General View

I was glad to read this report about long-term series of measurements of carbonyl sulfide fluxes over boreal forests. This 5-years study delivers a large data set and a great overview about the seasonality and interannual variability. Furthermore, it greatly supports the process study-based knowledge that stomatal conductance is one of the keys to interpret the flux behavior under day and night conditions, though the final consumption of COS is light independent, based on the enzymatic degradation by several enzymes. The results of this strong project must be published. However, the current version can and should be improved. The weaknesses of the current parameterization should be discussed. This parameterization is only a first attempt to interpret the monthly fluxes and embed the measurement data into a wider description based on environmental factors, such as the measured photosynthetically active radiation, vapor pressure deficit, air temperature, and leaf area index. The procedure is valid to get an overview but this parameterization obviously still needs some fitting parameters. What does this fitting mean? Altogether, the fitted simulation matches the general exchange pattern over several seasons in a satisfactory manner. However, the total net flux is a result of a complex process which is affected by contributions of trees, cryptogamic covers and soils which may contribute significantly as it has been demonstrated earlier by several process-based studies (see review by Whelan et al. 2018). Here a complete overview about the environmental factors will be of help for the reader to understand.

Specific Requests

One of the most important products of this paper is the development of a parametrization for FCOS. The reader automatically searches for information under Material and Methods. However, there one finds only a short chapter about the fitting parameters. I propose to shift the description with the corresponding formulas from the Results-section to Material and Methods. Furthermore, the information that fitting parameter were retrieved with MATLAB is not really satisfying. I miss information about the basics, i.e. PAR, Ta, VPD, and total LAI data over the 32 months. This information is needed to get a feeling for interannual and seasonal fluctuations. Why not producing a figure giving an overview about these environmental keys in a similar manner as for the fluxes in figure 1? Thus, the reader would get a picture about the seasonal microclimate.

Eddy correlation measurements deliver data for fluxes under turbulent conditions. I miss some critical remarks concerning nighttime flux data. As discussed by the authors, higher vegetation with a well-adapted gas exchange under stomatal regulation can be ruled out, though not completely. However, there can be a strong uptake by soils and cryptogams during the night which becomes strongly visible under a stable nocturnal boundary layer. The authors interpret non-zero uptake value in the dark simply as the ecosystem nocturnal COS uptake. Could this be specified and related to soil and cryptogams? The finding that the temperature is the most important factor governing FCOS at least gives room to have a closer look.

The simple semi-empirical parametrization, as the authors call it, is very helpful. However, I miss a holistic overview based on this promising long-term study at Hyytiälä. Which information can be given about the meteorological background, plant development and seasonal behavior? Where are reports about simultaneous measurements of exchanges with canopy and soils and the atmosphere? I expect this diversity is of special importance to handle flux measurements above a boreal forest  which can present a quite open structure.

There are some disagreements between measured data and simulated ones as indicated in figure 1 for the years 2013 and 2014. This is only poorly addressed in the text. Are there any environmental factors to be made responsible? Meteorology? Plant development? See my comments above. A report about the environmental factors which are the basis for the parameterization will clearly help. Within this context, vapor pressure deficits not only affect higher plants with stomata but also the water content of cryptogamic tissues

The light saturation of FCOS (Fig. 2d) is interesting. How is this understood? The consumption of COS by the enzyme carbonic anhydrase is light independent and stomatal restriction under the given light intensities should not be expected. Besides carbonyl anhydrase, a contribution of the carboxylation enzymes phosphoenolpyruvate carboxylase and ribulose-1.5-bisphosphate carboxylase has been already reported  earlier (Protoschill-Krebs and Kesselmeier, Bot. Acta 105, 1992, 206-212) and may help to understand a light saturation. Interesting to see this within flux data. Is this light saturation incorporated into the parameterization?

A remark concerning the numbers derived. Sandoval-Soto et al. (Biogeosciences, 2, 125–132, 2005) made the first attempt to recalculate the global budget for the COS vegetation sink based on GPP but corrected by measured COS/CO2 uptake ratios (experimentally obtained deposition velocities). They used older primary productivity data (Whittaker, R. H. and Likens, G. E.: The biosphere and man, in: The Primary Productivity of the Biosphere, edited by: Lieth, H. and Whittaker, R. H., Springer Verlag, New York, 305–328, 1975) to upscale COS sinks. This attempt demonstrated a clear underestimation of the global vegetational sink at that time and initialized re-estimations. Within their data sets, they came to a number of 0.036-0.063 Tg [COS] per year for the uptake by boreal forests, which is equivalent to 19.2-33.6 Gg S per year. Interestingly, this number is ranging quite close to the 16.6 Gg as estimated by the authors on the current Hyytiälä data. It is interesting to see the developing numbers in view of all the uncertainties. Within this context, I would like to propose to change a principle of citation. In the first and second line on page 2 there is a number of papers cited to demonstrate the relationship between COS and gross carbon uptake. The choice looks arbitrary. I think it would be adequate better to skip all the citations and cite one recent review paper Whelan et al. (Biogeosciences, 15, 3625–3657, 2018) instead. The history and number of important contributions to this special topic are much larger and the proposed review is giving a more complete story.

The eddy flux covariance data from this research effort is a valuable contribution to our community. Vesala and his team have worked hard to advance our understanding of ecosystem carbonyl sulfide (OCS) exchange.  It would be appropriate to do more with this data and to set it into a greater context.

