Dependency of the impacts of geoengineering on the stratospheric sulfur injection strategy part 2: How changes in the hydrological cycle depend on injection rates and model?

Laakso, Anton; Visioni, Daniele; Niemeier, Ulrike; Tilmes, Simone; Kokkola, Harri

doi:https://doi.org/10.5194/egusphere-2023-2520

Preprints

https://doi.org/10.5194/egusphere-2023-2520

Preprints

15 Nov 2023

| 15 Nov 2023

Dependency of the impacts of geoengineering on the stratospheric sulfur injection strategy part 2: How changes in the hydrological cycle depend on injection rates and model?

Anton Laakso, Daniele Visioni, Ulrike Niemeier, Simone Tilmes, and Harri Kokkola

Abstract. This is the second of two papers where we study the dependency of the impacts of stratospheric sulfur injections on the used model and injection strategy. Here, aerosol optical properties from simulated stratospheric aerosol injections using two aerosol models (modal scheme M7 and sectional scheme SALSA), as described in Part 1, are implemented consistently into EC-Earth, MPI-ESM and CESM Earth System Models to simulate the climate impacts of different injection rates ranging from 2 to 100 Tg(S)yr⁻¹. Two sets of simulations were simulated with the three ESMs: 1) Regression simulations, where abrupt change in CO₂ concentration or stratospheric aerosols over preindustrial conditions were applied to quantify global mean fast temperature independent climate responses and quasi-linear dependence on temperature and 2) equilibrium simulations, where radiative forcing of aerosol injections with various magnitudes compensate the corresponding radiative forcing of CO₂ enhancement to study the dependence of precipitation on the injection magnitude; the latter also allow to explore the regional climatic responses. Large differences in SALSA and M7 simulated radiative forcings in Part 1 translated into large differences in the estimated surface temperature and precipitation changes in ESM simulations: e.g. an injection rate of 20 Tg(S)yr⁻¹ in CESM using M7 simulated aerosols led to only 2.2 K global mean cooling while EC-Earth – SALSA combination produced 5.2 K change. In equilibrium simulation, where aerosol injections were used to compensate for radiative forcing of 500 ppm atmospheric CO₂ concentration, global mean precipitation reduction varied between models from -0.7 to - 2.4 %. These precipitation changes can be explained by the fast precipitation response due to radiation changes caused by the stratospheric aerosols and CO₂ because global mean fast precipitation response is rather negatively correlated with global mean absorbed radiation. Our study shows that estimating the impact of stratospheric aerosol injection on climate is not straightforward. This is because the capability of the sulfate layer to reflect solar radiation and absorb LW radiation is sensitive to the injection rate as well as the aerosol model used to simulate the aerosol field. These findings emphasize the necessity for precise simulation of aerosol microphysics to accurately estimate the climate impacts of stratospheric sulfur intervention. This study also reveals gaps in our understanding and uncertainties that still exist related to these controversial techniques.

Received: 29 Oct 2023 – Discussion started: 15 Nov 2023

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.

Download & links

Preprint (PDF, 63902 KB)

Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
Preprint (63902 KB)

Supplement (11889 KB)

Download & links

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Journal article(s) based on this preprint

24 Apr 2024

| Highlight paper

Dependency of the impacts of geoengineering on the stratospheric sulfur injection strategy – Part 2: How changes in the hydrological cycle depend on the injection rate and model used

Anton Laakso, Daniele Visioni, Ulrike Niemeier, Simone Tilmes, and Harri Kokkola

Earth Syst. Dynam., 15, 405–427, https://doi.org/10.5194/esd-15-405-2024,https://doi.org/10.5194/esd-15-405-2024, 2024

Short summary Chief editor

Anton Laakso, Daniele Visioni, Ulrike Niemeier, Simone Tilmes, and Harri Kokkola

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2023-2520', Peter Irvine, 07 Dec 2023

