the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 13 Aug 2026
The MESSy chamber DWARF: a box model for simulating chamber experiments (based on MESSy v2.55.2)
Abstract. Numerical simulation of environmental chamber experiments is essential for improving models of atmospheric composition. This task often relies on box models, which are not integral to three-dimensional atmospheric chemical transport models. Here, we present an application of the Modular Earth Submodel System (MESSy) DWARF model for chamber simulation experiments. It was developed as a zero-dimensional chemical box model of atmospheric oxidation. Chamber-specific features—including the injection of trace gases, variable radiation conditions, gas- and aqueous-phase chemistry, wall emissions, and dilution—are represented using existing and adapted MESSy submodels. We describe the multi-phase kinetic framework, the estimation of aerosol liquid water content and a simplified treatment of wall losses. We tested the MESSy chamber DWARF setup with experiments from three environmental chambers: The SAPHIR and SAPHIR-STAR chambers at Forschungszentrum Jülich, and the BATCH chamber at the University of Bayreuth. These chambers cover diverse designs and experimental conditions. The model reproduces observed time series of radicals, nitrogen oxides and ozone during an experiment in SAPHIR in which 2-methyl-3-butene-2-ol was oxidised. In another test the model was used to check the consistency of the experimental boundary conditions in the BATCH chamber. Further chamber DWARF tests also show that the framework can predict organic aerosol mass concentration via kinetic partitioning between the gas phase and deliquescent particles in SAPHIR-STAR. Modeled organic mass concentration during this experiment where α-pinene was oxidised agrees well with observations, supporting the framework’s ability to simulate the chemical evolution in both the gaseous and aqueous phase. The MESSy chamber DWARF provides a useful tool for developing and evaluating kinetic models. Within this framework, we can test new laboratory findings directly for their atmospheric relevance. Thanks to the MESSy framework, we can use the same kinetic model for global simulations.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Geoscientific Model Development.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(2020 KB) - Metadata XML
-
Supplement
(586 KB) - BibTeX
- EndNote
Status: open (until 08 Oct 2026)
- RC1: 'Comment on egusphere-2026-4522', Anonymous Referee #1, 04 Sep 2026 reply
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 50 | 19 | 7 | 76 | 9 | 11 | 7 |
- HTML: 50
- PDF: 19
- XML: 7
- Total: 76
- Supplement: 9
- BibTeX: 11
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Summary
Do et al. (egusphere-2026-4522) present a model description paper for the model called MESSy chamber DWARF. I thank the authors for their hard work. I enjoyed many parts of the paper, including the concept, and it is generally well written and presented.
The novelty of MESSy chamber DWARF over some other box models is its ability to readily integrate into three-dimensional atmospheric chemical transport models. This is an important concept, though as I state below not entirely novel. It must come with limitations, though these are very unexplored in the paper.
In my opinion, the paper requires multiple major revisions. Until these points are addressed, either through disagreement in the author response, or through paper revisions, I am not convinced that the paper meets the requirements for a model description paper and therefore I cannot recommend publication in its current form. I describe my suggested major and minor revisions below.
Major points
For the introduction, because the significance of MESSy chamber DWARF is that it readily allows assessment of ambient significance of box model processes, there needs to be much better justification of why MESSy is a valuable framework to be contributing to. Currently line 51 explains that it is well-established and used in numerous works. But the reader needs much more than this. Please explain the significance, providing evidence of impact, of MESSy to the scientific community. Such an explanation will be very valuable to a reader, particularly one coming from a chamber focus, deciding whether MESSy chamber DWARF is appropriate for their task. I suggest this explanation deserves its own paragraph.
Paragraph starting line 31, I understand that the SSH-aerosol model has been applied to both zero- and three-dimensional problems, as described in https://doi.org/10.5194/gmd-19-389-2026, and can host chemical schemes of varying complexity. The introduction must address why MESSy chamber DWARF is novel when compared to SSH-aerosol.
Section 4.2.2 and Fig 3. I think the authors are trying to verify (using GMD description paper terminology) the model treatment of continuous injection and dilution. However, the chamber observations compared against include other processes such as chemistry and wall effects, neither of which has been described by this point. Therefore, I do not accept Fig. 3 as a verification for the model treatment of continuous injection and dilution. It may be used an evaluation of model performance (using GMD description paper terminology) once all the relevant simulated processes have been described.
