The manuscript compares net ecosystem CO2 exchange from a paludified spruce forest and an adjacent ombrotrophic bog in west Russia and analysed the main environmental controls on NEE and its component fluxes. The study addresses an important research question aiming at better understanding interannual variability in NEE in these understudied ecosystems. The manuscript is mainly well written but could benefit from writing improvements (e.g., grammar and wording). The methodology is sound and in general appropriate for this study. Overall, the study remains very descriptive, and, in my opinion, results should be strengthened by adding uncertainty estimates for fluxes and statistical test results when comparing bog and forest fluxes throughout the manuscript. Additionally, water availability and drought effects are discussed but observational evidence does not appear to support that these factors play a significant role. I think this could be further clarified.

I also have some more specific comments:

Line 47: The following reference could be relevant here too: Helbig et al., 2019; https://doi.org/10.1029/2019JG005090

Line 54: The following reference could be cited here too:  Moore et al., 2006; https://doi.org/10.1111/j.1365-2486.2006.01247.x

Line 61-64:  Helbig et al. (2019) might be relevant here too

Line 72: It might be insightful to include results from the SPRUCE experiment to the introduction and/or discussion (https://mnspruce.ornl.gov)

Line 78: Another paired flux tower study comparing forested and non-forested peatlands in the sporadic permafrost zone is published by Helbig et al. (2017; https://doi.org/10.1111/gcb.13638)

Line 83: Park et al (2021; https://doi.org/10.3390/atmos12080984) is another study on Russian peatlands.

Line 93: I think the latitude/longitude coordinates should be listed here for both sites.

Line 106: The growing season definition could already be introduced here.

Line 146: It is unclear what a “standard design” is. Please clarify.

Line 148: The tower height is 29 m, but trees reach up to 27 m. It seems as if the EC measurements could be most of the time in the roughness sublayer affecting the validity of the essential EC assumptions. Perhaps the authors could explain how this potential issue was addressed.

Line 200: Why was VPD not included in the GPP response?

Table 1 and other tables: At least for the long-term means, the standard deviation should be included in the table. It would also help to characterise how strong the climate anomalies were.

Line 262: Is there a relationship between precipitation and water table depth?

Line 273: This is one example where the claim that “strong dependence … was not found” should be backed up with statistical methods.

Line 284: It is unclear where this hypothesis is coming from and how it is backed up.

Table 3 and other flux tables: Should include uncertainties in aggregated fluxes.

Line 324: Leaf-on and leaf-off might not be the right terms for evergreen ecosystems. Start and end of growing season might be more accurate.

Line 355: It seems as if the Q10 model was fitted to the entire dataset. Did the authors consider analysing short-term variations in temperature sensitivity (see Reichstein et al., 2005;  https://doi.org/10.1111/j.1365-2486.2005.001002.x).

Conclusions: In my opinion, the conclusion would be more impactful if it was shortened and if the main take-home messages were highlighted here.

In “Interannual variability of the ecosystem CO2 fluxes at paludified spruce forest and ombrotrophic bog in southern taiga”, Mamkin et al. present CO2 flux data and analysis at two taiga peatland sites in western Russia. They highlight the interannual variability in the CO2 fluxes and driving meteorological and environmental conditions at and between both sites, with implications for the future net carbon balance of this region and ecosystem due to climate change.

Overall, this is an important topic and the study presented here is, for the most part, thoroughly and completely introduced, described, and discussed, with results placed in a proper context. The study data are great for long term climate trends in a sparsely monitored region, and the paper shows well how ecosystem warming has varied impacts depending on seasonal timing. However, many English-language errors greatly hinder the paper’s readability and must be corrected. Additionally, all discussion of uncertainty in the CO2 flux measurements and partitioning methods is absent and must be included prior to publication in ACP.

More specific comments and suggestions are listed below:

Figure 1a: This figure would be more useful with country borders, lat/lon descriptions, and more contrast in colors between the different land cover types.

Line 100: I wondered why air temperature was not used from MS site when introduced here. It is later mentioned to be not available, perhaps move this mention earlier.

Line 108: Add additional context for CMI range of values, for those not familiar.

Line 110: At which site or sites is this regional trend detected?

Figure 1b: Not cited in text.

Line 165: This paragraph continues description of OB site, but the paragraph break without further mention of OB makes this unclear.

Line 176: Should this be “2015-2020”?

Line 179: What about the “2” quality flag makes that flux worthy of being removed?

Lines 176-184: As mentioned above, this section must be expanded to include description of error and uncertainty associated with eddy flux measurement, calculation, and partitioning of GPP and TER from observed NEE. Perhaps a comparison of the derived TER and GPP from isolated NEE alone (section 2.4) with the automated partitioning would be useful. Further discussion later on should reference how the results could differ based on the potential errors and uncertainty.

Figures 2 and 3: It may be more effective to convey interannual variability in meteorology and CO2 flux as anomalies from a mean set of values, rather than a timeseries. This is especially true when referring to differences on a monthly scale, such as early snow-off in a particular year.

Line 269 and elsewhere: Considering add in mention of processes when referring to numbers such as NEE. Rather that or in addition to saying “NEE decreases”, one could say “net CO2 uptake increases”.

Line 370: Why does GPP determine the parameters between the sites? Because of relative constant Rg?

Lines 373-379: Was there similar (any?) interannual variability in the TER parameters as for GPP shown here?

Line 473: The predictive relationships between environmental drivers and CO2 fluxes mentioned here are not shown. A figure or statistics that illustrate these would be useful.