General comments
The authors evaluate the precipitation response to stratospheric aerosol injection (SAI) geoengineering, considering earth system model and aerosol microsphysical uncertainty. Prescribed aerosol fields were generated in an ESM with either a sectional or modal aerosol module, producing quite different aerosol properties and hence radiative forcings. These were fed into 3 different ESMs which simulated a range of combinations of CO2 and SAI injections. The fast, forcing-driven hydrological response was found to be quite different for the different aerosol modules as the modal module produced fewer, larger particles which absorbed more LW radiation. Despite being driven by the same aerosol field, the ESMs produced quite different radiative, temperature and precipitation responses. However, the largest differences in many respects arose from the microphysical representation. The study makes a detailed analysis of the various factors that shape the precipitation response to SAI, making clear that microphysical uncertainties are important.
This paper will make a substantial contribution to the literature, is generally well-written and has generally good quality analysis, and so I recommend that it be published after making relatively modest changes, outlined below.
The paper is generally clear and well-written, but the argument was a little hard to follow in places as the paper jumped back and forth between radiative forcing and precipitation several times. For example, section 4.4 is titled “simulated precipitation response…” but the opening page is about the reasons for a radiative mismatch. The authors may consider revising the order of analysis.
The figures and analysis are generally very good, but in places the analytical choices made things a little difficult to follow, e.g., Figure 6 was particularly challenging. I’ve made a series of suggestions for improvement in the specific comments below.
I was left not quite knowing the answer to a question that I think could help increase the impact of this study and I think that with a little work could be easily answered. There is a factor of ~2 difference in the SO2 amount needed to achieve the same cooling for the sectional and modal aerosol modules. This made me wonder: is the residual precipitation, or just fast precipitation, difference ~2x larger as well? Or does the fast effect of CO2 dominate this residual? More generally, could the authors comments on the relative scale of the precipitation differences compared to this injection amount? RMSE difference might be a simple metric that could be calculated to test this. Some take-away claim that relates these 2 key elements would make the paper more memorable and useful to the community.
Specific comments
L14 – reduction relative to what?
L16 – “rather negatively correlated” – why not just negatively correlated? And could you clarify what is meant by “absorbed radiation” here? Is that a new finding or a widely established result that you are referencing?
L30 – relative to what?
L31-34 – review phrasing.
L43 – clarify whether the same 2 aerosol modules were used in the 3 different models.
L50 – might be nice to indicate roughly the fractional changes here.
L23-50 – Might be worth indicating which aerosol scheme performs better at reproducing observed volcanic response if that can be determined, i.e., is the SALSA sectional model better but more expensive and M7 the poor-man’s alternative?
L59-60 – Does this apply in the same way to stratospheric heating as it does to tropospheric? Is stratospheric heating as effective as tropospheric heating at suppressing precipitation? If the absorption occurred up in the mesosphere, I imagine it would have little effect on the hydrological cycle.
L64-66 – perhaps note T-driven intensification under GHG case?
L75 – formatting of citations.
L78 – in the consequent precipitation responses.
L103 – add resolution in degrees.
L149 – from a preindustrial baseline with GHG and SAI perturbations applied?
Figure 1 – Great figure! Small suggestion: 6x climate responses instead of impacts.
L162 – logarithmic fit
Figure 2 – Another great figure. Wondered if it might make sense to use the shape to match models, e.g., diamonds = CESM. This might help the colorblind to follow along. Looks like that was done in Figure 3, but I’d suggest adding the shapes to the legend or caption.
L206 – will have changed when it does settle down?
L205 – 213 – a little repetitive.
L245-249 - phrasing a little unclear.
Figure 4 – Is it best to compare injection mass for Salsa and M7 directly in this way? I found myself a little confused until I remembered that 50Tg in Salsa has a much greater cooling effect than in M7. Perhaps some additional text or analysis could clarify this, e.g., normalizing the fast effect by the expected cooling magnitude or plotting against an x-axis that shows temperature or RF?
L295-296 – Would this non-linearity disappear if the axis was RF instead?
L347 – less precipitation = a greater reduction in precipitation relative to the baseline?
L350 – link back to earlier claim on reduced SO2 for same RF in EC-earth?
Figure 6 – Is this the best way to get this information across? I’m very confused by some of the analytical choices and by how complex it is. Why aren’t the points falling on the precise CO2 ppm values used before? Can the analysis be flipped so that they do? More information needed on c, to clarify modelled pairs. Panel d seems like it could have been a whole multi-panel figure of its own. I also wonder if a pure temperature adjustment is the best choice, couldn’t you also scale up or down the fast effect of SAI by the fractional change in cooling that’s needed? Presumably that would give a better fit.
L355 – conversely? should that be Additionally?
355-360 – this suggests switching axes on Figure 6, as CO2 is the dependent variable.
4.4 – Given the first page is about the radiative mismatch, should this be 2 sub-sections? And should the radiative discussion come here or earlier? This might help with the flow of the article.
L392 – global mean precipitation is more positive?
L393 – here you are referring to the effect after the fast effect, whereas in some studies it is meant to include the total effect.
L398-400 – I think making the correction I suggested and noting that the forcing mismatch produced this precipitation mismatch might lead to a more useful conclusion here.
L403-420 – Isn’t a big driver of the overcooling / residual warming seen in many stratospheric aerosol geoengineering experiments the distribution of aerosols? Might be useful to refer to that distribution here and remind the reader that it’s the same in each model (I may have forgotten myself by this point).
Figure 7 – maybe a note on how these pairings were chosen. It might be useful to extend the y axis and add a global mean temperature residual value to the legend.
Figure 8 – A bit difficult to read, would adding figure wide column and row labels make it easier to parse? You might also consider rearranging so that SALSA is as one block, M7 as another.
Figure 9 – missing labels. Panel a is quite difficult to read, some for previous figure. Is there another way to show this?
Figure 10, same comment as 8.
489-494 – not particularly clear or particularly logical flow at the end of this paragraph, consider revising.
504-505 – compared to what? Is the comparison to the baseline the most relevant? Should it be to the 500 ppm case? Given the amount of SO2 injected scales with CO2, this difference in injection amount should modulate that total precipitation response, which as a consequence shifts the net result.
L513-514 – See my earlier comment about making a full adjustment, i.e., what would have occurred if the correct amount had been chosen to keep temperature constant, rather than just the temperature adjustment (which excludes the change in fast forcing effect).
L495-513 – Here or elsewhere some comment on the relative scale of the precipitation differences compared to the required injection amounts would be useful. M7 suggests ~2x greater sulphate required, is the gross or net precipitation difference 2x greater too?
518 – more negative?
530 – consistently more negative?
538 – perhaps remind reader that they faced the same change in aerosol optical properties
543-547 – a little hard to follow.

