the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Aug 2025
A Middle Atmospheric Chemistry-Climate Model MACO: Model Description and Simulation Evaluation
Abstract. Developing chemistry-climate models (CCMs) is crucial for advancing our understanding of middle atmosphere and improving whole-atmosphere weather and climate forecasting. Chemical processes play a critical role in shaping the middle atmospheric thermodynamic and dynamic structures. Therefore, CCMs require explicit representation of atmospheric chemistry and must be fully coupled to physical climate components. To this end, a new middle atmosphere chemistry module based on the atmospheric general circulation model ECHAM6 has been developed, ultimately establishing a fully coupled chemistry-radiation-dynamics CCM named MACO (Middle Atmospheric Chemistry with Ozone). This paper introduces the model framework and systematically evaluates its performance through a historical simulation (1970–2014), validated vs. reanalysis and satellite datasets. MACO demonstrates robust capabilities in simulating the key dynamical and chemical processes in the middle atmosphere. The model reasonably reproduces the climatology distributions of major stratospheric chemical constituents, effectively captures the annual cycle of the Antarctic ozone hole and Arctic ozone depletion, and realistically simulates the historical evolution of stratospheric ozone and HCl. However, significant biases persist in the upper stratosphere and mesosphere. Analysis identifies two primary bias sources. First, dynamical biases, including a weak polar night jet and polar vortex, lead to a pronounced warm bias near the polar stratopause. This is likely linked to the model’s moderate vertical resolution, which impacts the representation of atmospheric wave dynamics. Second, biases in the chemical scheme primarily manifest as an overestimation of CH4 and N2O and an underestimation of H2O and O3 in the upper stratosphere and mesosphere. These chemical biases are partly attributed to the omission of photolysis below 177 nm, particularly the Lyman-α line. Future model development will prioritize enhancing vertical resolution and refining the photolysis scheme to address these identified shortcomings.
- Preprint
(4364 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2025-3636', Anonymous Referee #1, 12 Oct 2025
-
RC2: 'Comment on egusphere-2025-3636', Anonymous Referee #2, 27 Oct 2025
This manuscript introduces the Middle Atmospheric Chemistry with Ozone (MACO) model—a coupled chemical-climate model developed based on the ECHAM6 general circulation model—and presents an evaluation of its performance in simulating the middle atmosphere over the period 1980–2014. However, the MACO model has major flaws in core mechanisms and validation methods, resulting in insufficient credibility and scientific value of the model results. Furthermore, the manuscript suffers from a severe lack of innovation. The paper does not meet the basic standards for publication in Geoscientific Model Development, and rejection is recommended.
Major comments:
- The paper does not clarify the differences and innovative points between the MACO model and existing mature chemical-climate models (e.g., SOCOL4, UKCA, CESM2-WACCM): it neither compares the advantages and disadvantages of MACO with similar models in terms of simulation biases for core variables (e.g., ozone, temperature) nor explains whether its chemical mechanisms or parameterization schemes include original improvements. This makes it impossible to demonstrate the model’s scientific contributions. Such unsupported conclusions and ambiguous innovative positioning do not meet the basic requirements of an academic paper and cannot provide valuable references for research in this field.
- The paper explicitly states that the MACO model does not incorporate photolysis processes for wavelengths below 177 nm (including the Lyman-alpha line at 121.567 nm) and acknowledges that this flaw leads to overestimation of methane (CH₄) in the mesosphere (with biases exceeding 50%) and underestimation of water vapor (H₂O) and ozone (O₃) (with biases exceeding 20%). However, the authors have not conducted any quantitative analysis to assess the impact of this flaw on the model’s overall chemical mechanism: neither sensitivity tests were performed to compare simulation differences between "including/excluding Lyman-alpha line photolysis" nor were quantitative data from existing literature on the contribution of this wavelength band to mesospheric compositions (e.g., CH₄ oxidation, H₂O production) cited. This results in a fundamental gap in the core driving link of the model’s chemical processes. More critically, biases in mesospheric photolysis processes will have cascading effects on radiative transfer (e.g., the cooling effect of H₂O as a radiatively active gas) and dynamical processes (e.g., temperature biases at the stratopause). Yet the paper simply attributes stratopause cold biases to H₂O underestimation without establishing a logical chain of "photolysis flaws → composition biases → radiative-dynamical feedback." Such fundamental model defects cannot be remedied through subsequent parameter adjustments, directly rendering the model’s simulation results of the "chemical-radiation-dynamical" coupling processes in the middle atmosphere scientifically unreliable.
- The paper notes the widespread underestimation of hydrogen chloride (HCl), nitric acid (HNO₃), and H₂O in polar regions and speculates that this is related to overestimated simulations of polar stratospheric clouds (PSCs). However, no evidence is provided to support this speculation: first, the paper fails to compare PSC parameters simulated by the model (e.g., occurrence frequency, optical depth, particle type proportion) with satellite observation data (e.g., CALIPSO, MLS), making it impossible to verify the accuracy of PSC simulations; second, it does not analyze the coupling relationship between the weakened polar vortex and PSC simulation biases—polar vortex intensity directly affects the polar temperature field, which is a core condition for PSC formation, yet the authors avoid addressing the critical question of whether dynamical biases exacerbate chemical biases. Furthermore, Antarctic ozone holes and Arctic ozone depletion are core evaluation indicators for middle atmosphere models. However, the paper’s validation of polar ozone simulations is limited to zonal mean comparisons, with no analysis of biases in the vertical distribution and temporal evolution of ozone concentrations inside the polar vortex. It also fails to explain the significant discrepancy between the simulated timing of Arctic ozone depletion (2005) and observations (2011) after 2000. Such systematic biases indicate major flaws in the model’s depiction of "heterogeneous chemistry-dynamical processes" in polar regions, and the authors have not proposed any feasible improvement plans.