The paper mentions updates to MESSy such as at line 236. However, it is not clear whether these updates are included in v2.55.2 (the version number stated in the title). Please explicitly state somewhere, such as in the introduction, the version number of the updated MESSy that the presented model uses. This will reassure the reader that when they use that version they are certainly using the version that relates to the description in the paper.
On a similar note to the previous point, the authors explain that MESSy chamber DWARF is a configuration of MESSy DWARF, which Kerkweg et al. 2025 explains is a simplified base model for MESSy to use. Therefore, isn’t the model being described here a subset of MESSy DWARF? I think giving names to subsets of existing models is confusing and can allow an unbearable (at least for a human) number of model names to be generated. Because of this, why not change the title to something like: MESSy DWARF for box modelling of chamber experiments (based on version number of MESSy used once any updates described in the paper have been accounted for in the version number). Another reason for this is that it seems quite plausible that the model described in this paper could be applied to box modelling ambient or indoor environments in a similar way that AtChem has been, for example. Using “chamber” in the title therefore implies a certain limitation that is misleading.
If a subset of the MESSy DWARF base model is being described, shouldn’t MESSy DWARF have an associated version number too? In the same way that if ECHAM were the base model, it would have a version number alongside whatever version number of MESSy is used.
In section 5.1 gas-particle partitioning or organics is dependent on ALWC, but not on other absorbing masses like organics. This conflicts with absorptive partitioning theory, which appears to be the one they are using but currently only considering water as the absorbing mass. The authors should state on what theory they are working and justify any divergence from that theory. To add to this, the authors need to address the expected limitations for gas-particle partitioning representation with the current model described by Eqs. 4 and 5 when water solubility is the only driver. E.g., for scenarios where low water solubility organics have a substantial condensable fraction because they are absorbing to particle-phase organics rather than particle-phase water.
Lines 421-431. The assumption that the model is right leads the authors to the conclusion that experimentalists at BATCH, at least for the studied experiment, have estimates for fundamental properties out by up to a factor of 2. Whilst the authors are encouraged to explore the possibility of experimentalists being out, it is unacceptable that they do not explore the possibility of the model being wrong, particularly as they are conducting a model evaluation exercise that explores uncertainty ranges for the model and model limitations.
Line 433, the first sentence of section 6.3 deserves inclusion in the earlier section (I think it was in Section 5) around describing treatment of gas-particle partitioning. Perhaps a new paragraph in that section that describes the limits of the aqueous aerosol phase representation, including what happens when the organics are in a non-aqueous phase or when relative humidity is low. Things like, does the user need to assume that organics are in the aqueous phase when using the model?
Line 443, evaluation of the model for secondary organic aerosol evolution must at least achieve mass conservation for organics in the gas-, particle- and wall-phase. However, by constraining gas-phase alpha-pinene concentrations to measurements, mass conservation is compromised since the observed alpha-pinene does not necessarily represent the sources and sinks (sinks including oxidation to generate SOA) that the model equations and integrator would, and is, estimating.
Line 454, the assumption that SVOCs do not partition between the gas and wall requires justification based on previous literature. My understanding of previous literature is that the wall can represent a much greater absorbing mass than particles and therefore SVOC partitioning can be significant. I would therefore be much more convinced if SVOC gas-wall partitioning were on.
On constraining particle number size distributions to measurements: the authors must explicitly explain what processes are therefore accounted for, e.g. coagulation and particle deposition to walls. They must also explain whether the particle-phase mass of components, both organic and inorganic is coupled or uncoupled to the number size distribution. If it is coupled, does this mean that mass conservation is lost since any gas-particle partitioning of organics in the previous integration time step is overridden by constraining to measured number-size distributions? If uncoupled, then what is compromised? Either way I think something needs to be compromised because the number-size distribution is not being modelled in the same way that gas-particle partitioning is being modelled.
There is no discussion around how or why the SOA results in Fig. 7 give some agreement with observations. For example, the reader deserves to know explicitly from this paper whether the HOMs that have been reported as important SOA contributors in previous literature are represented in the gas- and particle-phase chemical schemes used in the simulations.
Figure 7b, I don’t think that presenting measurement-constrained alpha-pinene against observations is an efficient use of space. It would be far more insightful to provide something that says something about why the model gets the agreement and disagreement in SOA shown in subplot a. For example presenting modelled and observed O:C and H:C. More generally, although O/C is shown in Fig 10, it would be very insightful for demonstrating model capability and evaluation to see observed vs. simulated H/C.
