Summary

The authors use a set of CESM1 simulations (1980-2020 transient runs; 10 members for each experiment) to isolate the impacts of fossil fuel/industrial aerosol emissions from China+India (EastFF; where emissions have increased) and North America+Europe (WestFF; where emissions have decreased) on the tropical Pacific and Pacific Decadal Variations (PDV).  There are interesting differences between the two experiments, e.g., different North Pacific SST responses; EastFF drives a more El Nino-like SST pattern whereas WestFF drives a more CP-type El Nino SST pattern, etc.  The authors continue onwards and discuss the dynamical mechanisms.

Overall, the paper is interesting and adds to our understanding of the potential climate impacts associated with regional changes in aerosol emissions.

Comments

L23. “The competing effects of the heterogeneously distributed regional aerosol forcings are expected to be changed in the near future, which is likely to introduce opposite and more profound impacts of aerosol forcing on the Pacific multi-decadal changes.”  This is unclear.  If emissions decrease, this will reinforce the WestFF pattern showed here, but potentially flip the EastFF pattern showed here?  EastFF changes are likely complex, as emissions from China are likely to decrease (and have been decreasing, see below) whereas emissions from India may not, i.e., the east Asian dipole AOD pattern in Samset et al. 2019 (already referenced).  Moreover, the components of EastFF and WestFF aerosol is different, as EastFF has a larger aerosol absorption component that has unique impacts on circulation, precipitation, etc. (e.g., https://www.nature.com/articles/nature11097).

Linearity.  The authors show WestFF+EastFF relative to FF.  The FF AOD changes are well captured by WestFF+EastFF (Figure 1).  However, this is less true for the responses (e.g., Figure 2).  In particular, Figure 2 shows a much larger tropical Pacific response (the focus of this paper) under WestFF+EastFF as compared to FF.  Why might this be?  Is it related to aerosol outside these two regions?  I suppose this is unlikely given Figure 1.  Or is it related to differences between WestFF+EastFF versus West+EastFF (i.e., the latter is the signal from both perturbations simultaneously, which may be a better estimate for FF?).  I understand the authors have not performed the West+EastFF simulation, but some discussion on this matter is warranted.

Related to the comments above is L93. “One note here is that the climate changes in response to FF do not necessarily equal the simple combination of that in response to EastFF and WestFF because the FF results also contain anthropogenic aerosol forcings originated from other regions not covered by EastFF and WestFF (e.g., Africa and Arabian Peninsula, Fig. 1d). More details of the regional AA single forcing large ensemble simulations are described in Diao et al. (2021).”  This seems to suggest the differences in Figure 2c and Figure 2d are due to the very small differences in AOD from Figure 1d?  Could there not be nonlinearities?  Again, there are some sizable differences between Figure 2c and Figure 2d.

Biomass burning aerosols (BMB).  Are they not important, i.e., in driving tropical Pacific SST variations?  There is a very small discussion (e.g., Figure 6).  A recent paper (https://www.nature.com/articles/s41612-024-00602-8) focused on the AMOC, using the same CESM1 Large Ensemble, and showed that BMB aerosols drove significant changes in the AMOC (that were largely out of phase relative to AMOC variations under FF aerosols).

A caveat/limitation of this study is the use of a single model, which should be acknowledged and discussed.  The results presented here are no doubt model dependent.  The recently established Regional Aerosol Model Intercomparison Project (RAMIP) is one community effort designed to understand similar questions as addressed here (climate impacts of regional aerosol emissions changes) in a multi-model framework (https://gmd.copernicus.org/articles/16/4451/2023/gmd-16-4451-2023.html). 

Statistical significance and robustness of the ensemble mean signal.  This information is not provided in any figure.  We do not know what changes are significant, nor do we know the spread across individual realizations.  For example, maybe the multi-model mean response is driven by 1 or 2 ensemble members?  How robust are the responses shown here?

Emissions. I assume CMIP5 emissions are used here?  And they are extended to 2020 using RCP4.5?  I just wonder about the similarities (or dissimilarities) between real-world changes in aerosol emissions and what is used to drive the model, particularly in the context of East Asia (largely China) emissions, where there are known disagreements.  For example, it is noted in a few places that EastFF is associated with “continuous cooling” (e.g., L140), which presumably means progressively more negative ERF and/or decreasing near-surface air temperatures regionally and/or globally?  Is this the case, or is there an inflection point where cooling transitions to warming (or perhaps the cooling levels off)?  If so, the EastFF forcing may be more complicated than is currently expressed.  Maybe this is not true in the modeling realm, but it might be true in the real-world, which certainly has impacts on one’s ability to make/attempt attribution (e.g., Figure 3). 

In a similar vein, additional details on the CESM1 model (specifically details relevant to this study) should probably be included.  For example, CESM1 contains a relatively large aerosol ERF (this ties into the comment above on the fact this study uses one model).

After Figure 2, plots generally only include EastFF and WestFF results (although Figure 6 throws in some BMB panels).  What about FF?  Do we not also want to compare EastFF and WestFF (or the linear sum of the two) to FF?  One is left wondering if these two aerosol signals (or their sum) resemble the total FF signal.  And more generally, what about the ALL forcing signal?  Some discussion is perhaps warranted, e.g., does the ALL signal in any way look like FF and/or EastFF+WestFF?  In other words, how important is the EastFF+WestFF signal relative to ALL/FF forcings.  For example, Figure 5 and 6 attempt to do this, and it seems clear the FF dynamical response is quite different than BMB, as well as EastFF and WestFF (Fig. 5).

