the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 09 Aug 2024
Exploring the processes controlling secondary inorganic aerosol: Evaluating the global GEOS-Chem simulation using a suite of aircraft campaigns
Abstract. Secondary inorganic aerosols (sulfate, nitrate, and ammonium; SNA) are major contributors to fine particulate matter. Predicting concentrations of these species is complicated by the cascade of processes that control their abundance, including emissions, chemistry, thermodynamic partitioning, and removal. In this study, we use 11 flight campaigns to evaluate the GEOS-Chem model performance for SNA. Across all the campaigns, the model performance is best for sulfate (R2 = 0.51, NMB = 0.11) and worst for nitrate (R2 = 0.22, NMB = 1.76), indicating substantive model deficiencies in the nitrate simulation. Thermodynamic partitioning reproduces the total particulate nitrate well (R2 = 0.79 and NMB = 0.09), but actual partitioning (i.e., εNO3= NO3-/TNO3) is challenging to assess given limited ammonia observations. Model performance is sensitive to changes in emissions and dry and wet deposition, with modest improvements associated with the inclusion of different chemical loss and production pathways (i.e., acid uptake on dust, N2O5 uptake, and NO3- photolysis). However, these sensitivity tests show only modest reduction in the nitrate bias, with no improvement to the model skill (i.e., R2) implying that more work is needed to improve the description of loss and production of nitrate and SNA as a whole.
- Preprint
(6733 KB) - Metadata XML
-
Supplement
(4509 KB) - BibTeX
- EndNote
Status: closed
- RC1: 'Comment on egusphere-2024-2296', Anonymous Referee #1, 28 Aug 2024
-
RC2: 'Comment on egusphere-2024-2296', Anonymous Referee #2, 29 Aug 2024
Norman et al perform a comparison of observed and modeled (GEOS-Chem) sulfate-nitrate-ammonium (SNA) aerosol in order to address a longstanding issue in models of significant discrepancies in nitrate and ammonia. This is important to address in light of the role that SNA plays in PM2.5 abundance. Policies aimed at addressing PM2.5 pollution can best be made with an improved understanding of what controls their abundance. The authors perform a thorough comparison of the model with observations from 11 field campaigns in the US, Europe and east Asia with the goal of addressing SNA discrepancies in regions impacted by anthropogenic pollution. Similar to other model-observation comparisons, they find good model-obs agreement for sulfate, but the model generally overestimates nitrate and underestimates ammonia. They use GEOS-Chem and a stand-alone version of the aerosol thermodynamic model ISORROPIA to examine reasons for these model biases. They are able to run certain factors out (biases in transport, precipitation, thermodynamic partitioning of HNO3/NO3- and NH3/NH4+, dry deposition, chemistry) and find that uncertainties emissions and wet deposition play a larger role.
This paper is well written and scientifically sound and is thus suitable for publication in ACP. Their ruling out of processes that don’t impact the model-obs discrepancies should help to move the science forward. I have two minor suggestions for improvement:
In the abstract, it is not immediately clear how the partitioning of HNO3/NO3- is hard to assess with limited ammonia observations. It seems like you would need observations of both HNO3 and NO3-, not ammonia. By the end of the paper is it more clear what you mean, so perhaps you should include more information on this in the abstract.
Section 3.1 is missing a description of the sinks of SNA aerosol. Is wet and dry deposition the only sink? If so, say so. Similarly, it is unclear in Table 2 if the lifetime is determined by wet and dry deposition or also by other processes. How was the lifetime calculated?
Citation: https://doi.org/10.5194/egusphere-2024-2296-RC2 -
RC3: 'Comment on egusphere-2024-2296', Anonymous Referee #3, 31 Aug 2024
Norman et al. present an evaluation of GEOS-Chem model predictions of sulfate-nitrate-ammonium (SNA) for aircraft campaigns. Box modeling is used to investigate possible drivers of error and Figure 11 is a useful demonstration of what could drive error. The authors conclude total nitrate is likely overestimated which drivers overestimates in particulate nitrate.
