Upper tropospheric ice sensitivity to sulfate geoengineering

Visioni, Daniele; Pitari, Giovanni; di Genova, Glauco; Tilmes, Simone; Cionni, Irene

doi:https://doi.org/10.5194/acp-18-14867-2018

Articles | Volume 18, issue 20

https://doi.org/10.5194/acp-18-14867-2018

© Author(s) 2018. This work is distributed under
the Creative Commons Attribution 4.0 License.

Special issue:

The Geoengineering Model Intercomparison Project (GeoMIP):...

https://doi.org/10.5194/acp-18-14867-2018

© Author(s) 2018. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 18, issue 20

Research article

|

17 Oct 2018

Research article |

| 17 Oct 2018

Upper tropospheric ice sensitivity to sulfate geoengineering

Daniele Visioni, Giovanni Pitari, Glauco di Genova, Simone Tilmes, and Irene Cionni

Download

Final revised paper (published on 17 Oct 2018)
Supplement to the final revised paper
Preprint (discussion started on 05 Feb 2018)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review of Visioni et al.', Anonymous Referee #1, 15 Mar 2018
- AC1: 'Response to referee #1', Daniele Visioni, 29 May 2018
RC2: 'Review', Anonymous Referee #2, 18 Apr 2018
- AC2: 'Response to referee #2', Daniele Visioni, 29 May 2018

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Daniele Visioni on behalf of the Authors (04 Jun 2018) Author's response Manuscript

ED: Referee Nomination & Report Request started (16 Jun 2018) by Ulrike Lohmann

RR by Anonymous Referee #2 (29 Jun 2018)

RR by Anonymous Referee #3 (04 Jul 2018)

Suggestions for revision or reasons for rejection

The manuscript by Visioni et al. aims at investigating the impact of geoengineering by stratospheric sulfur injections on upper tropospheric cirrus formations using the ULAQ-CCM. My comments below refer only to the revised version of the manuscript. I was not involved in the review process of the first manuscript version, but from the authors’ response to the reviewer comments I understand that they reran the model simulations due to an inadequate model setup in the first round.

Overall comment:

My overall impression is that the study indeed targets an interesting and important side effect of sulfate geoengineering, namely changes in upper tropospheric cirrus clouds, but that the chosen methodology and the performed analysis of the model results are flawed. I am afraid that the ULAQ-CCM is simply not the right tool for such an investigation (details see below). At least from what is written in the paper I am not convinced that the results are valid and provide substantial new insights. In the present form I cannot recommend this paper for publication in ACP, but I would give the authors a further chance to convince me, either by a more appropriate model set-up or an improved analysis.

General comments:

1) Model approach: Although I agree with the authors that it is a common approach to use sea surface temperatures and sea ice coverage from coupled ocean-atmosphere models in chemistry-climate models, I think that this approach is not appropriate for all kind of research questions. In the present case I have strong doubts that the change in sea surface temperatures as simulated by CCSM-CAM4 for G4 (no chemistry, simplified ice cloud scheme) is consistent with ULAQ-CCM’s G4 aerosol distributions and the change in ozone, clouds, etc. From Fig. R2_2 and R2_3 it is obvious that both models show substantial differences in the SG aerosol distribution and AOD. Therefore, using the SSTs from CCSM-CAM4 is as consistent as applying an artificial negative SST anomaly. In my opinion the only meaningful approach would be to use a coupled ocean-atmosphere model with aerosol scheme and interactive chemistry. Such models are in the meantime available, although probably quite expensive. Even with such models I would expect a large spread in the upper tropospheric cirrus response to SG due to uncertainties in parameterized processes.

2) Sect. 2.3: The description of the experimental set up is very confusing and imprecise. Do you use sea surface temperatures and sea ice coverage only or also land surface temperatures from CCSM-CAM? Is it a nudging approach or a prescribed boundary condition? There are fundamental differences. Nudging is a Newtonian relaxation technique which adds non-physical terms to the models’ equations to “pull” certain variables like temperatures towards observed values. Scientific inaccuracy is a general problem throughout the whole manuscript.

3) ULAQ-CCM performance wrt ice clouds: p10, l 2-4: “…we are considering thin ice clouds…” is this because ULAQ-CCM does not consider thick ice clouds or is this because the authors did a subsampling of the model output, i.e. selected only cases with thin cirrus? In the first case I have (again) severe doubts that ULAQ-CCM is the appropriate model for this study. In the second case the evaluation does not make sense, because the authors compare apples (thin cirrus) with oranges (all ice clouds) (apart from uncertainties in the MERRA and MODIS derived quantities).

4) The interpretation of the model results is very much focused on the “vertical temperature gradient – homogeneous ice formation” relationship. I am not sure whether this is a remnant of the first set of simulations in which heterogeneous ice formation had been erroneously switched off, but I miss an open discussion of other potential feedback effects. For example, changes in background cloudiness or large-scale circulation changes.

Specific comments:

- p1, l3: Why only “homogeneous freezing”? How can you exclude SG effects on heterogeneous freezing? Or is this a remnant from the first draft, for which heterogeneous freezing was switched off in the simulations?

- p1, l6: How do you define “longwave” radiation? Aerosols also absorb incoming radiation in the near IR.

- p3, l24: Again – why only “homogeneous freezing”?

- p4, l6: … help to explain…

- p4, l9-11: To design SG experiments which meet certain climate targets it is, at least in my view, crucial to consider all aspects and feedback processes in a self-consistent manner, which is not done in the present study. So this argument is counterproductive.

- Sect. 2.1: Are the CCSM-CAM4 simulations ensemble runs or only one realization for each scenario? And what is the climate sensitivity of the model?

