the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Aug 2023
Sensitivity of global direct aerosol radiative forcing to uncertainties in aerosol optical properties
Abstract. New satellite missions promise global reductions in the uncertainties of aerosol optical properties but it is unclear how those reductions will propagate to uncertainties in the shortwave direct aerosol radiative effect (DARE) and radiative forcing (DARF), which are currently large, on the order of at least 20 %. In this work we build a Monte-Carlo framework to calculate the impact of uncertainties in aerosol optical depth (AOD), single scattering albedo (SSA) and asymmetry parameter on the uncertainty in shortwave DARE and DARF. This framework uses the results of over 2.3 million radiative transfer simulations to calculate global clear-sky DARE and DARF based on a range of aerosol optical property uncertainties, representative of existing and future global observing systems. We find the one-sigma uncertainty varies between ±0.23 to ±1.91 Wm-2 (5 and 42 %) for the top of atmosphere (TOA) clear-sky DARE and between ±0.08 to ±0.47 Wm-2 (9 and 52 %) for the TOA DARF. At the TOA, AOD uncertainty is the main contributor to overall uncertainty, except over bright surfaces where SSA uncertainty contributes most. We apply regionally varying uncertainties to represent current measurement uncertainties, finding that aerosol optical property uncertainties represent 24 % of TOA DARE and DARF. Reducing regionally varying optical property uncertainties by a factor of two would reduce their contributions to TOA DARE and DARF uncertainty proportionally. Scaling to all-sky conditions, aerosol optical property uncertainty contributes to about 25 % total uncertainty in TOA, all-sky DARE and DARF. Compared to previous studies which considered uncertainties in non-aerosol variables, our results suggest that the aerosol optical property uncertainty accounts for a third to a half of total uncertainty. Recent and future progress in constraining aerosol optical properties using ground-based or satellite retrievals could be translated into DARE and DARF uncertainty using our freely available framework.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(2789 KB)
-
Supplement
(456 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2789 KB) - Metadata XML
-
Supplement
(456 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-1096', Anonymous Referee #1, 11 Sep 2023
The manuscript addresses the impact of the uncertainties in the aerosol optical properties (aerosol optical depth, single scattering albedo, and asymmetry parameter) on the uncertainty in the surface and Top of the Atmosphere (TOA) direct aerosol radiative effect (DARE) and forcing (DARF). This is done by building and implementing a Monte-Carlo framework. The DARE and DARF uncertainties, despite the progress made in recent years, remain large. The manuscript contributes to the discussion in the scientific community regarding these uncertainties and the results (especially the geographic distributions) are interesting.
However, there is a need for clarifications to be provided, and the discussion should be expanded to include more details, especially in section 3.3. There are also some issues to be addressed, listed below.
l 12 It's better to use the term "aerosol optical properties" instead of "aerosol optical property"
Sect. 1. I suggest to provide here more information about the novelty of your study.
l 101: multiplying or dividing?
l 105: "These are therefore kept constant throughout the whole vertical profile." If I understand well you kept the SSA and g of the total aerosol load constant in the atmospheric column? According to Kinne et al., 2019 "the total values for SSA and g at each altitude depend on the relative AOD contributions. It was not possible to employ this methodology to calculate the vertical profiles of SSA and g? Am I missing something? The lack of knowledge of the vertical distribution of SSA and g induces some biases in DRE calculations (Thorsen et al., 2020)
ll 130-132: Do you mean "Fig. 2"? Why do you compare the DARE presented in Fig. 1 with the DARF (Fig. 7, Kinne 2019) and not Fig. 5 of Kinne 2019? Maybe you mean Fig. 2 of the manuscript?
l 146: "While there are biases of up to 5% spatially". Do you mean "locally"?
Section 2.3: To make it more accessible to readers less familiar with Monte-Carlo simulations, it would be beneficial to expand upon it.
ll 198-200: It's not clear to me how the employed methodology accounts for covariances between uncertainties in different optical properties.
ll 204-205: "A draw is taken from a Gaussian distribution centred on the global mean AOD with standard deviation σAOD equal to the AOD uncertainty." So, did you randomly choose an uncertainty from the range 0.005-0.05 (Table 2)? Also, please correct the typo "centred".
l 207: " applied depending on aerosol type". Can you elaborate on this? In the formula below the uncertainty depends on the surface type.
ll 218-220: "Since typical ... calculated" This phrase is somewhat confusing. What do you mean by "with perturbations transformed back into ω0"? You didn't apply perturbations directly to ω0? Please clarify this.
l 248: "Figure 5 shows an example of the global-annual mean TOA DARF for one set of input uncertainties". Which specific set?
l 263: "shown in Figure 1" Maybe you mean Figure 2 since you are referring to "forcing".