Citation: https://doi.org/10.5194/egusphere-2023-2520-RC1
- AC1: 'Reply on RC1', Anton Laakso, 09 Feb 2024
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-2520/egusphere-2023-2520-AC1-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-2520-AC1
RC2:
'Comment on egusphere-2023-2520', Anonymous Referee #2, 20 Dec 2023
The paper provides an extensive analysis of inter-model difference in the global (and regional) precipitation response to Stratospheric Aerosol Injection. On the whole, this is a sound piece of work, carefully analysed and well written, with clear plots.
A not too major criticism is that the paper is rather technical, and readability and possibly usability by a larger set of readers could be improved by adding some clarifications and physical interpretation here and there. I hope my comments can help. In addition, there is a handful of minor improvement points regarding things like figure captions and labels, listed below.

Interpretation and readability

Overall aims: Maybe the overall aim(s) could be stated in a small number of clear research questions at the end of section 1. Currently line 81 states: “We investigate how these impacts depend on the injection rate and the aerosol microphysics model”, which is relatively vague and mixes physics questions (how does precipitation change under different SAI intensities) with modelling questions (is there model uncertainty). Unless the main focus is strictly the model uncertainty part, the paper, which is now relatively technical, may profit here and there from a bit more physical interpretation.

Structure, especially Section 4: It would help the reader to get a short hint at the beginning of sect 4 what the subsequent pieces of analysis are meant to do and how they relate to each other and the overall aim / research questions of the paper. For example, it helps to know before sect. 4.3 that you first estimate the precip change based on the radiation diagnostics and then will compare the overall change to the actual model results in 4.4.

Fast response and absorbed radiation, line 56 ff. The paragraph could be clearer. Line 57 “Further precip change” -> further with respect to what? Line 59: rather than saying “Any change in X translates to a change in Y”, it is clearer to say e.g. “Any increase in X translates to a decrease in Y”, to immediately give the direction of change. The whole sentence seems unnecessarily long-winded. Most importantly, since some of the readers may not be experts in hydrological cycle but e.g. in SRM or impact modelling, it would be helpful to explain in a bit more detail 1) what fast and slow precip responses are (you mention the fast one but not what the slow one is) and 2), give a few sentences about the physical meaning of the link between absorption of radiation (up to which height?) and the global precip response. I appreciate you give several references, but seeing how central this information is to the whole paper, it increases readability of the piece to spend a few more sentences (and an equation or two).

