General comments

The manuscript entitled ‘Local processes with global impact: unraveling the dynamics of gas evasion in a step-and-pool configuration’ by Peruzzo et al. focuses on the mechanisms driving high gas exchange velocity in step-pool streams. The authors used an artificial step-pool flume to evaluate gas exchange by investigating the role hydrodynamics on energy dissipation. Gas exchange is driven by diffusivity but enhanced with turbulence and gas entrainment in bubbles. Turbulence and bubble-mediated gas exchange is not easy to separate, which the authors do here and is a key component of identifying the mechanism behind high gas-exchange in mountain streams. The authors found that gas exchange in the artificial step-pool was highly heterogenous.  The greatest gas-exchange was at the spout – where the water dropped from the step to the pool. Also, under steady-stream conditions, the area of bubbles created from the spout varied spatially over time.  Bubble mediated gas exchange varied and was 2.5 to 5-fold greater than turbulent gas exchange.  Overall, the authors found that bubble mediated gas exchange dominated in the ‘bubble’ zone whereas turbulent mediated gas exchange dominated in the ‘calmer’ zones of the flume with both types of gas exchange being highly spatially heterogenous. This is likely driven by the strong spatial heterogeneity of energy dissipation.

The authors discuss the importance of taking spatial heterogeneity into account in these types of streams, which I agree given their findings. However, I think scale is also highly dependent on how one might evaluate or estimate gas exchange. In field settings, we do rely on reach-scale gas exchange estimates in mountain streams – often from tracer gas experiments – as other methods are not feasible (i.e., night-time regression from continuous oxygen measurements, domes given the high turbulence do not often work – besides some of the work cited here, see also Hall Jr., Robert O., and Hilary L. Madinger. “Use of Argon to Measure Gas Exchange in Turbulent Mountain Streams.” Biogeosciences 15, no. 10 (May 18, 2018): 3085–92. https://doi.org/10.5194/bg-15-3085-2018).  The work here illustrated how heterogenous k can be in a 120cm2 flume. At some point, averaging does occur - even within this experiment.  I would caution suggesting that reach scale metrics in mountain streams are not adequete (if I understood the discussion correctly) - but rather selecting the reach with the heterogenity in mind would be a helpful step of estimating gas exchange in the field as we try to move towards some kind of scaling mechanism or even for any reach-specific studies.

Specific comments

L21: missing a closing paratheses after ‘such as CO2 and N2O to the atmosphere)

L36:  missing ‘s’ for ‘see e.g. Cirpka et al. 1993’

L196: I appreciate quantifying the bubbles as bubble mediated gas exchange can be so much greater than turbulent driven k.  The bubbles that were ‘hardly feasible’ to detect – would they not still have a significant effect on gas exchange significantly from turbulent driven k? On L264 it is discussed heterogenous size of the bubbles should not matter when estimating k. However, the diameter of the bubble is accounted for in equation 6. Perhaps as long as the bubbles are below 0.82 mm for radius, then bubble size should not matter as much? 

L202:  'The decrease in [energy dissipation] with the depth was more than linear'.  I don't quite follow 'more than linear' - do the authors mean the relationship was not linear?  I suggest expanding what kind of relatinship was evident here for clarification.

L205: I don’t quite follow ‘the energy dissipation rate was one order (at least two orders)…’ I suggest for clarity – if it was indeed 2-fold difference in energy dissipation between the two regions – state that it was two-fold.

L266: I absolutely agree that likely in mountain streams – when bubbles are present – bubble-mediated gas exchange drives k.  This work here is a great push forward to demonstrate bubble-vs-turbulent driven gas exchange by quantifying kb vs kt.  However, this is not the first work to propose kb may dominate gas exchange in highly turbulent streams, which I suggest be cited appropriately.

The article by Peruzzo et al explores the gas exchange dynamics in a step-and-pool situation, which is a common geomorphic feature in many streams with some relieve like mountain streams. To do this they use a lab setup, with a unique monitoring set that measures water velocity and turbulence in multiple directions and at a high spatial and temporal resolution. This is something I have rarely seen so the authors can explore experimentally at a great resolution the drivers of gas exchange in waters. They can also separate bubble mediated gas exchange which is important ad often ignored. They find a high spatial in the gas exchange, both turbulence and bubble mediated, with a very small area accounting for most of the gas exchange with the atmosphere. The article is well written and figures of high quality, but the technical level of some of the section may be hard to follow for not only the average reader of “Biogeosciences”, but also for someone who works with gas exchange in rivers without a hydrologic engineering background like me. With this I provide some suggestion to clarify some terms, notations, and  provide citations on those technical aspects of the work.

As said, the work is of high quality, but I do have a major issue with the take home message of the article, which can be summarised with the last sentence of the discussion: “Based on the above arguments, we propose that the use of the mass transfer rate, k, should be dismissed in cases in which the heterogeneity of the flow field controls the fraction of mass evaded into the atmosphere, as in our step-and-pool configuration.” I fully agree that gas transfer velocities are tricky to measure, highly variable in space and tricky to translate from one spatial scale to another. There are also multiple methodologies for different use cases, see the review by Hall and Ulseth 2020 (WIRES) for a great overview. In this case, a reach-scale measure of the gas transfer velocity using a gas tracer, where a inert gas is injected upstream and the loss is quantified downstream with reaches of 50-150 meters for example, is a perfectly valid method to quantify gas exchange in a stream with steps and pools.  Spatial aggregation is indeed important, and how those reach scale measurements are translated to catchment scales remain an unanswered issue for instance. I will agree with the authors that for example, using chambers in a step and pool system is not a good method, but their claim is a bit overreaching.

I will conclude this with an analogy to another physical and turbulent system. Temperature of a fluid is some kind of emergent property, which is related to the average movement of all molecules. You could use some great technology to track and quantify the movement of all particles, only to realise that the system is highly chaotic and heterogenous with a lot of eddies, indicating that is rally hard to quantify the movement of particles. Regardless, at a larger scale we have some other tools to estimate temperature of that system at larger scales that may be a simplification but work well enough. This study is a bit similar in this sense, it provides unique insight of the fine scale turbulent dynamics of water, suggests that is highly heterogeneous, but the link to the larger reach scale is a bit weak that may need some improvement or toning down in the text.

Below I detail some minor issues I found in the text:

Abstract

-L10: maybe is cleaner to put k in parenthesis?

Introduction

In the first and second paragraph, the authors focus very fast on mountain streams. Step and pools are common in other landscapes outside mountains, so maybe it would help rising the generality of the article to discuss this more broadly outside mountain streams.

Methods

-L121: The symbol of a bar with a dot above and one below may be unfamiliar to many readers of Biogeosciences. Define it in parentheses.

-L121-132:This whole paragraph is highly technical but still can be understood by a broader readership.This would be more likely with the support of more references as only Zappa et al 2007 is cited here. For example it would be helpful to have a citation after “Batchelor scale” (L124), “in its energy cascade (L126).

-L134: citation supporting this? Maybe Hall and Ulseth 2020 WIRES water?

-L161: A brief explanation of this solubility coefficient?

Results

-L177: “Provided by the code” might be better to say “provided by the model”

-L190 “Typo in “Negative values are observer” -> observed

-L205: It is unclear if it was one or two orders of magnitude. Might need to rephrase.

-L223: No need to say what you did or what you show in a figure. You can directly explain the observation and cite the figure in parenthesis.

Discussion

-the discussion is very well written, despite the main comment that mostly concerns the discussion I have no small issues.