Dynamical responses/teleconnections.  The focus here is on annual means, but atmospheric teleconnections tend to have strong seasonal variations.   Are these results (e.g., Fig. 5) largely boreal wintertime responses?  A related paper that also addresses aerosol changes and their impacts on atmospheric circulation/teleconnections (but focused on Pacific Coast precipitation), which should probably be cited, is here: https://iopscience.iop.org/article/10.1088/2752-5295/ac7d68/meta

Since the 1980s, the anthropogenic aerosols (AA) in the western hemisphere are reduced whereas those in the eastern hemisphere continue to increase. The study is to contrast the effect of regional AA from the western and eastern hemispheres in driving the Pacific climate change in the past four decades, using large ensemble regional AA forcing simulations. The analysis is straightforward and the manuscript is clearly written. I only have some minor comments.

1. Wang et al. (2024) show that the AA in China has been decreasing between 2007 and 2020. Is this also reflected in your experiment setup? Or is it assumed to have continuously been increasing? Uncertainties in the AA forcing make me question the validity of comparing regional AA simulations with observations. Caveats regarding the uncertainties in the AA forcing should be discussed.
2. Indicate the significance of the responses in the figures.
3. Section 3.2: I presume the EOF is applied to the ensemble-mean. What about applying the EOF to each ensemble, and then taking the mean? It is more comparable to contrast the ERA5 with each ensemble rather than with ensemble-mean. It’ll also be useful to see the spread of indices in Fig. 3e,f.
4. The authors explain the upper-level low anomaly over the extratropial North Pacific in EastFF as a result of a Rossby wave train excited from the enhanced convection over the central equatorial Pacific (Fig. 5a,c,e). The precipitation response in the tropical Pacific exhibits a comparable order of magnitude between EastFF and WestFF (Fig. 4c,d). In particular, both simulations show precipitation increase in the western tropical Pacific. I’d expect a similar Rossby wave pattern driven by the enhanced diabatic heating over the warm pool. However, the response of upper-level wave activity is much weaker in WestFF. The fact that the tropical precipitation response is similar between EastFF and WestFF makes me think that strong circulation responses in the extratropical Pacific in EastFF is due to local AA radiative forcing via modulating storm tracks. That said, I don’t follow the discussion in Section 3.4.

Review’s comments for the manuscript egusphere-2024-1920, entitled "Contrasting the roles of regional anthropogenic aerosols from the western and eastern Hemispheres in driving the 1980–2020 Pacific multi-decadal variations"

General comments

By using regional aerosol forcing large ensemble simulations based on CESM1 performed earlier, this study investigates impacts of increasing fossil fuel-related aerosol emission over Asia (EastFF) and the reduction in aerosol emission over North America and Europe (WestFF) on the Pacific circulations and SST changes since the 1980s. One major concern the reviewer has is that results presented in the current version of manuscript are lack of significant test. It is not clear whether responses to different regional aerosols forcing changes described in the study are statistically significant from the internal variability. In addition, there are some other comments that need to be addressed to improve the manuscript. Therefore, the paper needs a major revision before it can be considered for publication.

Major comments

1. It is not clear whether all results analysed are about annual changes. In data and method, there is no any information given.
2. All results presented in the current version of manuscript are lack of significant test. It is not clear whether responses described are statistically significant from the internal variability. This aspect needs to be improved through whole manuscript.
3. Many references cited in text are not in reference list. Please check them carefully.
4. The study is based on a set of single model simulations, some comments on this aspect would be helpful for readers.

Specific comments

1. Line 20. “an IPO-like SST pattern”. It is helpful if authors can clarify the phase.
2. Lines 23-25. The last sentence is very confusing. How the likely opposite responses to future aerosol emission changes are likely to introduce more profound impacts?
3. Lines 112-113. Are SST anomalies low frequency filtered? Please clarify.
4. Line 115. “global warming mode induced by the greenhouse gases (GHG)”. There are also other external forcings in model simulations.
5. Lines 195-197. Do authors really think that the discrepancy between FF response and sum of EastFF and WestFF responses is due to aerosol forcings outside two focused regions, which are small as suggested in Figure 1d?
6. Line 204. “SST anomalies”. See major comment 1 and specific comment 3.
7. Lines 254-256. These statements about changes in Walker circulation over the equatorial Atlantic are not convincing. Figure 4b shows anomalous ascent in the tropical Atlantic, being consistent with warming in the tropical Atlantic (Fig. 2b). However, it is hard to argue whether this change is statically significant or not since no such test is shown. See major comment 2. 
8. Lines 278-279. (Fig. 5c). Shall be (Fig. 5e)?
9. Line 288. “convection”, use another word.
10. Lines 298-303. These arguments about weak topical teleconnection are not very convincing. See major comment 2 and specific comment 7.
11. Lines 305-310. It is not clear what are the aims for showing BMB experiment in Fig. 6b, d, f. What wave trains authors describe here and they are similar to what? Fig. 6a, c, e are not refereed in text.
12. Lines 342-344. See major comment 2, and specific comments 7 and 10.
13. Lines 346-348. See specific comment 11.
14. Figure 5e, f and Figure 6e, f. Reverse colour scale for easy comparison with other panels.