Following up on the fast response and absorbed radiation relationship (see also eq 1 of O’Gorman 2021 which you cite): I am wondering about the direction of causality. Is it really such that changes in absorbed radiation determine precip, and not vice versa? After all, precipitation (and evaporation) changes may be related to changes in water vapour content or clouds, which may feed back on radiation budget. So it would be good to clarify whether the relationship is (largely) a causal one, or whether it should be seen as merely a diagnostic relationship. If the latter is the case, then of course it can still be used for e.g. the analysis in sect. 4.3, but I would then suggest to me more careful with statements such as “precipitation changes as a function of injection rate can be understood based on the absorbed radiation” (line 279), a formulation which to me suggests causality.

sect 4.2 ff: you focus strongly on the fast precip response. Obviously this is an important quantity, especially in scenarios where GMST changes and hence the slow response are eliminated by SAI. However, since other scenarios are conceivable (e.g., keeping GMST change at 1.5 degrees), it would be quite nice to know how the fast response compares with the slow response. This can be inferred from S6, but is not discussed much. Maybe summarise the results in an equation like “P = a C + b S + c T” where P is the precip change, C the radiative forcing from CO2 (GHG), S the forcing from SAI, and T the GMST change, and a,b,c, are the fit parameters that arise from this study, though admittedly, at least b will suffer from nonlinearities (fig. 4).

Fig 4a: you state in the main text that the slope differs little among models. However, could you comment also on whether the slope is the (approximately) same for SAI and CO2? at least for MPI-ESM and SALSA, this seems not certain to me from the plot.

line 373: is there any clear physical reason why GMST increases in two models despite radiative balance being closed?

line 395, fig 6d: You suggest that in EC-earth, the correction hydrological sensitivity (i.e. effect of residual GMST change) “slightly overadjusts” precip estimates. This seems rather optimistic. In fact, the error hardly shrinks, of even becomes worse, in some scenarios in EC-earth, even if the correction works nicely in CESM. So it seems to me that in EC-earth there is stuff going on that is not easily captured by your method… could you comment?

line 428: why is there the local radiative forcing peak at ≈50ºN and S? if I understand correctly, then the reference, Laasko 2020 sect. 3.1.2 explains nicely why the forcing effect is lower at the poles, but not why there is a local maximum between the subtropics and the poles.

Fig 7-10: how linear are the responses (within each model combination) at the local level? Is it possible, like you did on the global level, to understand the local response as approximately the sum of the slow response, fast GHG response and fast SAI response?

≈ line 440, fig 8, CESM-SALSA, SRM20 and (in supplement S8) CESM-M7 SRM50: is there an AMOC response in the north Atlantic?

Regarding precip changes (fig 10): For impact modellers, maybe Precip-Evaporation would also be meaningful. Not sure if this is inside the scope of the paper. However, often SAI scenarios reduce not just precip but also evaporation, so that the overall effect on water availability is much less than precip changes suggest.

Last paragraph: Quantify “significant uncertainties”. Is the inter-model discrepancy for the most relevant quantities (e.g., global precip change) of the order of 10% of the signal, or 50%, or is there even disagreement of the sign?

Minor Clarifications

fig. 4: Legend: the dash in “-SALSA” and “-M7” look like a minus, which is a little misleading. maybe write “for SALSA” or “(SALSA)” ? Also, please make more clear in the figure caption that in plots b-d, the symbols and the solid lines are independent, i.e., the solid line is the sum of the other lines (not: “total”) whereas the symbols are the actual total (modelled) impacts. It becomes clear from the main text, but the figure itself is not as clear as it could be due to the shortness of the caption.

Supplement figs S1, S2, S3, S5, maybe other equivalent ones: Please add unit to the y-axis label of plot a.

Fig. 6b: add “estimated” to the y-axis label (equiv to “modelled” in plot d). Clarify in caption that “hydrological sensitivity” (which is a quite general-sounding word with a much more specific meaning), refers to the effect of residual GMST change on precip.

Section 5: you write a rather substantial summary of your findings. This could be further supported by adding references back to the corresponding sections and figures so that one can quickly (re)check the corresponding results in detail.

