Comments for Paper entitled “Sensitivity of climate effects of hydrogen to leakage size, location, and chemical background”

This paper investigates the impacts on H2 and CH4 GWP100 due to changes in three different scenarios settings. These settings are : 1) changes in hydrogen emission perturbation, 2) pulse emissions of hydrogen at specific locations, and 3) three different SSP scenarios. They find that the H2 GWP100 does not depend on the magnitude of hydrogen emission perturbations. For specific locations of H2 emissions, the H2 GWP100 is different in locations far from soil uptake e.g. in the ocean and in Antarctica. For different SSP scenarios, the H2 GWP100 is dependent on both CH4 concentration and the NOx:CO ratio, both of which heavily influence OH and, by extension, H2 atmospheric lifetime and GWP100. Ultimately, however, the soil sink is the dominant driver factor for H2 soil rather than the H2 atmospheric lifetime. With the exception of the pulse experiments located in the ocean and Antarctica, all H2 GWP100 results are within one standard deviation of the GWP100 found in Sand et al. (2023).

This is a comprehensive study assessing the GWP100 of H2 under different situations. Hydrogen is an important topic in both in research and society and this is a valuable contribution to the hydrogen community in narrowing down the uncertainty of H2 and its impact on the climate.

General comments

- Some of the results would benefit from further quantitative analysis to state whether these values are statistically significant (see comments below).

- I think the authors need a further explanation as to why they’ve chosen their locations and to define what is meant by a “sink soil active area” as this is unclear to me.

- Given that the authors have described how interconnected the reactions are in the atmosphere, would be useful to have a forth experiment where both the H2 emissions and the CH4 concentration is enhanced. This would then clarify whether the enhancement due to H2 emissions and CH4 concentration is additive or if increasing both CH4 and H2 causes an further increase due to OH production/loss e.g. anthro1 for H2 and 10% increase in CH4 for present day. Author may have already taken this into consideration, in which case this should be clarified in the text

- The conclusion of linearity between GWP100 and H2 emission perturbation is currently misleading and confusing, especially as authors later say these two variables are independent. From Fig 3 it is difficult to see how these are linear as well

Specific comments

Abstract

lines 9-10 : This sentence is vague – specify the reactions (OH induced) and what the effect on CH4, O3 and H2O are

Line 14 : See later comments about linearity of GWP100 + size of emission perturbation. Also, if it is linear, this is only true when you consider the magnitude of emission perturbation as the emission perturbations chosen are logarithmically increasing

Line 20 : This is a strong statement and as you point out in the conclusions it isn’t taking into account soil sink. Can you add a clause that this is only considering OH sink of H2?

Introduction

line 25 : Specify how it will cause these greenhouse gases to change

line 44 : Missing refs. Either add more in, or give e.g.

Methods

line 75 : expand full model acronym at first instance

line 80 : Does the third experiment also have enhanced H2 emissions along with an enhanced CH4 concentration? Adding a sentence in to clarify would help this.

Line 86 : Merge into previous paragraph

Line 87 : Define what is meant by linearity here – if the emission perturbations are increasing by a magnitude of 10 in each experiment, “linearity” is misleading here unless you specifically refer to the logarithmic increase of perturbations

Line 87 : In line 79, the authors say they run 3 simulations to calculate GWP, and here they explain they have 3 sets of sensitivity tests (which include multiple simulations). I assume these are separate to the GWP runs described previously? Please could the authors add in a few sentences at the start of the methods section to summarise all the set of experiments they’re doing more clearly. Authors might consider moving the “GWP Calculation” section to after the “Sensitivity experiments” (after Lines 103-109) have been explained to help with layout of explaining their experiment setup

Line 100 : “correspond to two different control simulations, as hydrogen is concentration driven in the methane perturbation simulation.” Please expand on the differences between the control simulations.

Line 113 : “The OsloCTM3 model is used in a similar set up as in Sand et al. (2023)” Please give a brief description of what the set up is here.

Line 120 : Authors might consider using the corrected oceanic H2 emissions from Paulot et al. 2023 (Fig S3)

Results

Section 3.1: The title of this section is quite ambiguous at a first glance. Please could the authors rename it to be more informative

Line 163 : Similarly, in this sentence could the authors clarify “the results” (I assume they mean the H2 GWP100 values from the previous section?).

Line 165 : It’s difficult to see how the GWP100 values have a linear relationship wrt magnitude of emission. The anthro100 GWP actually looks like it is lower than the anthro10. These values are very close together – can the authors authors say whether or not these are significantly different enough for it to be considered linear?

Line 178 : How do the authors define an soil sink active area? Is it based on the average soil uptake over an area?