General comment: While Figure 11 is very convincing in showing nitrate errors likely drive SNA errors, the lack of role for partitioning errors wasn’t completely demonstrated. Figure 6 shows pNO3- is underestimated below 4km for FIREX while Figure S2 shows HNO3 is overestimated indicating a partitioning error. Consider that errors in partitioning will affect lifetime due to different remove rates of HNO3 vs accumulation mode particle deposition (e.g., Nenes et al. 2021). Recent work has shown treating SNA as non-equilibrium can reduce model bias (Rosanka et al. 2024). A clearer demonstration that partitioning is not the issue and/or some investigation to bound the role is needed. Consider a providing an HNO3 and NH3 budget in Table 2 as well as more information on total NO3 and total NHx in the main text. Consider adding a figure that synthesizes across all the sensitivity simulations so they can be more easily compared in terms of relative impact and direction of changes.
Specific comments:
- Introduction: consider mentioning how VOCs can modulate nitrate abundance (Womack et al. 2019).
- Page 2, near line 76: consider rewording to emphasize that ammonia isn’t reacting stoichiometrically first with sulfate then second with nitrate as bisulfate is a common form of sulfate. What is meant by the term neutralize? pH 7? Note aerosols always have charge balance when H+ and OH- are considered.
- Figure 1: Add years and/or months on the campaigns.
- Figure 3: What is the current accuracy of PM2.5 and/or OA predictions for the 2018 period?
- Figure 6: are nitrate measurements above 3km during CalNEX below the limit of detection? They seem below the 0.
- Clarify methods for Figure 7 and the box modeling. Is a forward ISORROPIA calculation always used? Consider renaming Figure 7 x-axis to “measured concentration”.
- Section 5.2: Did you consider how changes to VOC emissions may affect total nitrate?
References:
Nenes, A., Pandis, S. N., Kanakidou, M., Russell, A. G., Song, S., Vasilakos, P., and Weber, R. J.: Aerosol acidity and liquid water content regulate the dry deposition of inorganic reactive nitrogen, Atmos. Chem. Phys., 21, 6023–6033, https://doi.org/10.5194/acp-21-6023-2021, 2021.
Rosanka, S., Tost, H., Sander, R., Jöckel, P., Kerkweg, A., and Taraborrelli, D.: How non-equilibrium aerosol chemistry impacts particle acidity: the GMXe AERosol CHEMistry (GMXe–AERCHEM, v1.0) sub-submodel of MESSy, Geosci. Model Dev., 17, 2597–2615, https://doi.org/10.5194/gmd-17-2597-2024, 2024.
Womack, C. C., McDuffie, E. E., Edwards, P. M., Bares, R., de Gouw, J. A., Docherty, K. S., et al. (2019). An odd oxygen framework for wintertime ammonium nitrate aerosol pollution in urban areas: NOx and VOC control as mitigation strategies. Geophysical Research Letters, 46, 4971–4979. https://doi.org/10.1029/2019GL082028
Citation: https://doi.org/10.5194/egusphere-2024-2296-RC3 -
RC4: 'Comment on egusphere-2024-2296', Anonymous Referee #4, 03 Sep 2024
Norman et al. present a comparison between aircraft-based measurements of sulfate, nitrate, and ammonium aerosol concentrations and GEOS-Chem simulations. They find that the model has more skill in reproducing sulfate observations than nitrate observations. Using various sensitivity tests, they identify several mechanisms that the model is sensitive to, but none are able to fully correct for differences between observations and measurements. The manuscript is well-written and thorough. I offer just a few specific points and questions below that may serve to improve an already high-quality work.
The authors could use the findings as a call to routinely monitor ammonia concentrations, especially in aircraft campaigns
140: is only sub-micron SNA captured by the AMS? Does this match the modeled size cutoff in Geos-chem?
Table 2 (and elsewhere, e.g., line 618): the authors use the term “burden”—I believe replacing this with something more specific such as “concentration” would be more precise.
Figure 4 (& in the Discussion): it would be useful to test (and possibly present) R along with R2 to identify any anti-correlation
260: The authors mention scaling by the nitrate NMB. When this scaling is performed, how does it affect bias in modeled total PM2.5 at ground-based monitors? It may be useful to add a brief discussion about GEOS-Chem’s performance at ground-based monitors to establish whether the comparisons with aircraft campaigns are representative of previous model evaluations.
