the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 Jan 2024
Observation-Constrained Kinetic Modelling of Isoprene SOA Formation in the Atmosphere
Abstract. Isoprene has the largest global non-methane hydrocarbon emission, and the oxidation of isoprene plays a crucial role in the formation of secondary organic aerosols (SOA). Two primary processes are known to contribute to SOA formation from isoprene oxidation: (1) the reactive uptake of isoprene-derived epoxides on acidic or aqueous particle surfaces and (2) the absorptive gas-particle partitioning of low-volatility oxidation products. In this study, we developed a new multiphase condensed isoprene oxidation mechanism that include these processes with key molecular intermediates and products. The new mechanism was applied to simulate isoprene gas-phase oxidation products and SOA formation from previously published chamber experiments under a variety of conditions and atmospheric observations during the Southern Oxidant and Aerosol Studies (SOAS) field campaign. Our results show that SOA formation from most of the chamber experiments is reasonably reproduced using our mechanism except when the concentration ratios of initial nitric oxide to isoprene exceeds ~2. The SOAS simulations also reasonably agree with the measurements regarding the diurnal pattern and concentrations of different product categories. The molecular compositions of the modelled SOA indicate that multifunctional low-volatility products contribute to isoprene SOA more significantly than previously thought, with a median mass contribution of ~57 % to the total modelled isoprene SOA. This contribution, however, may vary greatly, mainly dependent on the volatility estimation and treatment of particle-phase processes (i.e., photolysis and hydrolysis). Our findings emphasize that the various pathways to produce these low-volatility species should be considered in models to more accurately predict isoprene SOA formation. The new condensed isoprene chemical mechanism can be further incorporated into regional-scale air quality models, such as the Community Multiscale Air Quality Modelling System (CMAQ), to assess isoprene SOA formation on a larger scale.
-
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
(4633 KB)
-
Supplement
(13342 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(4633 KB) - Metadata XML
-
Supplement
(13342 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-97', Anonymous Referee #1, 08 Feb 2024
This manuscript describes the evaluation of an isoprene oxidation mechanism designed to be compact enough for large-scale (regional to global) modeling while representing the pathways of secondary organic aerosol (SOA) formation in sufficient detail to simulate ambient and chamber observations across a wide range of conditions. The manuscript then compares box-modeled SOA using the new mechanism to a series of chamber experiments in both high and low NOx conditions as well as ambient observations from the SOAS campaign. Crucially, the mechanism enables comparisons of specific SOA constituents -- e.g. IEPOX-SOA, low volatility nitrogen-containing compounds, etc. -- with such speciated observations, allowing the authors to show that some subcategories compare well while others don't; notably, C5 nitrogen-containing low-volatility compounds seem overestimated compared to ambient observations, but SOA in high-NOx chamber experiments was underestimated, which poses an intriguing conundrum.Also of note, the commonly used AMS-PMF factor for IEPOX-SOA seems likely to also incorporate other non-IEPOX isoprene-derived SOA.
Overall, I found I had a lot of questions about the specifics of the mechanism which couldn't be answered with the level of detail provided. The authors should make the full mechanism available in the supplement or via a readily accessible public link so as to allow readers to answer these questions without having to request (and attempt to interpret) mechanism files. (Providing the tables of species and their SMILES strings in the SI, though, is a big plus!). The discussions of uncertainties, especially in the NO3 pathway (L 322-350) and in C* (L 412-430), are particular highlights -- I was very glad to see these included and described as well as they were. The manuscript is well-written and will be a valuable addition to the literature, and should be published after addressing some of the following concerns and providing the mechanism itself.
First, it's misleading to suggest that this UCR-ISOP mechanism represents a unique development in the incorporation of detailed isoprene chemistry into global models, or that previous global modeling studies have neglected to include detailed isoprene chemistry leading to SOA formation. Stadtler et al. (DOI 10.5194/gmd-11-3235-2018), Bates & Jacob (DOI 10.5194/acp-19-9613-2019), and Müller et al (DOI 10.5194/gmd-12-2307-2019) all included both IEPOX and non-IEPOX SOA pathways along with detailed gas-phase isoprene chemistry into global models, and showed the relative contributions of each of the pathways. While the mechanism in this manuscript includes more detailed particle-phase chemistry than any of those studies, the chemistry of the SOA precursors in Bates & Jacob appears to be of comparable complexity. At the very least these previous studies should be mentioned and their mechanisms cursorily compared; most helpful to readers and to the field in general would be some discussion of how the model outcomes compare, e.g. the relative contributions from the various SOA-forming pathways.