Typos, grammar etc
line 56: Changes … has -> have

line 60: change translate -> translates

line 254: sentence structure is a bit awkward

line 274-275: check sentence structure (missing “FOR larger injection…”?)

line 291ff: Awkward sentence. Comma missing after “figure shows”?

line 316: In case -> In this case?

line 490: issensitive -> is sensitive
Citation: https://doi.org/10.5194/egusphere-2023-2520-RC2
- AC2: 'Reply on RC2', Anton Laakso, 09 Feb 2024
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-2520/egusphere-2023-2520-AC2-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-2520-AC2

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2023-2520', Peter Irvine, 07 Dec 2023

General comments
The authors evaluate the precipitation response to stratospheric aerosol injection (SAI) geoengineering, considering earth system model and aerosol microsphysical uncertainty. Prescribed aerosol fields were generated in an ESM with either a sectional or modal aerosol module, producing quite different aerosol properties and hence radiative forcings. These were fed into 3 different ESMs which simulated a range of combinations of CO2 and SAI injections. The fast, forcing-driven hydrological response was found to be quite different for the different aerosol modules as the modal module produced fewer, larger particles which absorbed more LW radiation. Despite being driven by the same aerosol field, the ESMs produced quite different radiative, temperature and precipitation responses. However, the largest differences in many respects arose from the microphysical representation. The study makes a detailed analysis of the various factors that shape the precipitation response to SAI, making clear that microphysical uncertainties are important.
This paper will make a substantial contribution to the literature, is generally well-written and has generally good quality analysis, and so I recommend that it be published after making relatively modest changes, outlined below.
The paper is generally clear and well-written, but the argument was a little hard to follow in places as the paper jumped back and forth between radiative forcing and precipitation several times. For example, section 4.4 is titled “simulated precipitation response…” but the opening page is about the reasons for a radiative mismatch. The authors may consider revising the order of analysis.
The figures and analysis are generally very good, but in places the analytical choices made things a little difficult to follow, e.g., Figure 6 was particularly challenging. I’ve made a series of suggestions for improvement in the specific comments below.
I was left not quite knowing the answer to a question that I think could help increase the impact of this study and I think that with a little work could be easily answered. There is a factor of ~2 difference in the SO2 amount needed to achieve the same cooling for the sectional and modal aerosol modules. This made me wonder: is the residual precipitation, or just fast precipitation, difference ~2x larger as well? Or does the fast effect of CO2 dominate this residual? More generally, could the authors comments on the relative scale of the precipitation differences compared to this injection amount? RMSE difference might be a simple metric that could be calculated to test this. Some take-away claim that relates these 2 key elements would make the paper more memorable and useful to the community.
Specific comments
L14 – reduction relative to what?
L16 – “rather negatively correlated” – why not just negatively correlated? And could you clarify what is meant by “absorbed radiation” here? Is that a new finding or a widely established result that you are referencing?
L30 – relative to what?
L31-34 – review phrasing.
L43 – clarify whether the same 2 aerosol modules were used in the 3 different models.
L50 – might be nice to indicate roughly the fractional changes here.
L23-50 – Might be worth indicating which aerosol scheme performs better at reproducing observed volcanic response if that can be determined, i.e., is the SALSA sectional model better but more expensive and M7 the poor-man’s alternative?
L59-60 – Does this apply in the same way to stratospheric heating as it does to tropospheric? Is stratospheric heating as effective as tropospheric heating at suppressing precipitation? If the absorption occurred up in the mesosphere, I imagine it would have little effect on the hydrological cycle.
L64-66 – perhaps note T-driven intensification under GHG case?
L75 – formatting of citations.
L78 – in the consequent precipitation responses.
L103 – add resolution in degrees.
L149 – from a preindustrial baseline with GHG and SAI perturbations applied?
Figure 1 – Great figure! Small suggestion: 6x climate responses instead of impacts.
L162 – logarithmic fit
Figure 2 – Another great figure. Wondered if it might make sense to use the shape to match models, e.g., diamonds = CESM. This might help the colorblind to follow along. Looks like that was done in Figure 3, but I’d suggest adding the shapes to the legend or caption.
L206 – will have changed when it does settle down?
L205 – 213 – a little repetitive.
L245-249 - phrasing a little unclear.
Figure 4 – Is it best to compare injection mass for Salsa and M7 directly in this way? I found myself a little confused until I remembered that 50Tg in Salsa has a much greater cooling effect than in M7. Perhaps some additional text or analysis could clarify this, e.g., normalizing the fast effect by the expected cooling magnitude or plotting against an x-axis that shows temperature or RF?
L295-296 – Would this non-linearity disappear if the axis was RF instead?
L347 – less precipitation = a greater reduction in precipitation relative to the baseline?
L350 – link back to earlier claim on reduced SO2 for same RF in EC-earth?
Figure 6 – Is this the best way to get this information across? I’m very confused by some of the analytical choices and by how complex it is. Why aren’t the points falling on the precise CO2 ppm values used before? Can the analysis be flipped so that they do? More information needed on c, to clarify modelled pairs. Panel d seems like it could have been a whole multi-panel figure of its own. I also wonder if a pure temperature adjustment is the best choice, couldn’t you also scale up or down the fast effect of SAI by the fractional change in cooling that’s needed? Presumably that would give a better fit.
L355 – conversely? should that be Additionally?
355-360 – this suggests switching axes on Figure 6, as CO2 is the dependent variable.
4.4 – Given the first page is about the radiative mismatch, should this be 2 sub-sections? And should the radiative discussion come here or earlier? This might help with the flow of the article.
L392 – global mean precipitation is more positive?
L393 – here you are referring to the effect after the fast effect, whereas in some studies it is meant to include the total effect.
L398-400 – I think making the correction I suggested and noting that the forcing mismatch produced this precipitation mismatch might lead to a more useful conclusion here.
L403-420 – Isn’t a big driver of the overcooling / residual warming seen in many stratospheric aerosol geoengineering experiments the distribution of aerosols? Might be useful to refer to that distribution here and remind the reader that it’s the same in each model (I may have forgotten myself by this point).
Figure 7 – maybe a note on how these pairings were chosen. It might be useful to extend the y axis and add a global mean temperature residual value to the legend.
Figure 8 – A bit difficult to read, would adding figure wide column and row labels make it easier to parse? You might also consider rearranging so that SALSA is as one block, M7 as another.
Figure 9 – missing labels. Panel a is quite difficult to read, some for previous figure. Is there another way to show this?
Figure 10, same comment as 8.
489-494 – not particularly clear or particularly logical flow at the end of this paragraph, consider revising.
504-505 – compared to what? Is the comparison to the baseline the most relevant? Should it be to the 500 ppm case? Given the amount of SO2 injected scales with CO2, this difference in injection amount should modulate that total precipitation response, which as a consequence shifts the net result.
L513-514 – See my earlier comment about making a full adjustment, i.e., what would have occurred if the correct amount had been chosen to keep temperature constant, rather than just the temperature adjustment (which excludes the change in fast forcing effect).
L495-513 – Here or elsewhere some comment on the relative scale of the precipitation differences compared to the required injection amounts would be useful. M7 suggests ~2x greater sulphate required, is the gross or net precipitation difference 2x greater too?
518 – more negative?
530 – consistently more negative?
538 – perhaps remind reader that they faced the same change in aerosol optical properties
543-547 – a little hard to follow.

