the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Multi-model ensemble projection of global dust cycle by the end of 21st century using CMIP6 data
Yuan Zhao
Yang Cao
Jun Zhu
Chenguang Tian
Yuwen Chen
Yihan Hu
Weijie Fu
Xu Zhao
Abstract. As a natural aerosol with the largest emissions on land, dust has important impacts on atmospheric environment and climate systems. Both the emissions and transport of dust aerosols are tightly connected to meteorological conditions and as a result are confronted with strong modulations by the changing climate. Here, we project the changes of global dust emissions and loading by the end of the 21st century using an ensemble of model outputs from the Coupled Model Intercomparison Project version 6 (CMIP6) under four Shared Socioeconomic Pathways (SSPs). Based on the validations against site-level observations, we select 5 out of 10 models and estimate an ensemble global dust emission of 3311 Tg a−1 (1 Tg = 1012 g) at present day, in which 75 % is dry deposited and 25 % is wet deposited. Compared to 2005–2014, global dust emissions show varied responses with a reduction of 15.8 Tg a−1 under the SSP3-7.0 scenario but increased emissions up to 53.4 Tg a−1 under the SSP5-8.5 scenario at 2090–2099. For all scenarios, the most significant increase of dust emissions appears in North Africa (0.4 %–4.7 %) due to the combined effects of reduced relative humidity and precipitation but strengthened surface wind. In contrast, all scenarios show decreased emissions in central Asia (−0.6 % to −20 %) and Middle East (0 to −2.8 %) because of the increased precipitation but decreased wind speed regionally. The dust loading shows uniform increases over North Africa (1 %−12.5 %) and the downwind Atlantic following the increased emissions, but decreases over East Asia (−3.4 % to −15.2 %) and the downwind Pacific due to enhanced local precipitation that promotes wet deposition. As a result, global dust loading will increase by 2.1 %–9.3 % at the end of the 21st century under different climate scenarios, suggesting a likely strengthened radiative and climatic perturbations by dust aerosols in a warmer climate.
- Preprint
(2308 KB) -
Supplement
(4940 KB) - BibTeX
- EndNote
Yuan Zhao et al.
Status: final response (author comments only)
-
RC1: 'Comment on acp-2022-760', Anonymous Referee #1, 03 Mar 2023
The manuscript by Zhao et al. aims at estimating future changes in global distributions and the budget of mineral dust aerosol by making use of results from the CMIP6 model experiments. Although this is an interesting topic and the use of CMIP results is a good starting point, the study has significant weaknesses and I cannot recommend publication in ACP.
Major points:
- The authors point out that taking into account the effects of changes in vegetation cover on dust emissions may play an important role but using vegetation model scenario results make the dust aerosol trends more uncertain. Indeed such changes would have a significant impact on dust emissions. The authors do not clarify to which extent vegetation cover changes were considered in the CMIP model experiments that were used for this study, especially for the future scenarios. If they want to avoid uncertainties related to potential vegetation changes by considering only the effects of wind and precipitation changes, then only in regions that are vegetation-free in both the historical and future scenarios should be considered. (see e.g Mahowald et al., 2003, Woodward et al., 2005).
- The selection criteria that the authors use to decide what are the ‘good’ models to be used in this study is vague. Only surface concentration data of mineral dust are compared with station data at several locations, which are likely covering different time periods (not given here). For several sites none of the models reproduces the observed concentrations at certain seasons, which hints toward a fundamental problem of dust modelling at global scales. Given that mineral dust concentrations are highly variable in time and space including altitude of dust transport, to show that that the dust cycle is reproduced well by the models other data such as optical thickness from satellites are the AERONET network should be taken into account. It would also be interesting if the results on future trends would be different if all models would be considered.
- Another point regarding the selection of models used in this study: Rather than comparing multi-year average concentrations, the decadal temporal trends in dust aerosols simulated by the models would give a better indication for their suitability of predicting future changes (eg., Kok et al 2023).
- As shown in Figures S7-S11 in the supplemental material, the results for emission and deposition in selected individual models are differ greatly from each other (dust column loads should be added as additional figure). Ranges need to be given for all results. In those places where ranges are provided for scenario results (e.g. page 8 line 224 to 227) it is unclear what the range refers to – Standard deviations? Results from different Models?
- What is the reasoning behind focussing on relative humidity in addition to wind speed and precipitation as the main factors influencing dust trends? Relative humidity does not impact the dust cycle directly. Also in Figure 6 there appears to be a better correlation for precipitation than relative humidity anyway. If the relative humidity is supposed to represent drought condition one could instead use.
- In Figure 7 it hard to see the dotted areas which indicate significant changes in the figure. (I recommend to mask out areas with non-significant changes eg. by grey color). In any case, it appears that the focus regions downwind of East Asia do not contain significant changes and thus should not be highlighted in the paper.
- High resolution convection-resolving results show that wet convection driven dust emissions (cold pools) cannot be represented correctly, and that strengthening convective activity in future scenarios may enhance dust emissions in the southern Sahara in NH summer, but may lead to overestimating of low-level jet emissions (Garcia Carreras, 2021). It is questionable to which extent the results of coarse-resolved global models such as used in CMIP are suitable future change estimates, at least in regions that are strongly affected by convective activity.
Minor points:
- The CMIP experiments should be explained in some more detail. Why is only one ensemble member selected for each model? Why is the analysis limited to 10 years?