Abstract: I know you don't want to sell yourselves short, but it would be helpful to highlight more the way that C5-LVN appear to be the outlier in how well the model works. While you mention that the model-measurement agreement breaks down under high-NO conditions, you don't mention the direction of the bias, and then you gloss over the sharp disagreement with ambient data as shown in Figure 5C. I think mentioning these would motivate readers by showing what isn't yet well-understood, since it's clear the mechanism is good at representing what we *do* understand!
From Figure 1: which I realize is not meant to represent all the reactions in UCR-ISOP comprehensively, but it's all I have to go off -- it appears that many of the low volatility species are gas-phase "dead-ends", in that they lack loss pathways via reaction with OH. Is this the case, or were those reactions just left out of the figure? Either way, it would be helpful to describe whether such reactions exist in the mechanism and if so, how their rates and products were decided, since most such reactions have not been experimentally studied.
L 203-204: Any estimation of how much this assumption of homogeneity might influence your results -- i.e., how much the model outcomes could differ is particles were allowed to adopt core-shell morphologies, since that has been estimated and studied previously (see e.g. DOI 10.1021/acsearthspacechem.1c00156)?
L 214: Are other tertiary nitrates allowed to hydrolyze quickly with the Vasquez et al rate, or just 1,2-IHN? Following hydrolysis, are the products treated explicitly and allowed to repartition and react? (See e.g. DOI 10.1021/acs.est.1c04177)
L 231: Particle-phase hydroperoxide photolysis has been mentioned, but how were gas-phase photolysis rates applied? Were all hydroperoxides assumed to photolyze there too? What about carbonyl nitrates? (See e.g. DOI 10.5194/acp-14-2497-2014)
L 474: "were" should be "was"
L 489: "overpredict" should be "overpredicts"
L 497: This potential involvement of cloud interactions as a loss pathway is a great idea! Is there any way to test of parameterize this, e.g. simply using a cloud presence flag from the measurement site to see if the loss correlates with the presence of clouds?
L 569-577: The conclusions (a) that the widely used AMS-PMF factor for IEPOX-OA may in fact be convoluted with other isoprene SOA formation pathways, and (b) that total isoprene SOA remains drastically underestimated, both seem very important (particularly to people in this field that frequently come across the AMS-PMF analysis) and worth more highlighting. I would suggest that both of these merit a mention in the abstract and that some version of Figure S15 merits inclusion in the main manuscript. Does the AMS-PMF factor really include SOA from both LV pathways, as implied on L 575, or would the C5-NLV pathway be separable because the high-NOx products are from different chemical pathways (and presumably therefore have different temporal patterns)?
L 577: "suggests" should be "suggesting"
L 592: "that of that from" ... I think either needs to be "that of" or "that from", but not both?
L 594-595: This ratio of IHN to [ISOPOOH + IEPOX] doesn't seem at all to be an apples-to-apples comparison; the latter incorporates two generations of chemistry, and the lifetime of IEPOX with respect to OH oxidation is on the order of 5x longer than IHN. If the observed instantaneous ratio of IHN to [ISOPOOH + IEPOX] is 1:10, the ratio of the amount of carbon that goes through IHN to that which goes through the low-NOx, ISOPOOH+IEPOX pathway is probably closer to 1:2, given the difference in their lifetimes. While your point still stands that there's a surprisingly large amount to C5-NLV contributing to isoprene-derived SOA in a seemingly low-NOx environment, I would avoid characterizing that as IHN being 10x more potent on a per-carbon basis at SOA formation.