The motivation for studying OCS over terrestrial ecosystems is to be able to observe leaf-level behavior, specifically stomatal conductance, at scales larger than a leaf. For over a decade, the great promise of OCS has been to improve the representation of stomatal conductance in land surface models. However, there are other equally important questions to be answered, e.g. the contribution of nighttime stomatal conductance to the atmospheric water budget. Given the potential applications of OCS data, the interpretation and application of this dataset appears incomplete.

Vesala et al. present an impressive observational dataset. They proceed to optimize a set of equations that comprise a semi-empirical model of OCS (Eq 2-5, the results of the first optimization plotted in Figure 1). In order to compare to the land surface model SiB4, the optimization is performed with the same equations again, but this time using variables from SiB4 for the 0.5 x 0.5 degree grid cell containing the field site.  The observations are not on a 0.5 x 0.5 degree scale. SiB4 has OCS leaf and soil exchange coded into it. SiB4 can also be run on the site level. An explantation is needed as to why the optimization was performed twice in this manner. Why not compare the OCS exchange predicted directly by SiB4 to the output of semi-empirical model run with related inputs?

The largest departure between the optimized model and observations occurs in July 2014. It is noted that VPD and temperature were both high in the afternoons during this period. Was this a drought or typical variation? This could be an interesting time period to investigate improving the modeling effort. Error attribution would be helpful and/or comparing to another model like SiB4.

Some explanation as to the provenance of the process-informed equations would help a wider readership understand where they came from. For example, the VPD equation looks related to the one mentioned in Medlyn et al., 2011 which itself was based on earlier work. It would be useful to justify why this approach is used instead of Ball-Berry, for example. Also, I have not encountered the variable S before to be able to recognize it as a measure of phenology.  All this is not helped by some possible errors in rendering the equations themselves.

This analysis does not partition ecosystem fluxes between soil and leaf uptake.  Sun et al., 2018 measured and analyzed a substantial soil flux record at this site. While it is difficult to satisfactorily model the soil fluxes here, it is straightforward to average them over seasons and include them as an uncertainty for analyzing the ecosystem OCS fluxes. This may prove to be important since the semi-empirical model is based on leaf interactions alone. Related, examining the daytime/nighttime OCS fluxes could be an important foray into nighttime stomatal conductance if soils and other minor fluxes are taken into account.

There are many directions to improve the impact of this study as you suggest in section 3.5. Is there a way that we could use this data and this modeling effort to improve the performance or diagnose problems in SiB4 or another land surface model? Can we use this information to compare the many large-scale regional modeling efforts that have recently come out? Of note, the study by Hu et al., 2021 focusses on the Arctic and Remaud et al., 2021 finds a different pattern than the Ma et al., 2021 study.

Finally, some additional data does exist in the broader study area. Multiple efforts by Rastogi et al. in 2018 examined fluxes at a site that appears to be in the southwestern corner of Figure 5. It would seem prudent to compare your model performance with an additional observational dataset.

Minor comments

194-200 and Table 1: This effort might be better put into the supplement for a curious reader. The main modeling work here is not a multivariate linear regression and including this in the main text seems a bit confusing.

201-209: This analysis could be better incorporated into the overall narrative of this study. The wavelet analysis revealed the lack of time lag between OCS uptake through stomata and PAR, which is reassuring but not surprising. Kooijmans et al., 2019 did an excellent job investigating the role of VPD and OCS uptake at this site. It is not obvious what the wavelet analysis achieves here.

242: Were there bounds on the possible values for the empirical fits? Since many of the fittings parameters have multiple significant digits, it stands out that b=1000.

257: Can some of the increase in OCS uptake interannually be attributed to the canopy growing 2 m over this time period?

This is interesting and important work and I look forward to seeing its evolution.

Mary Whelan

Citations

Hu, Le, Stephen A. Montzka, Aleya Kaushik, Arlyn E. Andrews, Colm Sweeney, John Miller, Ian T. Baker, Scott Denning, Elliott Campbell, Yoichi P. Shiga, Pieter Tans, M. Carolina Siso, Molly Crotwell, Kathryn McKain, Kirk Thoning, Bradley Hall, Isaac Vimont, James W. Elkins, Mary E. Whelan, Parvadha Suntharalingam: COS-derived GPP relationships with temperature and light help explain high-latitude atmospheric CO2 seasonal cycle amplification, Proceedings of the National Academy of Sciences Aug 2021, 118 (33) e2103423118; DOI: 10.1073/pnas.2103423118

Rastogi, B., Berkelhammer, M., Wharton, S., Whelan, M. E., Meinzer, F. C., Noone, D., and Still, C. J.: Ecosystem fluxes of carbonyl sulfide in an old-growth forest: temporal dynamics and responses to diffuse radiation and heat waves, Biogeosciences, 15, 7127–7139, 2018. https://doi.org/10.5194/bg-15-7127-2018

Rastogi, B., Berkelhammer, M., Wharton, S., Whelan, M. E., Itter, M. S., Leen, J. B., Gupta, M. X., Noone, D., and Still, C. J.: Large Uptake of Atmospheric OCS Observed at a Moist Old Growth Forest: Controls and Implications for Carbon Cycle Applications, J. Geophys. Res. Biogeosci., 123, 3424–3438, 2018. https://doi.org/10.1029/2018JG004430

Remaud, M., Chevallier, F., Maignan, F., Belviso, S., Berchet, A., Parouffe, A., Abadie, C., Bacour, C., Lennartz, S., and Peylin, P.: Plant gross primary production, plant respiration and carbonyl sulfide emissions over the globe inferred by atmospheric inverse modelling, Atmos. Chem. Phys. Discuss. [preprint], https://doi.org/10.5194/acp-2021-326, in review, 2021.

See attached file