Round 1

Reviewer 1 Report

Review of Aerosol Effective Radiative Forcing in the Online Aerosol Coupled CAS-FGOALS-f3-L Climate Model by Wang et al.

In this paper, the authors present their results from simulations coupling the CAS-FGOALS-f3-L model to the SPRINTARS online aerosol module. They show that their model is able to simulate a plausible aerosol ERF, with the radiation and cloud components comparing favorably to other CMIP6 models.

This paper is an important part of evaluating the FGOALS contribution for CMIP6 and will be a useful reference. There are some modifications I would suggest, primarily to the terminology and some of the comparisons. The paper should also be more carefully proof-read before publication, as there are a large number of spelling and grammar errors (I have not tried to mark them all). Following this, I think this paper would be suitable for publications in Atmosphere.

Major points

The paper often references the direct radiative forcing (DRF) an the cloud radiative forcing (CRF). Is there any reason to prefer these terms over the effective radiative forcing from aerosol-radiation interactions (ERFari) and aerosol-cloud interactions (ERFaci), which are perhaps more common in recent work?

While the authors do a good job of comparing their model to other CMIP6 models, these models do not exist in a vacuum. There is a large body of literature looking at estimating the aerosol forcing, both from other model intercomparisons (e.g. AeroCom, CMIP5) and observations. A close comparison with CMIP6 may be desirable, but perhaps a better comparison would be with observational estimates? Some of these papers (e.g. Gryspeerdt et al., ACP, 2020) also make a clearer decomposition between liquid and ice clouds than older papers, which would help interpreting the results from the MRI model. For the overall uncertainty in the ERF, Bellouin et al (Rev. Geophys, 2020) is perhaps the most up-to-date complete estimate.

In addition, perhaps a relevant (and interesting) comparison would be for the CAS-FGOALS-f3-L model with the old aerosol scheme? This is obviously a lot of work if you don't have the output, so I wouldn't require it. However, it might be interesting to see the improvement in the model when moving to the new scheme.

Minor points

L22 - We incorporate aerosol-cloud interactions into

L33 - reasonably well.

L43 - this sentence is a bit of a jumble.

L88 - analyzed

L103 - graupel

L119 - This sentence need to be re-written

L176 - natural aerosols

L178 - I am not sure I understand the reasoning here - surely there is relatively little change in the surface temperature with the fixed SSTs, so there would not be expected to be much of a change in dust emission?

Fig. 1 - Perhaps consider labeling the third column with $\Delta$s - e.g. $\Delta$SU, $\Delta$CA etc.

L202 - are there some numbers for the South Asia statement, or is this just done by eye?

L22 - hydrophilic

L226 - What is meant by the 'real CCN concentration'? I assume this means the 'relevant CCN concentration', as the number of activated CCN changes with updraft speed.

L237 - 'small increases' is more typical than 'light increases'

L242 - Is it clear that the LWP changes are due to changes in precipitation? Precip is never shown here.

L243 - CDNC affects autoconversion in the Berry scheme. This is related to CCN, but not the same as CCN.

Fig. 2. - Again, consider $\Delta$s in the appropriate titles

Fig. 2 - Cloud fraction has been shown to have a potentially large impact on the ERFaci. Is the change in cloud fraction large in your model?

L271 - The s.d. is larger, but the mean is also larger. Is it not expected to have a larger standard deviation?

L288 - 'More than 4 times' what is this in reference to?

L292 - A short comment on the LW forcing might be appropriate (as you have calculated it). How does it compare between the models (even if it is always very close to zero (the last IPCC report suggested it was positive).

L302 - FGOALS and SPRINTARS are spelled out again, but FAMIL is not.

L338 - The last sentence doesn't make sense

L359 - Abdul-Razzak and Ghan activation is also partially empirical. I know this paper is not about the BCC model though, so you don't need to investigate further (and you don't need to say you will either, unless you want to!)

Author Response

"Please see the attachment."

Author Response File: Author Response.docx

Reviewer 2 Report

Review of paper:

Aerosol Effective Radiative Forcing in the online aerosol coupled CAS-FGOALS-f3-L climate model.                                                                                                                                                                   by H.Wang et al.

Positives

• global climate effects by anthropogenic aerosol … as expected
• aerosol impacts split into components: DRF and CRF (and SRF)
• comparison of global distribution to output from other global models
• focused – no too long
  
  

Concerns

• avg are biased and std.deviation (uncertainty?) are exaggerated by biased outliers
• limitations to shortwave radiative effects (without ‘justified explanation’)
• missed opportunity on (dis-?) agreement to other climate models using the same SPRINTARS aerosol module ?

General comments:

The paper presents model simulations of the shortwave (SW) TOA radiative impacts by today’s anthropogenic aerosol. The simulations add the SPRINTARS aerosol component module (which also considers aerosol-watercloud interactions) to the Chinese CAS-FGOALS-f3-L climate model. The presented global average annual estimates and annual spatial distributions are close to expectations as demonstrated by comparisons to estimates to other climate model of the CMIP6 exercise. Based on the global model spread ‘uncertainty’ is suggested. Aside from general issues on this assumption (model result may better agree (1) due to offsetting errors or (2) because of same critical possibly false assumptions), the applied statistics (average, std deviation) suggest less negative CRF and a more negative DRF and larger ‘uncertainty’ by including admittedly biased outliers. For DRF the outlier is the GFDL-ESM4 with an unlikely strong absorption. And for the CRF the MRI-ESM2-0 should not be considered with added aerosol/ice-cloud interactions, because eventually the larger shortwave SW forcing is partially compensated by the non-zero LW forcing (as listed in Table 1). To understand DRF differences a comparison of SSA assumed for anthropogenic aerosol might be interesting and for the CRF I would plot in Figure 4 the total (SW +LW, assuming that CRFice is close to zero). Only SW radiative effects are discussed. So for climate implications LW effects are ignored. This can be justified by the (reasonable) assumptions that anthropogenic aerosol only contributes by the fine-mode sizes (those smaller than 1um) and that only indirect effect through low altitude water clouds matter. But these arguments are missed. So, when comparing SW radiative forcing comparison (of Figure 5), models that include aerosol impacts on ice-clouds, should be removed. An interesting aspect also would be to compare DRF and CRF models that apply the same SPRINTARS aerosol module, to investigate diversity introduced by the carrier model and /or carrier model resolution.
Overall, though this is a nice compact contribution.  

Minor comments:

Page 3/119       the dry (why dry?) modal radii for OC and OC/BC are awfully small (factor 10?)

Page 3/121       drove à driven

Page 4/153       are 30 years sufficient for significance (e.g. CRF related LWC changes) ?

Page 5/240       the CNRM_CM6-1 lack apparent aerosol absorption (strong cooling near continental sources) and in GFDL-ESM4 aerosol absorption is too strong (possibly aided by the assumption of elevated absorbing aerosol above clouds…?)

Page 6/211       the table is nice, also showing that the adjustment SRF is small & slightly positive

Page 7/237       arguments to explain the Twomey effect need to be reformulated (CDNC yes … but LWC increase?)

Page 8/266       without the MRI model the global average CRF is well below -1.0W/m2. In that context the CRF (and also the associated ERF) by the new model is relatively negative (too strong climate cooling) compared to most other global models.

Author Response

Please see the attachment.

Author Response File: Author Response.docx