Fig 10 & ll 322-323: I can observe an SSA uncertainty domination at TOA also in the Arctic (albeit not as pronounced as over deserts).
l 324-327: "It may also be due to limitations ... for this analysis".
To my opinion, some clarifications are needed here.
Doesn't the larger AOD at longer wavelengths increase the sensitivity of DARE to σAOD over desert regions? Also, you mention that "increased scattering from SSA offset by increased backscatter from asymmetry". Both increased SSA and increased backscatter result in a decrease of the TOA warming/ increase of TOA cooling effect. Why do you use the term "offset"? Also, I would expect that in case of an "offset", the uncertainty of the DARE would decrease and not increase.
l 335: "particularly the strong anthropogenic forcing over East Asia".
Do you mean that "the sensitivities of DARF uncertainty to AOD and SSA are very similar for most regions, particularly over East Asia where there is a strong anthropogenic forcing"? If yes, please rephrase accordingly, as the current sentence is confusing.
l 337: "This reflects the generally even shape of the contours on Figure" Please elaborate more on this.
l 340: In the phrase "In both cases" it would be better to clarify to which cases you are referring because the current phrasing is somewhat ambiguous
ll 350-352: "For DARF, the main contributor varies regionally, but AOD generally dominates, except again over bright surfaces, and over remote, low-AOD regions where scattering dominates due to high SSA sea-salt aerosol, where g dominates."
There are some issues here:
1) According to fig. 12a SSA dominates also in central Africa and most of the Indian subcontinent.
2) g dominates not only over remote regions but also in the region of the Saharan dust outflow (tropical North Atlantic).
Please correct the sentence and provide some explanation for the results.
l 361: When referring to "lookup tables," which lookup tables are you referring to? This is the first time this term is used in the manuscript. Please provide more details.
ll 371: "Since B13 only defines uncertainties ... for this case.". It would be useful here to expand the discussion by providing more information about the uncertainties.
ll 384-386: Does the phrase "scaling the values and their associated uncertainties by the global-mean AOD" imply that the AOD-associated uncertainty is linearly proportional to the AOD?
ll 397-401: The discussion here is too brief. Please provide more information, especially regarding the non-linearities.
Sect. 4, Table 3: Why did you choose to perform the analysis only using the optical properties uncertainties from Bellouin et al. (2013)?
ll 472-474: These are interesting results. To my opinion, they need to be discussed by providing more details (in Sect. 3.3).
References
Kinne, S., 2019. Aerosol radiative effects with MACv2. Atmospheric Chemistry and Physics, 19(16), pp.10919-10959. Doi: 10.5194/acp-19-10919-2019
Thorsen, T.J., Ferrare, R.A., Kato, S. and Winker, D.M., 2020. Aerosol direct radiative effect sensitivity analysis. Journal of Climate, 33(14), pp.6119-6139. doi: 10.1175/JCLI-D-19-0669.1
Citation: https://doi.org/10.5194/egusphere-2023-1096-RC1 - RC2: 'Comment on egusphere-2023-1096', Stefan Kinne, 13 Sep 2023
- AC1: 'Response to reviewers', Claire Ryder, 27 Feb 2024
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-1096', Anonymous Referee #1, 11 Sep 2023
The manuscript addresses the impact of the uncertainties in the aerosol optical properties (aerosol optical depth, single scattering albedo, and asymmetry parameter) on the uncertainty in the surface and Top of the Atmosphere (TOA) direct aerosol radiative effect (DARE) and forcing (DARF). This is done by building and implementing a Monte-Carlo framework. The DARE and DARF uncertainties, despite the progress made in recent years, remain large. The manuscript contributes to the discussion in the scientific community regarding these uncertainties and the results (especially the geographic distributions) are interesting.
However, there is a need for clarifications to be provided, and the discussion should be expanded to include more details, especially in section 3.3. There are also some issues to be addressed, listed below.
l 12 It's better to use the term "aerosol optical properties" instead of "aerosol optical property"
Sect. 1. I suggest to provide here more information about the novelty of your study.
l 101: multiplying or dividing?
l 105: "These are therefore kept constant throughout the whole vertical profile." If I understand well you kept the SSA and g of the total aerosol load constant in the atmospheric column? According to Kinne et al., 2019 "the total values for SSA and g at each altitude depend on the relative AOD contributions. It was not possible to employ this methodology to calculate the vertical profiles of SSA and g? Am I missing something? The lack of knowledge of the vertical distribution of SSA and g induces some biases in DRE calculations (Thorsen et al., 2020)
ll 130-132: Do you mean "Fig. 2"? Why do you compare the DARE presented in Fig. 1 with the DARF (Fig. 7, Kinne 2019) and not Fig. 5 of Kinne 2019? Maybe you mean Fig. 2 of the manuscript?
l 146: "While there are biases of up to 5% spatially". Do you mean "locally"?