Section 5.1.1: what effect is the missing dust cations expected to have on nitrate partitioning?
475: it would be helpful to have a theoretical reason for using the cumulative nitrate and ammonium NMB—this approach weights the biases of the two chemicals equally even though they make up different fractions of PM mass.
Citation: https://doi.org/10.5194/egusphere-2024-2296-RC4 - AC1: 'Comment on egusphere-2024-2296', Olivia Norman, 03 Nov 2024
Status: closed
- RC1: 'Comment on egusphere-2024-2296', Anonymous Referee #1, 28 Aug 2024
-
RC2: 'Comment on egusphere-2024-2296', Anonymous Referee #2, 29 Aug 2024
Norman et al perform a comparison of observed and modeled (GEOS-Chem) sulfate-nitrate-ammonium (SNA) aerosol in order to address a longstanding issue in models of significant discrepancies in nitrate and ammonia. This is important to address in light of the role that SNA plays in PM2.5 abundance. Policies aimed at addressing PM2.5 pollution can best be made with an improved understanding of what controls their abundance. The authors perform a thorough comparison of the model with observations from 11 field campaigns in the US, Europe and east Asia with the goal of addressing SNA discrepancies in regions impacted by anthropogenic pollution. Similar to other model-observation comparisons, they find good model-obs agreement for sulfate, but the model generally overestimates nitrate and underestimates ammonia. They use GEOS-Chem and a stand-alone version of the aerosol thermodynamic model ISORROPIA to examine reasons for these model biases. They are able to run certain factors out (biases in transport, precipitation, thermodynamic partitioning of HNO3/NO3- and NH3/NH4+, dry deposition, chemistry) and find that uncertainties emissions and wet deposition play a larger role.
This paper is well written and scientifically sound and is thus suitable for publication in ACP. Their ruling out of processes that don’t impact the model-obs discrepancies should help to move the science forward. I have two minor suggestions for improvement:
In the abstract, it is not immediately clear how the partitioning of HNO3/NO3- is hard to assess with limited ammonia observations. It seems like you would need observations of both HNO3 and NO3-, not ammonia. By the end of the paper is it more clear what you mean, so perhaps you should include more information on this in the abstract.
Section 3.1 is missing a description of the sinks of SNA aerosol. Is wet and dry deposition the only sink? If so, say so. Similarly, it is unclear in Table 2 if the lifetime is determined by wet and dry deposition or also by other processes. How was the lifetime calculated?
Citation: https://doi.org/10.5194/egusphere-2024-2296-RC2 -
RC3: 'Comment on egusphere-2024-2296', Anonymous Referee #3, 31 Aug 2024
Norman et al. present an evaluation of GEOS-Chem model predictions of sulfate-nitrate-ammonium (SNA) for aircraft campaigns. Box modeling is used to investigate possible drivers of error and Figure 11 is a useful demonstration of what could drive error. The authors conclude total nitrate is likely overestimated which drivers overestimates in particulate nitrate.
General comment: While Figure 11 is very convincing in showing nitrate errors likely drive SNA errors, the lack of role for partitioning errors wasn’t completely demonstrated. Figure 6 shows pNO3- is underestimated below 4km for FIREX while Figure S2 shows HNO3 is overestimated indicating a partitioning error. Consider that errors in partitioning will affect lifetime due to different remove rates of HNO3 vs accumulation mode particle deposition (e.g., Nenes et al. 2021). Recent work has shown treating SNA as non-equilibrium can reduce model bias (Rosanka et al. 2024). A clearer demonstration that partitioning is not the issue and/or some investigation to bound the role is needed. Consider a providing an HNO3 and NH3 budget in Table 2 as well as more information on total NO3 and total NHx in the main text. Consider adding a figure that synthesizes across all the sensitivity simulations so they can be more easily compared in terms of relative impact and direction of changes.
Specific comments:
- Introduction: consider mentioning how VOCs can modulate nitrate abundance (Womack et al. 2019).