Citation: https://doi.org/10.5194/egusphere-2023-2520-RC1
- AC1: 'Reply on RC1', Anton Laakso, 09 Feb 2024
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-2520/egusphere-2023-2520-AC1-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-2520-AC1
RC2:
'Comment on egusphere-2023-2520', Anonymous Referee #2, 20 Dec 2023
The paper provides an extensive analysis of inter-model difference in the global (and regional) precipitation response to Stratospheric Aerosol Injection. On the whole, this is a sound piece of work, carefully analysed and well written, with clear plots.
A not too major criticism is that the paper is rather technical, and readability and possibly usability by a larger set of readers could be improved by adding some clarifications and physical interpretation here and there. I hope my comments can help. In addition, there is a handful of minor improvement points regarding things like figure captions and labels, listed below.

Interpretation and readability

Overall aims: Maybe the overall aim(s) could be stated in a small number of clear research questions at the end of section 1. Currently line 81 states: “We investigate how these impacts depend on the injection rate and the aerosol microphysics model”, which is relatively vague and mixes physics questions (how does precipitation change under different SAI intensities) with modelling questions (is there model uncertainty). Unless the main focus is strictly the model uncertainty part, the paper, which is now relatively technical, may profit here and there from a bit more physical interpretation.

Structure, especially Section 4: It would help the reader to get a short hint at the beginning of sect 4 what the subsequent pieces of analysis are meant to do and how they relate to each other and the overall aim / research questions of the paper. For example, it helps to know before sect. 4.3 that you first estimate the precip change based on the radiation diagnostics and then will compare the overall change to the actual model results in 4.4.