L 596: "is resulted" should be "results"
Citation: https://doi.org/10.5194/egusphere-2024-97-RC1 -
RC2: 'Comment on egusphere-2024-97', Anonymous Referee #2, 11 Feb 2024
This paper introduces a new isoprene SOA multiphase chemical mechanism and provides detailed analysis of the efficacy of the mechanism against both laboratory and field studies under a variety of different conditions. The results are very detailed and thorough with much discussion into both what the model does well and its limitations. However, I think a lack of detail on the specifics of the mechanism led to confusion about what the goal of introducing this mechanism is and how it is advancing the science beyond the currently existing isoprene SOA mechanisms. Overall the work fits well within the scope of ACP, the results are thorough, and the limitations are clear and well-described. I would recommend publication in ACP once a few comments are addressed.
- How were the reactions and species that were included in the model determined? Was the larger MCM reduced or were selected species/reactions added to SAPRC07? Either way additional description on how this was done is warranted.
- Is any comparison possible with machine learning-based reduced mechanisms (e.g. Wiser et al)? If direct comparison is not possible, I think a discussion of the pros/cons of this reduced model compared to machine learning-based mechanisms would be useful.
- More discussion would be helpful on how the mechanisms discussed in this work differ from each other beyond just the size of the mechanism. Since all the mechanisms seem to perform similarly in their ability to predict isoprene SOA, it is important for the authors to indicate what this new mechanism is adding beyond those that already exist.
- In particular, the comparison in Fig 2 seems to show that there is very little improvement in the UCR model from the SAPRC07 mechanism. In light of that, what is the additional chemistry representing better?
- Does the mechanism include gas-phase RO2+RO2 reactions to form dimers? This is mentioned as a possible pathway during the results, but it was not introduced in the methods, which led to some confusion for me.
- The resistor model for uptake coefficient assumes no particle-phase diffusion to uptake, however should IEPOX-SOA form a viscous organic shell around an acidic core, this may affect the uptake. While this would not help the modeled underestimation of IEPOX-SOA, do the authors expect this to impact the results in this work at all?
- In Fig 2, what was considered an “outlier”? Beyond the discussed issue reproducing high NOx experiments, was there any consistency in these outliers that were not represented well?
Minor comment
- Fig 4. Please check the +/-50% lines as one appears to be 1.5:1 and one appears to be 1:2. I assume these were meant to be the same ratio.
Reference
Wiser, F., Place, B. K., Sen, S., Pye, H. O. T., Yang, B., Westervelt, D. M., Henze, D. K., Fiore, A. M., and McNeill, V. F.: AMORE-Isoprene v1.0: a new reduced mechanism for gas-phase isoprene oxidation, Geosci. Model Dev., 16, 1801–1821, https://doi.org/10.5194/gmd-16-1801-2023, 2023.
Citation: https://doi.org/10.5194/egusphere-2024-97-RC2 -
AC1: 'Reply to Comments on egusphere-2024-97', Haofei Zhang, 17 Mar 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-97/egusphere-2024-97-AC1-supplement.pdf
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-97', Anonymous Referee #1, 08 Feb 2024
This manuscript describes the evaluation of an isoprene oxidation mechanism designed to be compact enough for large-scale (regional to global) modeling while representing the pathways of secondary organic aerosol (SOA) formation in sufficient detail to simulate ambient and chamber observations across a wide range of conditions. The manuscript then compares box-modeled SOA using the new mechanism to a series of chamber experiments in both high and low NOx conditions as well as ambient observations from the SOAS campaign. Crucially, the mechanism enables comparisons of specific SOA constituents -- e.g. IEPOX-SOA, low volatility nitrogen-containing compounds, etc. -- with such speciated observations, allowing the authors to show that some subcategories compare well while others don't; notably, C5 nitrogen-containing low-volatility compounds seem overestimated compared to ambient observations, but SOA in high-NOx chamber experiments was underestimated, which poses an intriguing conundrum.Also of note, the commonly used AMS-PMF factor for IEPOX-SOA seems likely to also incorporate other non-IEPOX isoprene-derived SOA.
Overall, I found I had a lot of questions about the specifics of the mechanism which couldn't be answered with the level of detail provided. The authors should make the full mechanism available in the supplement or via a readily accessible public link so as to allow readers to answer these questions without having to request (and attempt to interpret) mechanism files. (Providing the tables of species and their SMILES strings in the SI, though, is a big plus!). The discussions of uncertainties, especially in the NO3 pathway (L 322-350) and in C* (L 412-430), are particular highlights -- I was very glad to see these included and described as well as they were. The manuscript is well-written and will be a valuable addition to the literature, and should be published after addressing some of the following concerns and providing the mechanism itself.