Line 180 : Can the authors show these differences are statistically significant from one another? I think this would strengthen their argument

Line 185 : Referring to Fig. 1b at the end of this sentence is somewhat confusing as the numerical values (5.2 and 5.4) refer to Fig. 4a. The authors could move this Fig reference to a more suitable place in the sentence or leave it out entirely.

Line 195 : Could authors give an equation for the feedback factor than rather a word description and also define lifetime of perturbation for a complete explanation

Line 204 : Can the authors expand more on what is significant of the lowest feedback factor being 0.76 would mean in the larger context?

Line 201 : There doesn’t look like there is much change in forcing if there is an increase in H2 burden from Fig 4. and all the GWP values are similar. Are these statistically significant values from each other to make this statement? Authors say that it is within the uncertainty range of values from Sand et al. 2023 so perhaps not?

Line 236 : Worth mentioning that the dominant chemical loss of H2 is via OH and/or refer to equation 1

Line 240-253 : This is nicely explained

Line 253: Can authors comment on the effectiveness of NOx: CO ratio and changes in CH4 of H2 lifetime?

Line 174 : Can authors suggest why the ozone contributes are greater in SSP119 than the others? Or is it that the ozone contribution is the same in all scenarios, but the contributions from strat. H2O and CH4 are lower in SSP126, resulting in a large proportion from ozone?

Line 297 : I might be misunderstanding this, but earlier in lines 163-6 authors say results are linear (which is unconvincing from the graph), but also say they are independent with respect to the magnitude of emission perturbation. These two conclusions seem mutually exclusive. Can authors describe what they mean by “independent” of hydrogen emission perturbation, but have a “linear response”?

Line 306 : Can authors clarify how the longitudinal dependence links to soil sink active areas (e.g. inland vs by the coast?)

Line 315 : Given that the authors are looking at particular locations, it’s not surprising that GWP values are outside the standard deviation values from Sand et al., especially as these are at the extremes which won’t be captured in a standard deviation range. Consider rephrasing the sentence 314-316 so it is less defensive of this result!

Line 335 : Rephrase to “lifetime decreased by a range of 0.19-1.1 years” or equivalent

Line 368-9 : Can authors rephrase this sentence to include quantitative values to support their argument?

Line 370-1 : “due to process understanding of the hydrogen budget.” I assume the authors specifically mean the soil sink budget?

Table 2 : State what the starting concentration is for CH4 concentration in present-day without 10% increase

Fig 1b : Could the authors also provide a figure of soil deposition in units commonly used in other papers e.g. cm s-1 so it can be compared to other soil deposition models (e.g. Paulot et al. 2021, Bertagni et al. 2023).

Fig 2a: I assume the authors mean the surface methane concentration – can you add this into the caption for clarity

Fig 3: Are these values averaged over one year or multiple years?

Technical comments

line 56 : NOx vs NO_x

line 68 : “as well as”

line 86 : Combine sentence with previous paragraph

line 87 : First sentence is a repeat from paragraph before – remove

Fig 1b: Move the caption or change the colour in the high northern latitudes as it is difficult to read

Table 1: missing “s” in “magnitude of the emission perturbation”

line 276 : Missing % after 3.0

line 296 : No need to reference Sand et al. At the end given the start of the sentence

line 360 : NO_x vs NOx

line 370 : process → processes

This manuscript describes an assessment of the climate impacts of hydrogen leakages and the sensitivity of these to emissions magnitude and location under different future scenarios for atmospheric composition. The study concludes that the climate impacts are largely independent of the magnitude of emission changes, that there is some dependence on location (although this is small for expected emission locations), and that the impacts will change under different future atmospheric composition, but by only a small margin. This is the first time that these effects have been explored in a consistent way in a global chemistry transport model, and the results are valuable, giving confidence to previous assessments of global warming potential. The methodology used is sound, although there are a number of places where it could be described more clearly. I have a number of minor concerns (detailed below), but in other respects this is an interesting and valuable paper, and merits publication in ACP once the following issues have been addressed.

General Comments

The paper refers heavily to the study of Sand et al 2023 and is largely presented as an extension of this previous work. While reference to this useful study is certainly merited, the paper would be stronger if presented in a more independent manner so that the reader does not feel that they must read the earlier study first. Frequent reference to the Sand et al. paper detracts from the novelty of this work.

The differences in GWP due to location are attributed to the H2 soil sink. While this effect is likely to be dominant, the abundance of OH in these locations also differs substantially. What are the relative contributions of these two sinks for the applied perturbations? No evidence is presented that the soil sink is the sole cause of the differences.

