the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Future water storage changes over the Mediterranean, Middle East, and North Africa in response to global warming and stratospheric aerosol intervention
Abolfazl Rezaei
Khalil Karami
Simone Tilmes
John C. Moore
Abstract. Water storage plays a profound role in the lives of people across the Middle East and North Africa (MENA) as it is the most water stressed region worldwide. The lands around the Caspian and Mediterranean Seas are simulated to be very sensitive to future climate warming. Available water capacity depends on hydroclimate variables such as temperature and precipitation that will depend on socioeconomic pathways and changes in climate. This work explores changes in both the mean and extreme terrestrial water storage (TWS) under an unmitigated greenhouse gas (GHG) scenario (SSP5-8.5) and stratospheric aerosol intervention (SAI) designed to offset GHG-induced warming above 1.5 °C and compares both with historical period simulations. Both mean and extreme TWS are projected to significantly decrease under SSP5-8.5 over the domain, except for the Arabian Peninsula, particularly in the wetter lands around the Caspian and Mediterranean Seas. Relative to global warming, SAI partially ameliorates the decreased mean TWS in the wet regions while it has no significant effect on the increased TWS in drier lands. In the entire domain studied, the mean TWS is larger under SAI than pure greenhouse gas forcing, mainly due to the significant cooling, and in turn, a substantial decrease of evapotranspiration under SAI relative to SSP5-8.5. Changes in extreme water storage excursions under global warming are reduced by SAI. Extreme TWS under both future climate scenarios are larger than throughout the historical period across Iran, Iraq, and the Arabian Peninsula, but the response of the more continental eastern North Africa hyper-arid climate is different from the neighboring dry lands.
- Preprint
(1521 KB) - Metadata XML
-
Supplement
(1515 KB) - BibTeX
- EndNote
Abolfazl Rezaei et al.
Status: open (until 04 Oct 2023)
-
RC1: 'Comment on egusphere-2023-1654', Anonymous Referee #1, 30 Aug 2023
reply
General comments:
Rezaei et al. use climate model simulations to investigate the impacts of stratospheric aerosol injection (SAI) on the Middle East and North African (MENA) region. The study looks specifically at total water storage, and associated hydrology variables, in model simulations with SAI and climate change. MENA is an understudied region in the context of SAI, with important potential impacts on the water cycle, so this paper is a welcome addition to the literature. The paper overall needs reorganization and edits to the text for clarity, as well as modifications to the figures and explanation of statistical methods to better communicate the results. General comments are listed first, and further specific comments below.
- I enjoyed reading about the study area in section 2.1 (in particular, the first two paragraphs). It would be informative to return to this context in the discussion section and talk about how the results in this paper might impact the regional climate more broadly.
- It would be helpful to give additional context for why these specific CESM simulations were used here (e.g., instead of the large ensemble simulations like GLENS/ARISE). I think a novelty of Tilmes et al. 2020 was the overshoot scenario but I don’t believe those simulations were used here. I suggest adding some additional text in section 2.2 for context on the model simulations.
- Generally, the paper does a good job specifying which comparison is being made (e.g., SSP relative to historical or SAI relative to historical) but there are some additional places to clarify the text in the results and discussion sections, which will make these sections easier to follow and interpret. I have some specific suggestions in the comments below.
- The discussion section currently recaps/repeats many of the results (e.g., Lines 485-507). I suggest focusing on interpreting the results, highlighting particularly interesting results, and connecting with previous studies. Some of this is already present in the discussion section (e.g., Lines 438-442, 475-483, 509-518), so I think the section just needs some edits to move results into the results section and focus the discussion.
Specific comments:
Lines 68-72: This paragraph seems out of place. Suggest moving this down to where projected future changes in the Mediterranean are discussed (e.g., Line 96).
Lines 74-84: Moving this paragraph down to after the discussion on climate change impacts (e.g., Line 108) would then transition to an introduction to SRM and associated impacts.
Line 90: Are there any more recent modeling studies (e.g., CMIP5 or CMIP6) that discuss projected changes in the MENA region? If not, it is worth pointing that out.
Line 102: An explanation of “soil moisture z-scores” is needed here.
