In "Results of a Long-Term International Comparison of Greenhouse Gas and Isotope Measurements at the Global Atmosphere Watch (GAW) Observatory in Alert, Nunavut, Canada" Douglas Worthy and co-authors present the results of long time series of atmospheric flask samples analysed for greenhouse gases to check the whether different laboratories succeed in obtaining the WMO GAW recommended compatibility goals for greenhouse gases, as recommended in WMO (2020), and specifically whether the observations at Alert observatory analysed by the Canadian ECCC are compatible with 5 other global labs. 

This important paper is clear and well written, but unfortunately way too long. I appreciate the level of detail and accuracy invested to document all the important details that are useful and needed to understand the conclusions, but in its current shape it is hard for the reader to see the tress for the forest.  Also it is good to have all data now available in one place for future analyses. So I recommend to not remove but move many details (e.g. those in section 2.3 and 3.x inlcuding most of the figures) to a supplement and replace those sections with more general and short descriptions of the commonalities.   

I also agree with the first reviewer that the abstract also need to be shortened (by about 50%). 

I would also appreciate an explanation why the documentation and analysis took so long (first results are from 1999 and latest results analysed are from 2017) and depending on the reasons would like to see recommendations on how this could be improved in the future, as timely information on scale differences between labs is crucial to detect and improve compatibility issues. And what would this mean for the global GAW network, should all associated labs cooperate in these regular comparisons and with which frequence and minimal delay time should the data be analysed?  Would a central database like now set up by NOAA be helpful in coordinating the results?

WMO (2020): GAW report #255, 20th WMO/IAEA Meeting on Carbon Dioxide, Other Greenhouse Gases and Related Measurement Techniques (GGMT-2019). Available at: https://library.wmo.int/index.php?lvl=notice_display&id=21758

As I assume a major rewite is needed I only have a few smaller textual comments and suggestions:

L94: please quote the most recent WMO GAW GHG Bulletin (2022) available at https://library.wmo.int/index.php?lvl=notice_display&id=22149

L104: remove being

L107 : world data centers -> the world data center

L115: Start new line with: However, 

L124 In this regard, -> Fore these reasons, 

L245: explain co-sponsored (what support is provided?)

review of manuscript amt-2023-99

Title: Results of a Long-Term International Comparison of Greenhouse Gas and Isotope Measurements at the Global Atmosphere Watch (GAW) Observatory in Alert, Nunavut, Canada by Worthy et al.

General comments:

The manuscript presents a comprehensive analysis of results from a co-located long-term flask sampling comparison in the high Arctic. The flask samples cover a period of 17 years. Seven international institutions, all well renowned in the atmospheric monitoring of greenhouse gases, participated in the comparison. The paper is well structured and written. It is very comprehensive and provides an excellent overview of what can be achieved in terms of precision and accuracy by the leading experts in greenhouse gas observations. The manuscript includes an excessive number of tables. The authors may consider moving Tables 6 to 22 into the supplementary material for the sake of conciseness of the main part.

The paper is definitely within the scope of ‘Atmospheric Measurement Techniques’. I have only a few minor comments and suggestions for improvement, all listed below.

Chapter 2 briefly summarizes the different sampling approaches and the analytical techniques used. I do not suggest to elaborate on many additional details here since relevant references to the existing literature are provided in the bibliography. However, I was wondering if some specific sampling and analytical methods do systematically contribute to the differences described in the results section. Factors could be the pressure in the flasks during filling, drying/no drying/the way of drying (cryo vs. Mg(ClO4)2) of the air prior to sampling, the time between sampling and analysis (which I assume can be rather long considering the remoteness of the site; rough numbers could be given), effects of single-stopcock vs. double-stopcock flasks (i.e. flushing of the flasks), or (changes) in the analytical technique. Relevant specifications should be mentioned accordingly. What are the most striking lessons-learnt after all these years?

For example, the change of the analytical technique from NDIR to CRDS at SIO in 2012 seems to be observable in Fig. 1 b (panel with the purple symbols, less noise in the difference in the more recent data).

Chapter 3 is rather descriptive. There, reference could be made to the different sampling and analysis methods (see my comment just above) if they can explain some of the observed differences.

Due to the characteristics of ALT, almost exclusively clean-air samples were analyzed and compared. Can you make a (qualitative) statement about the to-be-expected agreement under more polluted conditions?

At least for some of the 17 years, continuous in-situ measurements (for some of the species) were performed. How do the continuous data compare with the flask data? Please add at least a short (quantitative) statement, if applicable.

Are the sampling procedures at MLO and CGO identical to the described methods for ALT?

Specific comments:

Line 182ff.: low operating costs are mentioned as a benefit of the flask sampling program. I tend to agree in terms of investment, but the logistics, shipping costs, and manpower requirements for the off-line analysis are significant. Can you somehow quantify the effort over all these years? 8'000 flasks are mentioned in the abstract but I assume that there are also other quite impressive numbers. Numbers of boxes shipped to ALT? Total distance travelled by the flasks? … Any idea about the manpower required for flask preparation, sampling, and analysis? Total costs of the endeavour? No reference is made to any funding agency in the acknowledgements. This makes it even more impressive (that you were able to keep up this comparison exercise over all these years).

Line 281 – 283: "Flask air samples were collected at Alert during persistent southwesterly wind conditions, when wind speeds were greater than 1.5 m s-1 for several hours prior to sample air collection." In other words, there was no sampling when the wind conditions were inappropriate? If so, sampling was retried as soon as the conditioned matched the requirements, or did you skip this sampling and stick to the regular (weekly, bi-weekly) schedule (the following week)?

Lines 344 – 345: "When meteorological conditions are favorable for sampling, the NOAA sampler is taken outside and several meters away from the GAW laboratory to collect the air samples." What happened when conditions were unfavorable? Flasks for the other institutions were filled but no NOAA flasks? How is "favorable" defined?

Lines 367 ff.: Is it correct that SIO flasks are not pressurized at all (only the valve of the evacuated is opened for filling)?

Line 387: "… pressurized, vented and re-pressurized …" To which pressure?

Line 398 "double-stopcock flasks" vs. line 405 "double-valve stopcock flasks". Please harmonize the wording.

Table 4 reads "GC/ECD" etc. while "GC-ECD" etc. is used in the text.

Calibration scales given in Table 4: most scales are WMO/GAW scales hosted by the GAW Central Calibration Laboratory at NOAA (X2007, X2004A, …); SIO uses X08A, which does not seem to be an official WMO/GAW scale. Is there a reference that documents the WMO CO2 X2007 to X08A consistency?