Fast response and absorbed radiation, line 56 ff. The paragraph could be clearer. Line 57 “Further precip change” -> further with respect to what? Line 59: rather than saying “Any change in X translates to a change in Y”, it is clearer to say e.g. “Any increase in X translates to a decrease in Y”, to immediately give the direction of change. The whole sentence seems unnecessarily long-winded. Most importantly, since some of the readers may not be experts in hydrological cycle but e.g. in SRM or impact modelling, it would be helpful to explain in a bit more detail 1) what fast and slow precip responses are (you mention the fast one but not what the slow one is) and 2), give a few sentences about the physical meaning of the link between absorption of radiation (up to which height?) and the global precip response. I appreciate you give several references, but seeing how central this information is to the whole paper, it increases readability of the piece to spend a few more sentences (and an equation or two).

Following up on the fast response and absorbed radiation relationship (see also eq 1 of O’Gorman 2021 which you cite): I am wondering about the direction of causality. Is it really such that changes in absorbed radiation determine precip, and not vice versa? After all, precipitation (and evaporation) changes may be related to changes in water vapour content or clouds, which may feed back on radiation budget. So it would be good to clarify whether the relationship is (largely) a causal one, or whether it should be seen as merely a diagnostic relationship. If the latter is the case, then of course it can still be used for e.g. the analysis in sect. 4.3, but I would then suggest to me more careful with statements such as “precipitation changes as a function of injection rate can be understood based on the absorbed radiation” (line 279), a formulation which to me suggests causality.

sect 4.2 ff: you focus strongly on the fast precip response. Obviously this is an important quantity, especially in scenarios where GMST changes and hence the slow response are eliminated by SAI. However, since other scenarios are conceivable (e.g., keeping GMST change at 1.5 degrees), it would be quite nice to know how the fast response compares with the slow response. This can be inferred from S6, but is not discussed much. Maybe summarise the results in an equation like “P = a C + b S + c T” where P is the precip change, C the radiative forcing from CO2 (GHG), S the forcing from SAI, and T the GMST change, and a,b,c, are the fit parameters that arise from this study, though admittedly, at least b will suffer from nonlinearities (fig. 4).

Fig 4a: you state in the main text that the slope differs little among models. However, could you comment also on whether the slope is the (approximately) same for SAI and CO2? at least for MPI-ESM and SALSA, this seems not certain to me from the plot.

line 373: is there any clear physical reason why GMST increases in two models despite radiative balance being closed?

line 395, fig 6d: You suggest that in EC-earth, the correction hydrological sensitivity (i.e. effect of residual GMST change) “slightly overadjusts” precip estimates. This seems rather optimistic. In fact, the error hardly shrinks, of even becomes worse, in some scenarios in EC-earth, even if the correction works nicely in CESM. So it seems to me that in EC-earth there is stuff going on that is not easily captured by your method… could you comment?

line 428: why is there the local radiative forcing peak at ≈50ºN and S? if I understand correctly, then the reference, Laasko 2020 sect. 3.1.2 explains nicely why the forcing effect is lower at the poles, but not why there is a local maximum between the subtropics and the poles.

Fig 7-10: how linear are the responses (within each model combination) at the local level? Is it possible, like you did on the global level, to understand the local response as approximately the sum of the slow response, fast GHG response and fast SAI response?

≈ line 440, fig 8, CESM-SALSA, SRM20 and (in supplement S8) CESM-M7 SRM50: is there an AMOC response in the north Atlantic?

Regarding precip changes (fig 10): For impact modellers, maybe Precip-Evaporation would also be meaningful. Not sure if this is inside the scope of the paper. However, often SAI scenarios reduce not just precip but also evaporation, so that the overall effect on water availability is much less than precip changes suggest.

Last paragraph: Quantify “significant uncertainties”. Is the inter-model discrepancy for the most relevant quantities (e.g., global precip change) of the order of 10% of the signal, or 50%, or is there even disagreement of the sign?

Minor Clarifications

fig. 4: Legend: the dash in “-SALSA” and “-M7” look like a minus, which is a little misleading. maybe write “for SALSA” or “(SALSA)” ? Also, please make more clear in the figure caption that in plots b-d, the symbols and the solid lines are independent, i.e., the solid line is the sum of the other lines (not: “total”) whereas the symbols are the actual total (modelled) impacts. It becomes clear from the main text, but the figure itself is not as clear as it could be due to the shortness of the caption.

