The paper “New directions in atmospheric research offered by research infrastructures combined with open and data-intensive science” provides useful view on the future of the atmospheric research. The paper contains a lot of information and in its current form it rather looks like a comprehensive review than an opinion. It will take a reader 25 pages of reading until the authors go to the point.

I would suggest that the author shorten the paper by at least 30% , which still would allow to make a point about future research approaches. In particular, this applies to the references and description of the infrastructures where the whole paragraphs are just copied from the referred publications.

I find the authors opinion interesting, though it would be useful to have authors opinion on two points which are currently not reflected in section 5. The first one: currently the performance of scientists is judged by a number of publications they produce. Within the envisioned new research environment, the research will be done as a collective one. This diminishes the role of the individual researcher and the current evaluation framework by publication number (as in the envisioned research environment the summaries can be easily generated by AI). The second question is related to the role of innovation (e.g. regarding analytical methods or measurement techniques and instruments) within the highly standardized reserach environment.

It would be useful if the authors reflect on these points in the next revision of the paper.

General note on the use of language: the sentences are too long, and the idea is often get lost before one reads to the end. It would be magnificent if authors use more concise language.

Additional comments:

Line 22: the term “variability” may be a bit better than “periodicity”

Line 31: “which evolve from”; it is not clear from what opportunities evolve as you do not provide initial state of affairs. Maybe it would be better to use another works here, like “emerge”

Line 32: what do you mean with “emerging service ecosystem”

Line 44: “economic processes in the Earth system” – economic processes refer to human society rather than to the Earth system

Line 50: GCOS is a co-sponsored programe of WMO, IOC-UNESCO, UNEP and International Science Council

Line 55-56:  do you consider vertical profile measurement as ground based or as in situ?

Line 62-66: lots of repetitions here

Line 67: actually the threats are posed to humanity rather than to the planet

Line 94-94: I would disagree with the comparison with astronomy infrastructure. Unlike astronomy, environmental infrastructure has an immediate value for multiple applications, including climate services for mitigation and adaptation, health and agriculture applications, hence this poses different requirements for timeliness of data availability.

Line 120-126: it is not clear what point you are trying to make in this paragraph

Line 146-151: not clear what the relevance of this paragraph to this paper

Line 154: air quality is a part of atmospheric composition

Line 170-176: another paragraph of repetitions

Line 187-190: the name of infrastructures should be spelled out here, rather than later in the text (lines 205-210)

Line 189: I was under impression that ACTRIS also recently became ERIC

Line 203: it is not necessary to spell out the title of the book, reference would be sufficient

Line 214: “in providing” (“in” is missing)

Figure one: the name of the phenomena is “sand and dust storms” not “desert storms”

Table one: “aerosol particles” – use either of the words, not both

Comma is missing after O3 in IAGOS section (Table 1)

Line 237: what do you mean with “research infrastructure process”?

Line 240-241: “Before the consolidation phase, data have been locked in silos and were poorly standardised over decades, hampering scientific progress severely” – I disagree with this statement as it diminishes the role of the programmes like the Global Atmosphere Watch

Line 251: “periodicity in the presence of” is better to reformulate as “variability in atmospheric levels/burdens of”

Line 259-296: what are you trying to demonstrate with the presented three examples?

Line 277: “Data which were collected before ICOS measurement protocols were put in place, have been secured physically”. Could you please explain this statement? Most of ICOS atmospheric data are part of GAW and those data were and are achieved in the World Data Center for Greenhouse Gases supported by Japan

Line 286: the referred “latest assessment” is 5 years old. Are there updates on this publication?

Line 299-300: could you please clarify the statement “Trends are detectable in the data, possible only due to collection of data with consistent quality and operating procedures over long time periods.”

Line 319: ICOS is a regional network

Line 322: “which investigates the processes” – who investigate the processes?

Line 323: what do you meant with “controlled atmospheres”?

Line 331: the chemical species are also interacting with each other

Figure 4: I disagree with the proposed adjustment of the figure as it completely changes its sense. The original figure was designed to resent a limited set of key climate indicators and each box includes only one variable, while the added box includes tens of variables disturbing the intended meaning. I would suggest to remove this figure from the paper.