- It is unclear what information content the regional budget column in Table 3 has. Again as everywhere, ranges should be shown for the results in this table.
- Page 7, line 189: the sites offshore East Asia would certainly not be impacted from Middle Eastern dust sources
- Page 9 line 280 – “column concentration” should rather be named “column load”
- Some more discussion of the reasons of future precipitation and surface winds would be good.
References:
Garcia-Carreras, L., Marsham, J.H., Stratton, R.A. et al. Capturing convection essential for projections of climate change in African dust emission. npj Clim Atmos Sci 4, 44 (2021). https://doi.org/10.1038/s41612-021-00201-x
Kok, J.F., Storelvmo, T., Karydis, V.A. et al. Mineral dust aerosol impacts on global climate and climate change. Nat Rev Earth Environ 4, 71–86 (2023). https://doi.org/10.1038/s43017-022-00379-5
Mahowald, N. M., and Luo, C. (2003), A less dusty future? Geophys. Res. Lett., 30, 1903, doi:10.1029/2003GL017880, 17.
Woodward, S., Roberts, D. L., and Betts, R. A. (2005), A simulation of the effect of climate change–induced desertification on mineral dust aerosol, Geophys. Res. Lett., 32, L18810, doi:10.1029/2005GL023482.
Citation: https://doi.org/10.5194/acp-2022-760-RC1 -
RC2: 'Comments on acp-2022-760', Anonymous Referee #2, 22 Mar 2023
There is a growing concern about future dust change induced by global climate change and human activity. The study be Zhao et al. has presented the future changes in global dust cycles based on the five CMIP6 models. Ten models are first used for model evaluation, and five of these models with better performance are selected for the projection. They also investigate the change in surface wind and precipitation/relative humidity, the factors associated with the dust changes in the future. The conclusions can provide a good reference to the relevant community. Most of the manuscript is well written and clearly presented. I have some comments for the authors to consider. In particular, if possible, please provide more information on the uncertainty in the model simulation of dust cycle and discuss whether the changes are significantly large in the future.
Major comments:
- Line 366 (solid): The change of future vegetation change due to both climate change and human activity is not considered in this study, which may induce large uncertainty in the projection of future dust change. I suggest the vegetation change should be considered as well. If the impacts of vegetation change are not included, the authors should add some discussions on this.
- Introduction and Conclusions and discussion: Some studies on dust cycle using CMIP6 models should be included for discussion:
Checa-Garcia et al. (2021, https://acp.copernicus.org/articles/21/10295/2021/),
Le and Bae (2022, https://acp.copernicus.org/articles/22/5253/2022/),
Li and Wang (2022, https://acp.copernicus.org/articles/22/7843/2022/),
Maki et al. (2022, https://www.jstage.jst.go.jp/article/sola/18/0/18_2022-035/_article/-char/ja/),
Woodward et al. (2022, https://acp.copernicus.org/articles/22/14503/2022/).
- Selection of models for future projection: UKESM1-0-LL may produce too much dust emission compared to other models, according to Figure S7. I am wondering if it is reasonable to select UKESM1-0-LL.
- Uncertainty: As Table 3, the values of the range should be also provided for understanding the uncertainty. In addition, please provide the values for each models in supplemental files to compare different models.
- Line 259: significant: How to determine the regions with significant changes?
Specific comments:
Line 22: meteorological conditions: meteorological conditions can affect the vegetation cover, which further affects the dust emission. But the impacts of vegetation change on dust emission are not mentioned in the study. Please clarify.
Line 33: relative humidity: I think soil moisture is the variable more closely related to dust emission.
Line 35 (central Asia and Taklimakan): The regions are not correctly named. According to Figure 3a, central Asia and Taklimakan should be East Asia (at least Gobi Deserts are not located in central Asia); Middle East should be Middle East and central Asia.
Line 39: due to: I think it is “partly due to”.
Line 40: “As a result” should be “In total”?
Lines 65-67: First, according to Munktsetseg et al. (2016), it is more precise to say "soil moisture". Second, soil moisture alone does not control threshold friction velocity and dust emission intensity. Many factors including soil moisture determines them.
Line 114: All: it may be better to mention the date when the data are accessed to, as more data may come out later.
Lines 161-163: It is not clear to me. Please check.
Lines 176-177: It is hard for me to check in Fig. 2b. Perhaps also provide a table with these values in supplemental file.
Line 194: Figure 3 captions: Please also mention the latitudes and longitudes for the three regions.
Line 230: dominates: It is not clear to me. And it is hard for me to read this from Figure 4c.
Lines 232-233: But dust emission may be sensitive to precipitation change. Please clarify.
Lines 242-244: 18 regions: please check whether the numbers are correct.
Line 252: limited changes: It is not clear to me.
Lines 323-325: The sentence does not read clearly. Please revise
Lines 579-580: Not exactly red/blue (but light red & blue). I think the colors are too light to distinguish easily.
Figures 6 and 8: Could you make the zero lines bolder? It is not easy to see.
Table 2: u_*t and u_t are different. Please clarify.
Citation: https://doi.org/10.5194/acp-2022-760-RC2
Yuan Zhao et al.
Yuan Zhao et al.
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
266 | 88 | 8 | 362 | 28 | 1 | 2 |
- HTML: 266
- PDF: 88
- XML: 8
- Total: 362
- Supplement: 28
- BibTeX: 1
- EndNote: 2
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1