Figure 10c, the performance of the simulated O/C and H/C should be discussed in the context of past work, including why similarities/discrepancies arise. For example, Fig 2c of https://doi.org/10.1038/s41467-019-12338-8 and Figure 2 of https://doi.org/10.5194/acp-26-2769-2026.
The conclusion needs to fairly reflect the main text much better, for example the summary of the BATCH experiment needs to acknowledge that there is an unresolved discrepancy for which the cause (model or measurement) has not yet been identified. Additionally, the main text and Figs. do not conclusively show that the model predicts a shift in functional groups as stated at line 545, nor that this shift affected wall loss and therefore SOA. The final sentence of the conclusion presents future work that does not seem to be justified by the main text, for example, there is no evidence presented in favour of better representing partitioning of organic vapors in the condensed phase. Generally, the conclusion does not provide any summary of model limitations. Below I provide a guideline from a recent review of conclusion content. The conclusion should restate the significance of having available a model that can be readily applied to both chambers and to the three-dimensional Earth.
Summary: Summarize the main results and relate them to the objectives, questions, or hypotheses of the study. The summary should include the main quantitative results.
Throughout the paper, and certainly in the introduction, there needs to be a better acknowledgment of the limitations of applying a model suitable for three-dimensional simulation to chambers. For example, can the model help us identify missing or poorly-represented processes in chambers? If not, what is the advantage of this model? I think the authors should acknowledge in the introduction that a workflow for transferring information from chambers to three-dimensional simulation exists, e.g. https://acp.copernicus.org/articles/20/10889/2020/, and that detailed models, some unsuitable for 3D application are an important part of that workflow. Though if the authors disagree they should explain their disagreement in the introduction.
Throughout the paper description of the model equations and how they are solved is either light or omitted. Because this is a description paper linked to a MESSy special issue, it is acceptable to address this by referencing previous publications that contain these details, but there are sections where this is not yet done, e.g. the equations and solver for dilution and injections.
Minor points
Line 25, I think the purpose of the Fuchs et al. 2026 reference is to provide an example of previous literature describing, in general terms, environmental chambers. However, the Fuchs et al. 2026 paper is specifically about atmospheric simulation chambers. It seems that atmospheric simulation chambers are the objective of the described model, and therefore I recommend changing from “environmental chambers” in the main text to “atmospheric simulation chambers”. If the authors choose to keep “environmental chambers”, then the main text should describe why the model is relevant to experiments beyond those of atmospheric relevance, e.g. air-water and air-soil exchange.
Line 28, remove “,”
Line 29, the final sentence of the paragraph is too vague, meaning I don’t understand what “chemical kinetic model” is, is it the box model, the Earth model, or something else?
Line 32 “and some also of aerosol microphysics” is too difficult to read and understand, please edit.
Line 33, I suggest “chemical” be changed to “physico-chemical” mechanisms. I also suggest another aim is to help design chamber experiments.
Line 37, the wording here about “Other models” suggests that the EASY, PyCHAM and F0AM models do not provide an interactive user interface. However, PyCHAM at least does.
Lines 46-59 in the introduction: this paragraph is too long and delivers multiple different, and important points, please separate out into multiple paragraphs.
Line 48, please explain what “base model” means, this will also be useful for understanding its use in later sections.
Lines 58 and 44, these lines may suggest that the described model allows a seamless way to represent box model processes in three-dimensional simulations. If this is the case, I suggest being more explicit, e.g. changing “for the transfer” in line 59 to “for the seamless transfer”
Line 60, although the terminology can be debated, I ask the authors to consider carefully whether “tested” is the most appropriate word for a GMD model description article. For example, in the Manuscript Types page (https://www.geoscientific-model-development.net/about/manuscript_types.html#item1), model description papers should provide “verification” that equations are solved correctly and “evaluation” that the model is a good representation of the real system. I therefore encourage the authors to change “tested” to whichever of “verified” and/or “evaluated” is appropriate.
Line 68, state where BATCH is located to be consistent with previous bullet points
Line 69, think it should be stated that the aim relates to MESSy chamber DWARF, because it currently reads as if the aim is to describe chamber simulations for any kind of simulator.
Title of Section 2: because the whole paper has been submitted as a Model Description paper, it is recommended to change the section title to something less confusing, perhaps “Model outline”.