Line 346: “this approach” – which one?

Line 348: could you please provide the source for the definition of “essential variable”. Is it introduced by the authors of the paper?

Line 353: where does this number of 12 application areas come from. There are 10 application areas under atmospheric domain as one can see in the OSCAR database https://space.oscar.wmo.int/applicationareas

Line 379: OSCAR is a database, not a table

Table 2a: is the word “distribution” missing in the “aerosol number and size”? In GAW most of the ozone depleting substances are included under greenhouse gases

Table 2b: GCOS does none use volume mixing ratio for ozone variables. The correct variable is mole fraction

Line 405: please use either upper air or upper atmosphere, not both

Figure 5: typo at the end of the first line (“to;” is not needed)

Line 419: I guess “hundreds of stations” applies to temperature observations as this is not the case for greenhouse gases

Line 430: the term “validation” is not used in the Paris agreement

Line 444: what do you mean with “organised the implementation of the FAIR principles”. The principles can be applied to something, rather implemented

Line 448: “FAIR enabling resources need to be implemented” – resources cab be used, rather than implemented

Line 459-468: is this paragraph needed?

Line 468: “composed of” or use another word instead of “composed”

Line 487: how satellite and modelling data re integrated in this cloud?

Line 496: could you please address the risks of the data misuse and misinterpretation by citizens without appropriate background? Training platforms cannot substitute years of professional education.

Line 497: please spell out SMEs

Line 501: could you please use another word for “uses”

Line 505: please spell out abbreviations here, it reads like a lot of slang

Line 515: “metadata catalogue to access information from the RI of interest”

Line 517: “composability” -please clarify

Line 533: “on the use of license on data”

Line 578-579: “for machines, knowledge recorded in scientific articles is not accessible” – this is incorrect and the text below referring to the AI had been used for this purpose. Later in the text (line 599) the authors refer to the publication of the scientific papers, which according to the initial statement would be a waste of resources

This manuscript offers an opinion on “new directions in atmospheric research offered by research infrastructures (RIs) combined with open and data-intensive science”. The European author group is well qualified to discuss the present and future of research infrastructures based on their involvement and experience. Recognizing the current state of affairs of observations, data and models related to climate resilience and anticipating future needs and demands on the RIs of the world is the basis of a valuable narrative. The manuscript is suitable for publication after the authors consider the following comments.

1) The manuscript would be improved by substantially shortening its length. It is a slow read given the abundance of detail that is primarily background material to the main thesis. In shortening, I suggest a focus on the promised ‘new directions’, ie to make them standout better in this narrative.

2) The figures (except 6 and 7) and Table 2 seem to be unnecessary information for this opinion piece; hence I suggest the authors considering deleting.

3) The most interesting and relevant text was in sections 4 and 5 where the reader finds the most new thinking about this topic.

4) The recommendation that the GCIs be modified to include SLCFs appears rather casually on ln 337. The GCIs are formally put forth by WMO/GCOS so a change would be a major consideration. While a quite reasonable suggestion, this proposed change could alone be the topic of an opinion piece. In the present context, it is not clear what the authors would like the reader to think about this proposal, eg what is the next step in advancing this idea or is it only meant to be a marginal comment?

5) Another casual remark is on ln 430 : ‘One prime example for research-based information for climate policy is the validation of emission reductions required as part

of the COP21 Paris Climate Agreement of 2015.’ The failure to achieve GHG emissions and concentration reductions in the next 2 decades is perhaps the greatest threat to the future health of human society and ecosystems. The role of RIs is essential to have an efficient, effective and verifiable global emissions reductions. I could easily see an opinion piece focused on this essential RI role under the label of New Directions and hence standout more in this opinion piece.

6) Para at ln 249. It is important to note that using long timeseries of observations for trend analysis requires a measurement infrastructure that guarantees intercomparability regarding observation precision and accuracy over the time period of interest. Such intercomparability generally requires constant vigilance and support.

7) Minor point: suggest replacing ‘periodicity’ with ‘seasonality’ since the former is non-standard usage.