Line 162: This is covered in the introduction; suggest removing this sentence and moving the second sentence of this paragraph into the preceding paragraphs on the discussion of regional climate.
Line 174: In addition to defining “real evapotranspiration”, please add an explanation on “potential evapotranspiration” and how that is calculated since that is also listed in Table 1. Is real ET a model output and potential ET is calculated from model output?
Table 1: The caption mentions historical model output as the data source for this table – please add this to the text as well (e.g., Line 172). It might also make sense for Table 1 to come after the model is introduced in section 2.2.
Line 209: Please clarify the text here – what is meant by “in turn”?
Line 241: Was this correlation calculated or made by eye using the plots in Figure S2? And was the correlation tested for the other variable combinations? Calculating the correlations and reporting them in the paper would make for a stronger justification of the MLR model inputs.
Line 246: How often / how many outliers were removed using this method?
Line 254: Please add more details on the method here – specifically what is meant by “independent variable-order average over average contributions…” and “impacts adjusted for other regressors”.
Figure 2: It is difficult to see the trends in the anomalies with the strong seasonal cycle in certain regions. Suggest removing the seasonal cycle and/or some other method of filtering out noise here (e.g., running yearly means).
Lines 282-295: Please use the region labels (e.g., R1-R6) in the text here to ease the interpretation of Figure 3.
Line 292: Clarify “Mean TWS” here – is that the temporal mean, ensemble mean, spatial mean, or some combination?
Figure 3 (and others): For the difference plots (e.g., panels b-d), I recommend choosing a different color scale with a clear divergence at 0. With this yellow/blue color scale it is difficult to discern positive vs. negative regions of change. Same comment for Figures S2, S4. The color scale in Figure S1 works better for difference plots.
Lines 304-311: Use percentage values instead of absolute kg/m^2 changes in the text here to match the black labels in Figure 4. Or use absolute labels in Figure 4, which would match the y-axis.
Figure 4: In addition to the partial reversals (R1, R3, R4) and the overcompensation (R2), SAI also has an amplifying effect in R5 and a slight overcompensation in R6 – it is worth noting these responses in the text (even if to say they are not significant).
Figure 4: Why are there three p-values shown at the bottom of each panel? I assume two of the values denote the significance of the changes in SSP and SAI relative to historical, but what does the other value represent? Please clarify in the figure caption. Same comment for Figures S3 and S5.
Line 330: Similar question to Line 241 – was significance calculated here or by eye? Why do non-overlapping curves imply significance?
Lines 344-360: Please use the region labels R1-R6 in text here to ease comparison to Table 2.
Line 346 and following: Please clarify “decreases the TWS extremes” – does this mean a decrease in positive extremes (i.e., fewer wetter conditions) or negative extremes (fewer drier conditions) or both?
Lines 378-380: Please clarify here whether this is referring to the most important variable under SAI or SSP (or both).
Lines 386-389: Please clarify what is meant by “due to evapotranspiration” if this is looking only at temperature and precipitation (“with just temperature and precipitation as independent variables”). Are there results that look at subsets of these three variables and are they included somewhere?
Line 399: Please include the specific variance explained values for the MLR models somewhere in the text or figures (e.g., the bars of Figure 6-7).
Figures 6-7: Here, or perhaps in the methods section, please provide some context for the importance values (y-axis). Is this unitless, and if so, should the individual variable contributions total to 1 if all the appropriate variables were sampled? Are interactions considered?
Lines 444-457: Most of this paragraph should go in the results section, as the supplemental figures have not yet been discussed. The last sentence gets to a comparison with other studies which is appropriate for the discussion section and can be merged with another paragraph.
Line 446: Please specify which simulation “The TWS decreasing patterns” refers to.
Line 461: Related to vegetation, it is worth discussing the competing impacts of high CO2 and less solar radiation in the SAI scenario. These impacts could also be contributing to the overall ET, soil moisture, and TWS responses. The regions discussed here have varying amounts of vegetation and that could be contributing to the range of regional responses.
Figure S3: I thought the middle row of this plot (TWS) would be same as Figure 4, but it appears to be different. What is plotted here and what is the difference with Figure 4?