Line 98, please explain what “number of aerosol phases” means, e.g. does it mean gas/particle?
Line 103: currently reads as if measurement data can only be used to constrain and then compare against, therefore please clarify if measurement data can be used just for comparison whilst not being used for constraint? (since one would expect the model to reproduce variable values that had been used for constraint)
Figure 1: Numerical integration?
Figure 1: In Data Analysis box I recommend “Plotting for comparison” changed to “Plotting” as presumably you can plot variables that are not wanted to or able to be compared against anything such as observations.
Figure 1: I suggest removing the Modification text and arrow, and associated caption text in the figure as it suggests that modification is integral to the workflow. Actually modification is the user’s choice, which may be stated in the corresponding main text. It is a valuable outcome whether or not a given model setup achieves model-measurement agreement so long as the setup is justified and useful.
Line 107: built-in?
Line 114, when considering the constraint setup that allow change between two time steps, please state more clearly what you mean. E.g., if you mean that the variable value can change during the numerical integration over the prescribed time step, please say this is what can happen.
Section 3.4 Currently reads as if the model cannot solve the number size distribution independently, rather it constrains to measurements. If this is the case, it needs stating explicitly as it is a significant feature of the model. Similarly the effect on secondary organic aerosol needs stating, for example, is the SOA concentration then constrained to measured number size distributions?
Section 4, it is not obvious whether processes (e.g. dilution and gas-phase chemistry) are solved simultaneously, or whether they are split and solved in sequence. This must be clarified.
Line 170, the description of what SCALC does is too vague. Is it just scaling operations or numerical integration, and for what physico-chemical processes do the mathematical expressions apply?
Line 179, do you mean the pressure-driven inflow of ambient air rather than diffusion?
Section 4.1, how about dilution of particles?
Figure 2: currently unclear from this figure where chemistry is solved
Section 5 needs to state clearly whether chemistry schemes different to those specified (MOM and JAMOC) can be applied
Section 5 needs to state more explicitly whether the MOM and JAMOC chemistry schemes are the same as those available for use in three-dimensional MESSy simulations
Lines 240 and 250, does the Henry’s law coefficient also account for volatility, as well as solubility? If so, this needs stating, if not then a reason should be provided as why.
Line 263, should LWC in Eq. 4 be ALWC?
Line 339, please state what C_org and C_OA represent, if they are concentration, please state the type, e.g. number or mass and explain whether number or mass is consistent with interpretation of AMS results.
Line 339: Consistently -> Consistent
Section 5.2, I don’t see how the C* is converted into a k_f,wall and/or a k_b,wall, please either explain in a paragraph such as the one starting at line 357, or by referencing the relevant supplementary material section in that paragraph.
Section 5.2, the authors need to acknowledge that previous literature finds wall chemistry to be influential, and therefore they should also discuss what options, if any, users have for setting wall chemistry. E.g. “parameterization of sources for nitrous acid” in https://www.eurochamp.org/simulation-chambers/SAPHIR
Section 6, throughout please carefully consider whether you mean verify or evaluate the model and its constituents. See notes above about “testing” and consistency with the GMD description paper terminology
Line 391, if Eqs. 1 and 2 were used to set photolysis frequencies please state this explicitly, otherwise please explain why they were not used.
Lines 393-395. Please clarify whether gas-phase HONO concentrations were constrained to measurements or whether the model explicitly set its emission from the wall as a boundary condition, or something else. Because currently it is somewhat contradictory to say it was constrained to measurement and its emission from wall was included.
Line 430: identifying -> identify
Line 433, so what assumption is being made here? Is it that organics are present in the aqueous phase? If so, this needs to be justified with reference to past literature.
Caption of Table 3, do you mean before reactions, or before lights on? Since alpha-pinene and ozone will be reacting when present together.
Line 455, volatile -> volatility, also this sentence does not make sense
Figure 7, the caption explains the blue shaded area is model uncertainty, but also that it seems to be derived from observations. If this is really derived from measurement then either it should be named as measurement uncertainty, or a paragraph of main text is needed to explain how model uncertainty is quantified. Either way a description of what the shaded area represents is required for both subplots a and b.
Line 496, I expected this sentence to follow logically from the previous sentence, should it read that wall loss of RO2 was deactivated whilst low volatility closed-shell species wall loss was maintained?
Section 6, the reader needs to be told what ALWC estimation method was used for the results in Figs 7, 8, 9