- Page 2, near line 76: consider rewording to emphasize that ammonia isn’t reacting stoichiometrically first with sulfate then second with nitrate as bisulfate is a common form of sulfate. What is meant by the term neutralize? pH 7? Note aerosols always have charge balance when H+ and OH- are considered.
- Figure 1: Add years and/or months on the campaigns.
- Figure 3: What is the current accuracy of PM2.5 and/or OA predictions for the 2018 period?
- Figure 6: are nitrate measurements above 3km during CalNEX below the limit of detection? They seem below the 0.
- Clarify methods for Figure 7 and the box modeling. Is a forward ISORROPIA calculation always used? Consider renaming Figure 7 x-axis to “measured concentration”.
- Section 5.2: Did you consider how changes to VOC emissions may affect total nitrate?
References:
Nenes, A., Pandis, S. N., Kanakidou, M., Russell, A. G., Song, S., Vasilakos, P., and Weber, R. J.: Aerosol acidity and liquid water content regulate the dry deposition of inorganic reactive nitrogen, Atmos. Chem. Phys., 21, 6023–6033, https://doi.org/10.5194/acp-21-6023-2021, 2021.
Rosanka, S., Tost, H., Sander, R., Jöckel, P., Kerkweg, A., and Taraborrelli, D.: How non-equilibrium aerosol chemistry impacts particle acidity: the GMXe AERosol CHEMistry (GMXe–AERCHEM, v1.0) sub-submodel of MESSy, Geosci. Model Dev., 17, 2597–2615, https://doi.org/10.5194/gmd-17-2597-2024, 2024.
Womack, C. C., McDuffie, E. E., Edwards, P. M., Bares, R., de Gouw, J. A., Docherty, K. S., et al. (2019). An odd oxygen framework for wintertime ammonium nitrate aerosol pollution in urban areas: NOx and VOC control as mitigation strategies. Geophysical Research Letters, 46, 4971–4979. https://doi.org/10.1029/2019GL082028
Citation: https://doi.org/10.5194/egusphere-2024-2296-RC3 -
RC4: 'Comment on egusphere-2024-2296', Anonymous Referee #4, 03 Sep 2024
Norman et al. present a comparison between aircraft-based measurements of sulfate, nitrate, and ammonium aerosol concentrations and GEOS-Chem simulations. They find that the model has more skill in reproducing sulfate observations than nitrate observations. Using various sensitivity tests, they identify several mechanisms that the model is sensitive to, but none are able to fully correct for differences between observations and measurements. The manuscript is well-written and thorough. I offer just a few specific points and questions below that may serve to improve an already high-quality work.
The authors could use the findings as a call to routinely monitor ammonia concentrations, especially in aircraft campaigns
140: is only sub-micron SNA captured by the AMS? Does this match the modeled size cutoff in Geos-chem?
Table 2 (and elsewhere, e.g., line 618): the authors use the term “burden”—I believe replacing this with something more specific such as “concentration” would be more precise.
Figure 4 (& in the Discussion): it would be useful to test (and possibly present) R along with R2 to identify any anti-correlation
260: The authors mention scaling by the nitrate NMB. When this scaling is performed, how does it affect bias in modeled total PM2.5 at ground-based monitors? It may be useful to add a brief discussion about GEOS-Chem’s performance at ground-based monitors to establish whether the comparisons with aircraft campaigns are representative of previous model evaluations.
Section 5.1.1: what effect is the missing dust cations expected to have on nitrate partitioning?
475: it would be helpful to have a theoretical reason for using the cumulative nitrate and ammonium NMB—this approach weights the biases of the two chemicals equally even though they make up different fractions of PM mass.
Citation: https://doi.org/10.5194/egusphere-2024-2296-RC4 - AC1: 'Comment on egusphere-2024-2296', Olivia Norman, 03 Nov 2024
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 463 | 109 | 18 | 590 | 35 | 7 | 4 |
- HTML: 463
- PDF: 109
- XML: 18
- Total: 590
- Supplement: 35
- BibTeX: 7
- EndNote: 4
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1