The term "feedback factor" is not used correctly here. Methane is sufficiently long-lived to be well-mixed in the troposphere, and is a substantial sink of OH, and therefore has a chemical feedback on its own lifetime. A methane perturbation thus decays more slowly than expected as global OH is perturbed.  Hydrogen has a similar, but smaller effect on OH, and therefore has a feedback factor slightly greater than one for a well-mixed global perturbation. The different perturbation lifetimes for hydrogen identified here are due to differences in distribution (and thus to the relative balance of sinks) and not to a specific feedback process. The soil sink, in particular, is first-order, so no feedback is possible. While the tropospheric mean lifetime will increase or decrease under different perturbations, depending on location, it is misleading to describe this as a feedback factor. The perturbation lifetime is simply different from the global lifetime, as is true for most reactive gases. Different terminology is needed here.

The paper notes that the locations with the largest GWP100 values are not very relevant for the future hydrogen economy. What conclusions can you draw from the sites that are most relevant for the hydrogen economy? If future emissions increase in populated continental regions, will GWP100 be lower than estimates for current conditions?

Specific Comments

Line 23: "only water vapour is emitted": this is not true if hydrogen is burned, so please rephrase.

Line 74: "steady-state perturbation approach" is not immediately clear to readers who haven't read the Sands et al. paper. It would be clearer to use "emissions perturbation approach" here and then explain the steady state aspect in the next paragraph.

Line 82: please explain why the methane concentration is fixed at the surface, and note that this prevents it responding to the changes in OH due to hydrogen. The reader needs to appreciate this to understand why the third simulation is needed. Is the enhanced methane concentration in addition to the enhanced hydrogen emissions, or instead of them? Explain how the methane adjustment is made (or are these all +10%?)

Line 93: please introduce reader to the seven locations before referring to them (perhaps just reference Fig 1 at end of the previous sentence).

Lines 114-115: "has been updated", "has been corrected": please explain how (briefly) so that the reader can understand the changes.

Line 119: Which version of GFED? version 4?

Line 129: please explain (briefly) what the adjustment term is for. 

Fig 1: Please adjust the position of the site labels so that they are legible. Given that the units are the same, why are the emissions presented on a log scale but the deposition on a linear scale? This is misleading and makes the panels difficult to compare. 

Line 162: "Are the results linear" is unclear as a subtitle given the different sets of simulations described. Please use "Is the response to emissions size linear?" (or something similar).  The conclusion of this paragraph is that the response is linear, but yet the results differ slightly; what is the uncertainty in the GWP calculation?  The CH4 response starts to introduce nonlinearity at higher H2 (evident in Fig 4b), so large H2 emissions (1000 Tg/yr) are expected to cause the GWP to fall. The response is therefore only linear up to a certain emission size. Please make this clear on line 166.

Line 181: global mean surface hydrogen concentration?

Line 195: "lifetime of the atmospheric component" isn't clear here; should this be the steady-state lifetime?

Line 227: Note here that the chemical loss of hydrogen is also reduced (as described in the next paragraph), but this effect is smaller, so the changes in chemical production dominate the overall change.

Table 5: The methane lifetime here is very short; the observation-based estimate is about 9.1 years, based on a chemical lifetime to OH of 11.2 years (Prather et al). This indicates that OH concentrations are too high. What are the likely impacts of this on the GWP estimates presented for hydrogen?

Line 279: These changes in methane feedback factor look very large. Please place these in the context of results from other studies in the literature (this is done later on lines 323-327, but the implications for the results presented here remain unclear).

Line 296: "within one model": the model used in this study, or a different one?

Line 297: As noted above, a caveat is needed on the conclusion that GWP100 is independent of emission perturbation size, as only small (realistic?) changes have been explored in this study.

All the location differences are attributed to the soil sink, but what about proximity to high OH regions? This is not mentioned in the manuscript.

Technical Corrections

Line 13: "perturbation" not needed

Line 14: "is linear with respect to" -> "scales linearly with the"

Line 15: compositions -> composition (and L.77)

Line 16: "and it" -> "which"  (but whole sentence needs revising)

Line 19-20:  Final sentence unclear: reverse ordering and remove negative.

Line 23: move "as an energy carrier" to the end of the sentence

Line 25: composition -> abundance

Line 36: capture -> captures

Line 45: "land-ocean fraction" -> "greater landmass"

Line 46: "higher...in Southern Hemisphere" -> "lower...im Northern Hemisphere"

Line 54: as -> such as

Line 55: leads -> lead

Line 59: remove "as"

Line 69-70: reverse sentence: sensitivity of GWP to atmospheric composition

Line 89: anthropogenic -> global anthropogenic (?)

Line 92: emission -> emissions

Line 98: "first 11" -> "first two sets of"  (much clearer to reader!)

Line 102: on -> to

Line 185: has -> show

Line 352: "it depends somewhat" -> "are somewhat dependent"

Line 370: "due to process understanding of" -> "due to weaknesses in understanding of the processes controlling"  (or something similar)

There is an additional reference (Aamaas et al.) at the end of the reference list which is out of place.