- The paper contains critical omissions in validating the model results. The evaluation of all trace gases (e.g., CH₄, N₂O, O₃) relies solely on zonal mean data comparisons, with no regional-scale bias analysis (e.g., in the intertropical convergence zone or polar vortex edges). These regions are key areas for "chemical-dynamical" interactions, and zonal means cannot reflect the model’s performance in critical regions; second, no quantitative statistical indicators (e.g., root mean square error, time series correlation coefficient, percentage trend bias) are calculated. The assessment of model performance relies solely on qualitative descriptions such as "biases less than 5%" or "good agreement," lacking objective data support; third, regarding the uncertainty of MSR-2 total ozone column data in polar regions, the authors did not conduct cross-validation using other independent datasets (e.g., SBUV/SBUV2 series, OMI satellite data), making it impossible to confirm the credibility of polar ozone simulation results. Most seriously, when evaluating tracers such as CH₄ and N₂O, the paper fails to compare the vertical gradients of the substances simulated by the model with observations. Vertical gradients are core indicators reflecting atmospheric transport processes, and their biases directly indicate the rationality of the model’s dynamical transport parameterization. The authors completely avoid this critical validation step, making it impossible to judge whether the coupling between the model’s dynamical framework and chemical module is reasonable.
Minor comments:
- Lines 357-358: A 10-year spin-up period may be insufficient for middle atmospheric chemistry models, especially those involving long-lived species (e.g., N₂O, CFCs), as it might not allow stratospheric chemistry to reach an equilibrium state.
- For Section 2, it is recommended to add a flowchart to briefly illustrate the main model increments added in this study and their logical relationships.
- Section 3.1 focuses on experimental design rather than results and should be moved to Section 2.
- Lines 236-239: The content belongs to result discussion and does not belong in Section 2.
Citation: https://doi.org/10.5194/egusphere-2025-3636-RC2
Data sets
Middle Atmospheric Chemistry-Climate Model MACO Code and Simulation Data Zhipeng Zhu https://doi.org/10.5281/zenodo.16452132
Model code and software
Middle Atmospheric Chemistry-Climate Model MACO Code and Simulation Data Zhipeng Zhu https://doi.org/10.5281/zenodo.16452132
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 1,803 | 43 | 17 | 1,863 | 17 | 22 |
- HTML: 1,803
- PDF: 43
- XML: 17
- Total: 1,863
- BibTeX: 17
- EndNote: 22
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The manuscript presents the description and evaluation of the chemistry-climate model MACO. The subject is relevant to the journal's scope. The authors describe the main components of the model and perform a model run covering the 1970-2014 period. The model results were compared against several datasets, including MERRA2 reanalysis and satellite data (MSR2, GOZCARDS, and ACE-FTS), to evaluate the model's performance. The authors have done a tremendous job trying to identify strong and weak model features. If this manuscript were submitted in 2015, I would not have any doubts recommending the publication. But in 2025, the MACO does not look really new. The used ECHAM6.1 is not the latest version and is not supported by MPI. Most of the discovered problems (e.g., absence of Lyman alpha line, overestimated oxygen photolysis, simplified deposition schemes) are rather known and can be easily implemented in the model. Without these corrections, the simulated fields of methane, water vapor, and ozone are not satisfactorily accurate. The tropospheric chemistry, which contributes to the ozone trends, is missing. The authors know the problems, but are not willing to present a more reasonable version. Even the integration period does not cover years after 2014. Therefore, I can not recommend publication of the manuscript.
Minor problems:
Around line 65: I do not agree that “CCMI emerged from the merger of CCMVal and the Atmospheric Chemistry and Climate Model Intercomparison Project (ACCMIP; Lamarque et al., 2013)”. I think that these programs existed independently.
Around line 85: The look-up table method can be characterized as online, because it uses real-time atmospheric and orbital data.
Around line 115: I do not see that Fast-J has been widely implemented. It seems that only up to 4 models use it.
Around line 170: ECHAM6 in the standard set-up cannot reproduce the stratospheric quasibiennial oscillation (QBO).
Around line 215: Please, add a reference to the Rosenbrock method.
Around line 240: Do all species transported separately, or are they grouped by families?
Around line 270: The deposition treatment is outdated.
Around line 315: Tropospheric chemistry is virtually absent. Lightning NOx emissions can help to get better ozone.
Around line 345: For Fast-J assimetry factor is not enough. The code requires a more complete phase function.
Around line 370: Why only up to 2014? ACE-FTS v3.5 is outdated. There is a newer version
Fig.4: statistical significance is not shown. Methane does not have a double-pick structure. It would be more instructive to show different seasons,
Around line 495: I do not understand why “The situation is similar for N2O, where photolysis at shorter wavelengths is also underestimated”. What is the mechanism? N2O does not have absorption in Lyman alpha.
Fig.7 – No relative differences
Around line 540: How does wet HNO3 scavenging affect the stratosphere?