This paper presents an evaluation of the IFS-COMPO model (applied in CAMS forecasts) version Cy49R1 simulations of stratospheric composition, with stratospheric chemistry represented by the BASCOE module. The evaluation is structured into a global assessment during a quiescent period (July 2023 to May 2024), polar chemistry during a series of Arctic and Antarctic winters, and the volcanically perturbed years after the Pinatubo eruption in 1991. The model performance is discussed in detail for ozone, water vapor, long-lived species (N2O, CH4), sulphate, and a suite of other chemicals.

Overall, I find this a valuable contribution for the community interested in simulating atmospheric composition and particularly stratospheric chemistry, which falls clearly within the scope of the journal. The paper is well written and structured, however somewhat lengthy (see minor comment below). I especially like the scorecard summary, which allows for easy identification of the relevant parts of the results, but would ask the authors to carefully think about the evaluation grades here (see minor comment). My recommendation is publication in GMD after after taking into account the following minor and specific comments.

Minor comments:

1.) Length of paper: Overall, the paper is a bit long. I understand that it is important to present and discuss the model performance for several different species. However, making the text more concise and reducing the number of figures could improve readability, potentially increasing both the paper’s readership and its impact. I don't have special recommendations here but would just encourage the authors to think about such potential improvements.

2.) I'm wondering why the model moist biases in the extratropical lowermost stratosphere (below about 100hPa) are not discussed at all (around L542). These are the largest biases in the profiles shown in Fig. 5, and are similar to known moist biases in climate models (e.g. Charlesworth et al., 2023,  https://doi.org/10.1038/s41467-023-39559-2), and in IFS have recently been shown to contribute to UTLS cold biases (Bland et al., 2024, https://doi.org/10.1002/qj.4873). I'd find it good to discuss these issues briefly here.

3.) Scorecard grading: I really like the summary of results in the scorecard in Sect. 7. But I'd suggest to be somewhat more careful with giving particularly high scores here, given the remaining biases in parts of the profiles. Such high scores could be misleading if quick readers don't look into specific details in the related subsections. A few examples where I'm sceptical about the choice of score are:

Fig. 19, U.S.: CH4, H2O, O3, ... show significant biases above about 10hPa (Figs. 4, 5, 9), so that I'm unsure whether "good performance" is suitable here.

Fig. 19, Tropical M.S./O3: Also in the tropical profile (Fig. 9) the bias increases above 10hPa, such that I wouldn't rate the performance "very good".

Fig. 19, Mid-lat. M.S./H2O: For H2O the mid-latitude correlation in Fig. 5 is very low, so that also here I'm wondering about the "good performance".

Abstract, L30: "very good performance for O3, HC4, N2O and H2O..." perhaps too strong given the remaining biases in parts of the profiles.

Conclusions, L1019: "very good performance for CH4, N2O and H2O" I find too positive.

Related to these comments, it's not obvious to me that ACE-FTS is the better reference dataset for stratospheric water vapor (as chosen in Fig. 19 grading). MLS also provides a very good stratospheric water vapor product, and compared to MLS the IFS biases are generally larger.  

Specific comments:

L138: How is the volcanic injection of sulphate species treated in the model? Would be good to mention here or point to the relevant place in the paper.

L212 (Fig. 1): I don't understand the distinction between SO2 in CB05 and BASCOE. Please clarify in caption or text.

L225: The variable "c" in Eq. 11 needs to be explained.

L403: I agree that the highest mean age values in the stratosphere are below 10 years. However, age spectrum tails extend well beyond. Hence, the statement "oldest air encountered in the stratosphere" is not correct and should be changed.

L507ff: How is the upper boundary condition treated? Can't this also be a source of bias in the upper stratosphere? Please add some explanation and discussion here.

L524: Adding age of air tracers to IFS would indeed be very interesting for future work.

L890: What means "By elimination..." here? 

Technical corrections:

L72: ... CAMS was upgraded ....

L101: Lagrangian

L142: blank between "aerosols  as"

L197: Cy48R1 - there are also other places where the "R" is lower-case (e.g. L201, L375, etc). Please check the entire manuscript again.

L307: one "solar" too much.

L403: "oldest age" sounds awkward, better "oldest air" or "largest/highest age values", etc.

L415: Would change "Let us compare..." to "In the following, we compare ...", or similar.

L431: Would change "It will be interesting to see how..." to "It is a particularly interesting question how ...", or similar.

L664: blank missing "inthe".

L691: number missing: "~150 and ~hPa".

L748: "in the two control runs".

L804: blank missing "afterwardan", and better "afterwards ..."

L884: blank missing "quitesimilar"

L894: Just simplify to: "To conclude, IFS-COMPO ..."

L939: "stratospheric"

L962: blank missing "theagreement"

L983: blank missing "thisunderestimation"

L987: "Northern"

L1044: "spring"

Figures 4, 5, 6, 7, 9: Is the legend labelling of red solid and dashed lines correct? I guess the solid line should be Cy48... (not Cy49...), as for the blue lines? 

This manuscript presents a thorough evaluation of the stratospheric chemistry aspect of the IFS-COMPO Cy49R1 system. These updates will help the Copernicus Atmosphere Monitoring Service (CAMS) move towards a better stratosphere-troposphere chemistry scheme. The authors detail recent improvements, specifically focusing on advances in the heterogeneous chemistry part of the stratospheric chemistry. They attempt to evaluate the updated scheme using the Mt. Pinatubo-related perturbation as a test case.

These enhancements and detailed documentation will be highly beneficial to the large scientific community that utilizes CAMS data and will serve as a crucial guide for future model development. The manuscript's primary strength lies in the documentation of the updated scheme with some evaluation, rather than new scientific discoveries. This perfectly aligns with the scope and aims of Geoscientific Model Development. The authors have effectively detailed the upgrades and assessed their impact on stratospheric composition, making this a very suitable contribution to the journal. Overall, I recommend the manuscript for publication with minor modifications.

Comments:

1. Decluttering Figures:

   Many figures in the manuscript are currently difficult to read due to an excessive amount of text. To improve clarity and readability, I strongly recommend that the authors move explanatory text, detailed descriptions, and other information from the figures into the figure captions or the main body of the text. This will allow for larger tick labels and legends, making the figures much easier to interpret at a glance.

2. Climatology:

   I failed to understand the mean biases of modeled trace gases in a climatological context. Please consider adding an additional figure that compares key long-lived tracers/reservoir species (e.g., O3, HNO3, H2O, N2O, HCl, CFC-11, CFC-12) against well-established, publicly available observation-based datasets. There are many selective line plots but it is very difficult to understand basic strength and weakness in the chemistry scheme.

   I suggest showing 5 to 10 mean latitude/height or pressure cross-sections (e.g., see Chapter 6 in the CCMVal 2010 report; tracer-tracer correlation plots in that report also provide a better guideline to understand issues with transport/chemistry). This approach would allow for a clearer assessment of mean biases across different regions of the stratosphere. For example, a 2x5 type figure could be used, with the first column displaying modeled climatological values from the lower stratosphere to the upper stratosphere, and the second column showing absolute or percent differences with respect to SPARC Data Initiative products (e.g., Hegglin et al., 2021). Alternatively, the authors could utilize TCOM (TOMCAT CTM and Occultation Measurement based data), as described in Dhomse et al., 2023 (data for all these species is available on Zenodo). While the current manuscript includes comparisons with daily satellite retrievals (ACE-FTS, Aura-MLS) and reanalyses (BRAM3), an additional comparison against observation-based climatologies would help readers understand mean biases in the updated chemistry scheme.