Supplement figs S1, S2, S3, S5, maybe other equivalent ones: Please add unit to the y-axis label of plot a.

Fig. 6b: add “estimated” to the y-axis label (equiv to “modelled” in plot d). Clarify in caption that “hydrological sensitivity” (which is a quite general-sounding word with a much more specific meaning), refers to the effect of residual GMST change on precip.

Section 5: you write a rather substantial summary of your findings. This could be further supported by adding references back to the corresponding sections and figures so that one can quickly (re)check the corresponding results in detail.

Typos, grammar etc
line 56: Changes … has -> have

line 60: change translate -> translates

line 254: sentence structure is a bit awkward

line 274-275: check sentence structure (missing “FOR larger injection…”?)

line 291ff: Awkward sentence. Comma missing after “figure shows”?

line 316: In case -> In this case?

line 490: issensitive -> is sensitive
Citation: https://doi.org/10.5194/egusphere-2023-2520-RC2
- AC2: 'Reply on RC2', Anton Laakso, 09 Feb 2024
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-2520/egusphere-2023-2520-AC2-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-2520-AC2

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

ED: Publish subject to minor revisions (review by editor) (15 Feb 2024) by Michel Crucifix

AR by Anton Laakso on behalf of the Authors (16 Feb 2024) Author's response Author's tracked changes Manuscript

ED: Publish subject to technical corrections (27 Feb 2024) by Michel Crucifix

AR by Anton Laakso on behalf of the Authors (07 Mar 2024) Author's response Manuscript

Journal article(s) based on this preprint

24 Apr 2024

| Highlight paper

Dependency of the impacts of geoengineering on the stratospheric sulfur injection strategy – Part 2: How changes in the hydrological cycle depend on the injection rate and model used

Anton Laakso, Daniele Visioni, Ulrike Niemeier, Simone Tilmes, and Harri Kokkola

Earth Syst. Dynam., 15, 405–427, https://doi.org/10.5194/esd-15-405-2024,https://doi.org/10.5194/esd-15-405-2024, 2024

Short summary Chief editor

Anton Laakso, Daniele Visioni, Ulrike Niemeier, Simone Tilmes, and Harri Kokkola

Supplement

https://doi.org/10.5194/egusphere-2023-2520-supplement

Anton Laakso, Daniele Visioni, Ulrike Niemeier, Simone Tilmes, and Harri Kokkola

Viewed

Total article views: 280 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
199	61	20	280	26	10	9

HTML: 199
PDF: 61
XML: 20
Total: 280
Supplement: 26
BibTeX: 10
EndNote: 9

Views and downloads (calculated since 15 Nov 2023)

Month	HTML	PDF	XML	Total
Nov 2023	26	7	0	33
Dec 2023	73	14	8	95
Jan 2024	28	13	2	43
Feb 2024	41	12	5	58
Mar 2024	25	9	1	35
Apr 2024	6	6	4	16
May 2024	0
Jun 2024	0
Jul 2024	0
Aug 2024	0

Cumulative views and downloads (calculated since 15 Nov 2023)

Month	HTML	PDF	XML	Total
Nov 2023	26	7	0	33
Dec 2023	73	14	8	95
Jan 2024	28	13	2	43
Feb 2024	41	12	5	58
Mar 2024	25	9	1	35
Apr 2024	6	6	4	16
May 2024	0
Jun 2024	0
Jul 2024	0
Aug 2024	0

Viewed (geographical distribution)

Total article views: 276 (including HTML, PDF, and XML) Thereof 276 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 01 Sep 2024

Download

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Preprint (63902 KB)
Metadata XML

This is the second of two papers that study how the used model and injection strategy affect the impacts of stratospheric sulfur injections. The paper highlights the role of microphysics and of injection rate on the results of stratospheric sulfur interventions. The findings emphasize the need for precise simulation of aerosol microphysics to accurately estimate the climatic effects of stratospheric sulfur interventions.

Short summary

This study is the second in a two-part series in which we explore the dependency of the impacts of stratospheric sulfur injections on both the model employed and the strategy of injection utilized. The study uncovers uncertainties associated with these techniques to cool climate, highlighting how the simulated climate impacts are dependent on both the selected model and the magnitude of the injections. We also show that estimating precipitation impacts of aerosol injection is a complex task.


Total:	0
HTML:	0
PDF:	0
XML:	0