undoubtedly valuable, I believe there is a greater opportunity here: to understand why the best-performing models are successful. This deeper insight could yield more valuable scientific knowledge.
Specific comments:
*) line 42 : I would appreciate a clear definition of VOCs : when is an organic compound volatile?
*) lines 45-47 : I would move the descriptions of different compounds forming VOCs just after the definition (at the beginning of the text).
*) line 77 : Would be great to add a references on this statement (i.e. nucleation of VOCs' oxidation products)
*) line 82 : "Despite" sounds odd to me in this sentence
*) line 94 : If I understood correctly, emissions are possible by conserving mass, carbon, moles or also reactivity, as described in detail few lines before.
*) line 98 : I would remove the reference to the VOCMIP, as this has not yet introduced (coming later in line 110)
*) line 114 : I found quite funny that you describe methane as "the most aboundant VOC in the atmosphere and the dominant source of key VOCs" , but you will neglect it in this work. I also do not really understand the connection with methane being prescribed at the surface. Actually, this would be even an advantage, as most of the dynamical-driven processes are forced and you could simply focus on the chemistry of methane. In addition, while not a major focus, it seems (table 1 and 2) that you will need exactly the same data of all the other components. So there is no real differences from the technical point of view between methane and other VOCs.
*) Line 132 : What is the motivation to select these emissions? What about different speiciation methods in the emissions? How do you assure that the same amount of VOCs (either in reactivity or mass) are actually emitted ?
*) Line 133 : "supplemented with natural emissions" is quite vague: if you have completely different emissions you cannot expect any intercomparison at all (or only partially).
*) Table 2: What is the definition of "Chemical destruction"? The total chemical sink (including photolysis)? Would be a speciation (e.g. OH oxidation or NO3 oxidation) be more informative?
*) Line 192 : Some of these acronyms were not introduced before (and some only in the Fig.2 label). Would be very interesting if you have a stragety to compare models with "BIGALK" with models that have a more (or less) chemical detail (i.e. VOCs speciation).
*) Line 195 : "within carious project" sounds a bit vague.