First, it's misleading to suggest that this UCR-ISOP mechanism represents a unique development in the incorporation of detailed isoprene chemistry into global models, or that previous global modeling studies have neglected to include detailed isoprene chemistry leading to SOA formation. Stadtler et al. (DOI 10.5194/gmd-11-3235-2018), Bates & Jacob (DOI 10.5194/acp-19-9613-2019), and Müller et al (DOI 10.5194/gmd-12-2307-2019) all included both IEPOX and non-IEPOX SOA pathways along with detailed gas-phase isoprene chemistry into global models, and showed the relative contributions of each of the pathways. While the mechanism in this manuscript includes more detailed particle-phase chemistry than any of those studies, the chemistry of the SOA precursors in Bates & Jacob appears to be of comparable complexity. At the very least these previous studies should be mentioned and their mechanisms cursorily compared; most helpful to readers and to the field in general would be some discussion of how the model outcomes compare, e.g. the relative contributions from the various SOA-forming pathways.
Abstract: I know you don't want to sell yourselves short, but it would be helpful to highlight more the way that C5-LVN appear to be the outlier in how well the model works. While you mention that the model-measurement agreement breaks down under high-NO conditions, you don't mention the direction of the bias, and then you gloss over the sharp disagreement with ambient data as shown in Figure 5C. I think mentioning these would motivate readers by showing what isn't yet well-understood, since it's clear the mechanism is good at representing what we *do* understand!
From Figure 1: which I realize is not meant to represent all the reactions in UCR-ISOP comprehensively, but it's all I have to go off -- it appears that many of the low volatility species are gas-phase "dead-ends", in that they lack loss pathways via reaction with OH. Is this the case, or were those reactions just left out of the figure? Either way, it would be helpful to describe whether such reactions exist in the mechanism and if so, how their rates and products were decided, since most such reactions have not been experimentally studied.
L 203-204: Any estimation of how much this assumption of homogeneity might influence your results -- i.e., how much the model outcomes could differ is particles were allowed to adopt core-shell morphologies, since that has been estimated and studied previously (see e.g. DOI 10.1021/acsearthspacechem.1c00156)?
L 214: Are other tertiary nitrates allowed to hydrolyze quickly with the Vasquez et al rate, or just 1,2-IHN? Following hydrolysis, are the products treated explicitly and allowed to repartition and react? (See e.g. DOI 10.1021/acs.est.1c04177)
L 231: Particle-phase hydroperoxide photolysis has been mentioned, but how were gas-phase photolysis rates applied? Were all hydroperoxides assumed to photolyze there too? What about carbonyl nitrates? (See e.g. DOI 10.5194/acp-14-2497-2014)
L 474: "were" should be "was"
L 489: "overpredict" should be "overpredicts"
L 497: This potential involvement of cloud interactions as a loss pathway is a great idea! Is there any way to test of parameterize this, e.g. simply using a cloud presence flag from the measurement site to see if the loss correlates with the presence of clouds?
L 569-577: The conclusions (a) that the widely used AMS-PMF factor for IEPOX-OA may in fact be convoluted with other isoprene SOA formation pathways, and (b) that total isoprene SOA remains drastically underestimated, both seem very important (particularly to people in this field that frequently come across the AMS-PMF analysis) and worth more highlighting. I would suggest that both of these merit a mention in the abstract and that some version of Figure S15 merits inclusion in the main manuscript. Does the AMS-PMF factor really include SOA from both LV pathways, as implied on L 575, or would the C5-NLV pathway be separable because the high-NOx products are from different chemical pathways (and presumably therefore have different temporal patterns)?
L 577: "suggests" should be "suggesting"
L 592: "that of that from" ... I think either needs to be "that of" or "that from", but not both?