- Table 1: equatorual -> equatorial; footnote 3 not used; do you use surface temperatures or SEA surface temperatures from CCSM-CAM4 in Base, G4 and G4K? That’s an important difference and needs clarification!

- p7, l17: This represents…

- p8, l5: What is an ice mass fraction? I guess you mean ice mass mixing ratio?

- p8, l23: What is Qext? Extinction efficiency coefficient?

- p8, l24: How is upper troposphere defined? Which altitude range?

- p8, l25: remove ij after rij

- p8, Fig.1: Do you use MERRA or MERRA-2? I assume MERRA-2. Please use a consistent nomenclature throughout the manuscript.

- p9, l5: Again - ice mass fraction?

- Fig. 3: What’s the purpose of showing this figure? I do not see the link to the present study, neither for model evaluation nor for the SG effects.

- Fig. 3: Add explanation of dashed and dash-dotted lines in panel b) to the caption.

- Fig. 4c): Again - what’s the purpose of showing this figure?

- Fig. 3, 4 and 7: Which seasons are shown (tropopause)? My year has 4 seasons, but only there are only two lines.

- Table 2: caption: homogenous -> homogeneous, row 2: (ULAQ-CCM) missing bracket; row 5: (HET) missing

- Sect. 2.3: As mentioned above Sect. 2.3 needs clarification, especially with respect to the treatment of surface or sea surface temperatures, nudging or prescribed lower boundary condition.

- p14, l32: Does G4 assume 5 Tg(SO2)/yr or 8 Tg(SO2)/yr, as stated 2 lines above?

- p15, l3: (Fig. 4-5) should read (Fig. 5-6)

- Fig 5: Wouldn’t it make more sense to show temperature anomalies from ULAQ-CCM instead from CCSM-CAM4 as this is the basis for the study? As far as I understand surface temperatures over land are calculated by the ULAQ-CCM?

- Fig. 6 and following: Why do you show averages for 2030-2039 when the simulations cover 2020-2069 (at least according to Table 1)?

- Fig. 7a,b: A ∆T = 0 contour line would be helpful to clearly identify regions with positive and negative temperature changes. Alternatively, a better color scale. And I would prefer to see the temperature changes starting at the surface.

- Discussion of Fig. 7a,b: The difference in surface temperatures between G4 and G4K has effect on the outgoing longwave energy and therefore the IR absorption by the SG aerosols. Furthermore I would expect general changes in cloudiness which also affect emission of terrestrial radiation. These aspects are not at all mentioned in the study.

- p18, l16ff: The authors state here that vertical motions caused by synoptic scale disturbances and gravity waves dominate the updraft velocities. Furthermore, they state that in G4 the vertical updrafts are reduced due to a reduced vertical temperature gradient, but how about the impact of changes in the meridional temperature gradient and subsequent changes in zonal winds and gravity waves?

- p18, l23: I assume the authors mean SO4 in the particulate phase, not in the gas phase.

- Discussion of Fig. 9: First of all, I do not understand why the LW heating rates in Fig. 9b) have been calculated with temperatures fixed at Base values as written in the caption? Furthermore, the authors do not mention potential changes in adiabatic heating rates due to a change in the Brewer-Dobson-Circulation (“decreased wave activity and a consequent decrease in poleward mass fluxes”) and their effects on lower stratospheric temperature changes. What is meant by “tropospheric convective cooling”? And how do the authors explain the neg. LW anomalies in G4 above ~26 km?

- Fig. 7/8: Are the displayed changes in vertical velocities statistically significant? The ±1𝝈 range (for which scenario? Base?) in Fig. 8 seems to be pretty wide compare to the differences between Base and G4.

- Fig. 8: Why are the respective results for G4K not included?

- p23, l2-4: As mentioned above I think the authors have to provide a more thorough evaluation of the model performance with respect to cirrus clouds.

- Fig. 11, 12, 13, 14: Please include ±1𝝈 range as done in other figures.

- Discussion of Fig. 12: What is the reason for the double-peak-structure of the change in ice extinctions shown in panel a), i.e. a more pronounced decrease in ice extinctions at 11 and 13 km, and a less pronounced signal at 12 km. This feature is similar in G4 and G4K.

- Sect. 3.2.2: How are the changes in RF calculated? Online during the model simulations or offline using the identical radiative transfer code as in ULAQ-CCM? In the first sentence it is written “online”, but it is not clear to me how the authors in that case distinguish between ice and background clouds. Furthermore, I miss a discussion of cloud changes in general and how they affect SW and LW radiation.

- Fig. 16: Geoengineered case: In my opinion the arrow “ more planetary radiation to space (ice)” is a bit misleading, because in that case the SG aerosols will absorb more IR radiation, and consequently emit more IR by themselves (upward and downward). So the sketch is overly simplified.

Hide

RR by Anonymous Referee #4 (18 Jul 2018)

ED: Reconsider after major revisions (23 Jul 2018) by Ulrike Lohmann

AR by Daniele Visioni on behalf of the Authors (14 Sep 2018) Author's response Manuscript

ED: Publish as is (03 Oct 2018) by Ulrike Lohmann

AR by Daniele Visioni on behalf of the Authors (03 Oct 2018) Manuscript

Download

Article (10299 KB)
Full-text XML

Short summary

Many side effects of sulfate geoengineering have to be analyzed before the world can even consider deploying this method of solar radiation management. In particular, we show that ice clouds in the upper troposphere are modified by a sulfate injection, producing a change that (by allowing for more planetary radiation to escape to space) would produce a further cooling. This might be important when considering the necessary amount of sulfate that needs to be injected to achieve a certain target.