The study by Pennington et al. exploits ozonesondes to quantify the bias and drift in several satellite records of tropospheric ozone. Overall, this is a nice study providing valuable constraints on TROPRESS products but especially on the methods to QC the ozonesonde data before and after the satellite operators have been applied to the sondes (i.e. allowing for like-for-like comparison between the two). Therefore, this manuscript is suitable for publication in ACP subject to some minor comments.

Minor Comments:

Section 2: CrIS, AIRS and OMI have already been defined in Section 1, so do not need to be redefined in Section 2.

Line 121: Space needed between “15” and “km”.

Line 162: Should “altitude” be “altitudes”?

Line 182: Should “compared to climatology” be “compared to a climatology”?

Line 247: “AIRS’ and CrIS’ error”…should this be error or bias? On line 247 you have used “bias” and “error”. However, they are subtly different things. Please be consistent.

Figure S5: The multiple lines for each colour (red, black or blue) represent lat bins at 10 degree intervals. However, it is not possible to know which profile is for which lat bin. Or is the main point that they are typically close together? If the latter, please make that clear in the Figure caption or discussion.

Line 260: The authors cite Miles et al., 2015 for the RAL GOME-2 data. A more up to date study in TOAR-2 is that from Pope et al., (2023 - https://acp.copernicus.org/articles/23/14933/2023/acp-23-14933-2023.html) using multiple RAL Space products. Thus, could be useful to cite that ref on line 260 as well.

Line 282: The authors say “A previous study” but it is unclear what the study is. Can you provide the reference in the text?

Line 318: “Joint AIRS+OMI does not include high ozone values in the Northern high latitudes due to the low joint AIRS+OMI data volume near the poles”. Can the authors please expand on this statement as unclear to follow. Why would it not include high ozone values because of low sampling? Would you not expect to have normal levels (be this low or high values relative to other latitude bands) but just a noisier signal as the data volumes are lower?

Line 320: I know the authors cite Manney et al., 2017 for “multiple tropopauses”, however, it might be useful to add a sentence explaining who you get multiple tropopauses.

Table 2-4: It is clear that N is the sample size. However, I cannot see this defined in the Table caption. Thus, for clarity, please do so. Also, e.g. JJA is defined as June/July/August in lots (in not all) instances. You probably only need to define the season(s) once and then use the abbreviations.

This paper evaluates the TRopospheric Ozone and its Precursors from Earth System Sounding (TROPESS) CrIS, AIRS, and AIRS+OMI ozone products.  It quantifies the biases and bias trends in the products compared to quality controlled ozonesonde data to assess the suitability of the products for trend calculations.  Multiple methods for quality controlling the ozonesonde data are also considered.  The analysis is rigorous and thorough and provides important validation of the TROPESS products.  I list general and specific comments below.

General Comments

1. While the main focus of the paper is on quantifying biases and trends in bias in the TROPESS products, it would be helpful to provide more discussion of the ozone trends themselves seen in the TROPESS products in order to provide more context for the discussion of biases.
2. Since MLS is used to quality control the sonde data, more discussion of how much uncertainty is present in MLS would be helpful to justify this choice.

Specific Comments

Line 92: Please explain why you focus specifically on SNPP-CrIS

Line 131: Is the Fu et al joint retrieval the same as what is implemented in MUSES?

Line 169: sonde operator or satellite operator?

Line 217: Please clarify if you are actually using the moving block bootstrapping in this study.

Line 239: Does “CrIS errors can be comparable…” mean some individual points in Fig. 5a are comparable, or something else?

Fig. S5: It’s a bit hard to interpret this figure where some of the red or blue lines cross the black lines.  Perhaps light to dark shading or different linestyles could be used to indicate which lines correspond to different latitude bands?