L 594-595: This ratio of IHN to [ISOPOOH + IEPOX] doesn't seem at all to be an apples-to-apples comparison; the latter incorporates two generations of chemistry, and the lifetime of IEPOX with respect to OH oxidation is on the order of 5x longer than IHN. If the observed instantaneous ratio of IHN to [ISOPOOH + IEPOX] is 1:10, the ratio of the amount of carbon that goes through IHN to that which goes through the low-NOx, ISOPOOH+IEPOX pathway is probably closer to 1:2, given the difference in their lifetimes. While your point still stands that there's a surprisingly large amount to C5-NLV contributing to isoprene-derived SOA in a seemingly low-NOx environment, I would avoid characterizing that as IHN being 10x more potent on a per-carbon basis at SOA formation.
L 596: "is resulted" should be "results"
Citation: https://doi.org/10.5194/egusphere-2024-97-RC1 -
RC2: 'Comment on egusphere-2024-97', Anonymous Referee #2, 11 Feb 2024
This paper introduces a new isoprene SOA multiphase chemical mechanism and provides detailed analysis of the efficacy of the mechanism against both laboratory and field studies under a variety of different conditions. The results are very detailed and thorough with much discussion into both what the model does well and its limitations. However, I think a lack of detail on the specifics of the mechanism led to confusion about what the goal of introducing this mechanism is and how it is advancing the science beyond the currently existing isoprene SOA mechanisms. Overall the work fits well within the scope of ACP, the results are thorough, and the limitations are clear and well-described. I would recommend publication in ACP once a few comments are addressed.
- How were the reactions and species that were included in the model determined? Was the larger MCM reduced or were selected species/reactions added to SAPRC07? Either way additional description on how this was done is warranted.
- Is any comparison possible with machine learning-based reduced mechanisms (e.g. Wiser et al)? If direct comparison is not possible, I think a discussion of the pros/cons of this reduced model compared to machine learning-based mechanisms would be useful.
- More discussion would be helpful on how the mechanisms discussed in this work differ from each other beyond just the size of the mechanism. Since all the mechanisms seem to perform similarly in their ability to predict isoprene SOA, it is important for the authors to indicate what this new mechanism is adding beyond those that already exist.
- In particular, the comparison in Fig 2 seems to show that there is very little improvement in the UCR model from the SAPRC07 mechanism. In light of that, what is the additional chemistry representing better?
- Does the mechanism include gas-phase RO2+RO2 reactions to form dimers? This is mentioned as a possible pathway during the results, but it was not introduced in the methods, which led to some confusion for me.
- The resistor model for uptake coefficient assumes no particle-phase diffusion to uptake, however should IEPOX-SOA form a viscous organic shell around an acidic core, this may affect the uptake. While this would not help the modeled underestimation of IEPOX-SOA, do the authors expect this to impact the results in this work at all?
- In Fig 2, what was considered an “outlier”? Beyond the discussed issue reproducing high NOx experiments, was there any consistency in these outliers that were not represented well?
Minor comment
- Fig 4. Please check the +/-50% lines as one appears to be 1.5:1 and one appears to be 1:2. I assume these were meant to be the same ratio.
Reference
Wiser, F., Place, B. K., Sen, S., Pye, H. O. T., Yang, B., Westervelt, D. M., Henze, D. K., Fiore, A. M., and McNeill, V. F.: AMORE-Isoprene v1.0: a new reduced mechanism for gas-phase isoprene oxidation, Geosci. Model Dev., 16, 1801–1821, https://doi.org/10.5194/gmd-16-1801-2023, 2023.
Citation: https://doi.org/10.5194/egusphere-2024-97-RC2 -
AC1: 'Reply to Comments on egusphere-2024-97', Haofei Zhang, 17 Mar 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-97/egusphere-2024-97-AC1-supplement.pdf
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 393 | 143 | 16 | 552 | 33 | 9 | 8 |
- HTML: 393
- PDF: 143
- XML: 16
- Total: 552
- Supplement: 33
- BibTeX: 9
- EndNote: 8
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Cited
1 citations as recorded by crossref.
Chuanyang Shen
Xiaoyan Yang
Joel Thornton
John Shilling
Chenyang Bi
Gabriel Isaacman-VanWertz
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(4633 KB) - Metadata XML
-
Supplement
(13342 KB) - BibTeX
- EndNote
- Final revised paper