Figure S4: For the middle row (temperature) difference plots, the color bar limits should be increased on both ends to better show the regional responses.
Data availability: Suggest providing some more information on how to access these specific CESM simulations via the ESGF website (e.g., Source ID, Experiment ID). Tilmes et al. 2020 also has a DOI for the SAI simulations which should be included if those experiments are not on ESGF: https://doi.org/10.26024/t49k-1016.
Technical corrections:
Line 39: Typo “Projected” should not be capitalized.
Lines 335-336: I think this should be “return levels” instead of “level returns”.
Lines 335-337: Should these sentences be combined?
Figure 5: Please add panel labels to the subplots and update caption to “(a to f)”.
Line 397: Typo “EV”
Line 467: Typo “EV”
Lines 520-522: I think “SAI” is missing after “with...and without” here.
Citation: https://doi.org/10.5194/egusphere-2023-1654-RC1 -
RC2: 'Comment on egusphere-2023-1654', Anonymous Referee #2, 24 Sep 2023
reply
Thank you for inviting me for reading this article. The authors evaluate terrestrial water storage under SSP5-8.5 and SSP5-8.5-SAI scenarios across the Middle East and North Africa. The results are useful for supporting aerosol intervention strategy against global warming and water resources management for Mediterranean, Middle East, and North Africa. I have some concerns about the methods and figures which may be helpful for improvement.
- Section 2.3. The authors calculate return periods from GEV distribution. However, GEV distribution is used to simulate maximum value in a certain period, instead of monthly values. The authors may give more details about how to apply GEV distribution. Did the authors calculate the annual maximum TWS values? In addition, authors may provide empirical probabilities and examine whether annual maximum TWS follows GEV distribution or other distributions.
- Line 211. The historical period is from 1985-2014, and future period is from 2071-2100. The authors do not analyze mid-21th century. The authors may explain why you do not analyze the full period from 1850-2100.
- Authors only select CESM2 for analysis. The authors may evaluate the performance CESM2 for historical climate over the study area to validate this model.
- Authors use the MLR model to predict TWS. Apart from potential ET, the actual ET is also correlated with temperature and precipitation. How to solve the collinearity between ET, temperature and precipitation?
- Authors remove outliers in the MLR model. This will artificially give better results. Please justify the removal of these values?
- The temporal autocorrelation is an important component in TWS evolution. Monthly TWS is not only impacted by concomitant precipitation and temperature, but also antecedent soil moisture and climatic variables. Authors may consider include climatic variable in previous months as predictors as well.
- Water storage include soil moisture, groundwater, snow, ice, and others. Figure S3 seems to indicate soil moisture is the dominant driver of TWS variations. It may be insightful for evaluate the relative contributions of other components in TWS.
- Figure S4 is important for interpreting current results. May consider to place this figure in main text.
- It may be useful to compare the results with previous evaluations (https://www.nature.com/articles/s41558-020-00972-w; Global terrestrial water storage and drought severity under climate change).
- Line 29, this sentence may be improved. May explain “more contnental” and “hyper-arid” climates? Specify what is different response ?
- Line 86, may place this paragraph earlier than the introduction of SRM, which is proposed to address climate change.
- Line 127. What is the regional consequence and hydrological cycle? May give more explanations
- Line 228 and Eq. (1). The authors give the equations for Xi = 0 in equation (2). It may be better to provide CDF when Xi (ξ)= 0 in Equation (1) as well. In addition, I think Eq.(1) is the CDF instead of PDF. It is better to clearly specify this.
- Line 272, this sentence may be improved.
- Figure 3. The colors for legend may be improved. For example, use two different hues to represent positive and negative values, and use white to repesent 0.
- Figure 5, it may be much better to show empirical probabilities of observed TWS and visually show the performance of GEV distribution.
- Figures 6 and 7. May add R-squared in the figures for better interpretation.
Citation: https://doi.org/10.5194/egusphere-2023-1654-RC2
Abolfazl Rezaei et al.
Abolfazl Rezaei et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 212 | 71 | 9 | 292 | 32 | 3 | 3 |
- HTML: 212
- PDF: 71
- XML: 9
- Total: 292
- Supplement: 32
- BibTeX: 3
- EndNote: 3
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1