Section 2.3: To make it more accessible to readers less familiar with Monte-Carlo simulations, it would be beneficial to expand upon it.
ll 198-200: It's not clear to me how the employed methodology accounts for covariances between uncertainties in different optical properties.
ll 204-205: "A draw is taken from a Gaussian distribution centred on the global mean AOD with standard deviation σAOD equal to the AOD uncertainty." So, did you randomly choose an uncertainty from the range 0.005-0.05 (Table 2)? Also, please correct the typo "centred".
l 207: " applied depending on aerosol type". Can you elaborate on this? In the formula below the uncertainty depends on the surface type.
ll 218-220: "Since typical ... calculated" This phrase is somewhat confusing. What do you mean by "with perturbations transformed back into ω0"? You didn't apply perturbations directly to ω0? Please clarify this.
l 248: "Figure 5 shows an example of the global-annual mean TOA DARF for one set of input uncertainties". Which specific set?
l 263: "shown in Figure 1" Maybe you mean Figure 2 since you are referring to "forcing".
Fig 10 & ll 322-323: I can observe an SSA uncertainty domination at TOA also in the Arctic (albeit not as pronounced as over deserts).
l 324-327: "It may also be due to limitations ... for this analysis".
To my opinion, some clarifications are needed here.
Doesn't the larger AOD at longer wavelengths increase the sensitivity of DARE to σAOD over desert regions? Also, you mention that "increased scattering from SSA offset by increased backscatter from asymmetry". Both increased SSA and increased backscatter result in a decrease of the TOA warming/ increase of TOA cooling effect. Why do you use the term "offset"? Also, I would expect that in case of an "offset", the uncertainty of the DARE would decrease and not increase.
l 335: "particularly the strong anthropogenic forcing over East Asia".
Do you mean that "the sensitivities of DARF uncertainty to AOD and SSA are very similar for most regions, particularly over East Asia where there is a strong anthropogenic forcing"? If yes, please rephrase accordingly, as the current sentence is confusing.
l 337: "This reflects the generally even shape of the contours on Figure" Please elaborate more on this.
l 340: In the phrase "In both cases" it would be better to clarify to which cases you are referring because the current phrasing is somewhat ambiguous
ll 350-352: "For DARF, the main contributor varies regionally, but AOD generally dominates, except again over bright surfaces, and over remote, low-AOD regions where scattering dominates due to high SSA sea-salt aerosol, where g dominates."
There are some issues here:
1) According to fig. 12a SSA dominates also in central Africa and most of the Indian subcontinent.
2) g dominates not only over remote regions but also in the region of the Saharan dust outflow (tropical North Atlantic).
Please correct the sentence and provide some explanation for the results.
l 361: When referring to "lookup tables," which lookup tables are you referring to? This is the first time this term is used in the manuscript. Please provide more details.
ll 371: "Since B13 only defines uncertainties ... for this case.". It would be useful here to expand the discussion by providing more information about the uncertainties.
ll 384-386: Does the phrase "scaling the values and their associated uncertainties by the global-mean AOD" imply that the AOD-associated uncertainty is linearly proportional to the AOD?
ll 397-401: The discussion here is too brief. Please provide more information, especially regarding the non-linearities.
Sect. 4, Table 3: Why did you choose to perform the analysis only using the optical properties uncertainties from Bellouin et al. (2013)?
ll 472-474: These are interesting results. To my opinion, they need to be discussed by providing more details (in Sect. 3.3).
References
Kinne, S., 2019. Aerosol radiative effects with MACv2. Atmospheric Chemistry and Physics, 19(16), pp.10919-10959. Doi: 10.5194/acp-19-10919-2019
Thorsen, T.J., Ferrare, R.A., Kato, S. and Winker, D.M., 2020. Aerosol direct radiative effect sensitivity analysis. Journal of Climate, 33(14), pp.6119-6139. doi: 10.1175/JCLI-D-19-0669.1
Citation: https://doi.org/10.5194/egusphere-2023-1096-RC1 - RC2: 'Comment on egusphere-2023-1096', Stefan Kinne, 13 Sep 2023
- AC1: 'Response to reviewers', Claire Ryder, 27 Feb 2024
Peer review completion
Journal article(s) based on this preprint
Data sets
Radiative effect and forcing sensitivity data and software Jonathan Elsey, Nicolas Bellouin, Claire Ryder https://doi.org/10.5281/zenodo.7958296
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 336 | 150 | 22 | 508 | 40 | 15 | 17 |
- HTML: 336
- PDF: 150
- XML: 22
- Total: 508
- Supplement: 40
- BibTeX: 15
- EndNote: 17
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Jonathan Elsey
Nicolas Bellouin
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2789 KB) - Metadata XML
-
Supplement
(456 KB) - BibTeX
- EndNote
- Final revised paper