Dear Authors,

An understanding of atmospheric OCS sources and sinks enable the ability to ascertain plant functioning on an integrated, regional scale inaccessible to other methods. This manuscript presents an effort in untangling OCS gross fluxes over a specific region. Some additional analysis and editing are needed to realize the potential of the study. Below I have some major questions followed by a few minor ones.

For the first investigation that Kooijmans et al published in 2016, there is an entire year of calibrated data, but the additional data presented here is at a single tower measurement height for 2 months without a calibration cylinder? Is there something missing in the description in the text?

In this study, the tower footprints over time were calculated by STILT and the concentrations of the tower were calculated based on assumed fluxes at the surface. Attribution was estimated on page 13 based on footprints during periods of trace gas enhancement.

If we want to do something truly powerful with this data, we can take the known flux estimates as priors and generate new maps of surface fluxes based on observed concentrations at the tower (averaged over afternoons where nighttime inversions have already been dispensed with. Calculating footprints when the PBL is on the move, e.g. at midnight, is error-prone.) This atmospheric inversion would give you a stronger, data-based hint about where the missing sources of the region are and requires no further field measurements.

That said, the uncertainty introduced by using the STILT model is not sufficiently addressed. Derek Mallia at the University of Utah writes articulately about the STILT model and it’s application to regional fluxes. The recent update to the STILT model – which version did you use? – makes the analysis more user friendly than previous versions. There has also been work done by Anna Michalak’s group in analyzing uncertainties in this type of analysis.

The analysis in 3.1 may belong in the supplement with Figure S1. It is a look at wind direction and deviation from a calculated seasonal average. Using the flux-gradient method or approach, a flux estimate could be made based on concentrations measured at two different heights (along with high frequency wind and temperature data).  However, the conclusion of the analysis here is unsatisfying – we are no closer to knowing the sources and sinks of OCS in this region, but rather again acknowledge that atmospheric mixing affects OCS concentrations. At the same time, it seems like a great effort was made to calculate nighttime fluxes with Rn, with no further use of the flux estimates.

This study misses some context. For example, there are other places in Europe collecting OCS concentration data and an extensive N American dataset that could be used to figure out the seasonal cycle. Some recent efforts to better quantify anthropogenic sources by Sauveur Belviso, who I see has already reviewed this manuscript, would be prudent to include in the interpretation.

In short, this project moves us towards answering several interesting questions in our community, but the analysis is incomplete.

Minor Comments

Figure 1 and site description: the site description gives context to the Lutjewad tower that is lacking in the map. Maps are difficult to make well and I found myself sketching a separate map to understand the greater context. It would be useful to mention that the ocean, aluminum smelting, wetlands, and winter wheat are all known sources of atmospheric OCS. Figure 1 and several other figures need a more robust caption.

Table 1: Is ploughing a source of OCS? Or is the ploughed soil?

P6, L18: Mentioning why these extra cylinders were collected would be helpful here, even if the details are included in the supplement.

P7, L17: Emission rather than exhalation? Or is this a term specific to Rn?

P7, L23-24: Is simply taking the average the “done” thing for dealing with Rn emission variability? Can you cite another group or two who have done this and perhaps did a sensitivity analysis or similar?

P7, L26-27: The methods are cited, however, can you give a 1 sentence explanation for why the method only works at night? It seems earlier in the paragraph there is a comparison between daytime and nighttime PBL. While we’d expect photosynthesis to cease at night, making the nighttime fluxes easier. Is that’s what’s happening here? I know you cite the papers that include more detail on the methods and I could read those and piece it together myself, though as it stands the paragraph here is confusing.

P8, L13: What does it mean for a footprint to be negligible?

P16, L25: For a guassian distribution to be a useful model here, certain assumptions must be met. A justification of these would be useful here.

P17, L19 and on: this goes into discussion rather than results.

P19, L20: The word “prove” is too strong here.

P19, L20-30: Too much faith is being put into the STILT analysis. Note that the model is run with imperfect data, the PBL height is often off and this effects the size of the “box”.

P20, L29-31: I’m not sure this conclusion is justified.

P20, the rest of section 4.2: This reads like speculation when you have an analysis with associated uncertainty to rely on.

P21, L26: In conclusion, this inversion analysis is incomplete.

Thank you for your efforts so far. This is an interesting dataset and is moving towards the most interesting application of atmospheric OCS observations. 

Mary Whelan