the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 Oct 2024
Implications of Reduced-Complexity Aerosol Thermodynamics on Organic Aerosol Mass Concentration and Composition over North America
Abstract. Atmospheric organic aerosol (OA) mass concentrations can be affected by water uptake through its impact on the gas–particle partitioning of semivolatile compounds. Current chemical transport models (CTMs) neglect this process. We have implemented the Binary Activity Thermodynamics model coupled to a volatility basis set partitioning scheme in the GEOS-Chem CTM, providing an efficient reduced-complexity OA model that predicts relative-humidity-dependent mixing and partitioning thermodynamics while limiting the impact on computational efficiency. We provide a quantitative assessment of this water-sensitive OA treatment, focusing on a subdomain over North America. The updated OA scheme predicts a spatiotemporal mean enhancement in surface-level OA mass concentration of 145 % for January 2019 and 76 % for July 2019 compared to GEOS-Chem's most advanced OA scheme. The temporal mean surface-level OA organic mass concentration can increase by up to ∼ 590 % for January 2019 and ∼ 280 % for July 2019, with the greatest enhancements occurring over the ocean. The updated OA scheme also quantifies the OA-associated water content. The simulations show how different OA precursors and related OA surrogates contribute and respond to water uptake, including due to changes in temperature and relative humidity over the diurnal cycle in selected winter and summer months. These results are independent of future CTM improvements involving updates to chemical reaction schemes and emission inventories. Our water-sensitive OA scheme allows for a better representation of the seasonal and regional variations of OA mass concentration in CTMs.
- Preprint
(6394 KB) - Metadata XML
-
Supplement
(29688 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-2712', Anonymous Referee #1, 21 Nov 2024
In this work, the authors present results from the integration of a recently developed organic aerosol (OA) thermodynamics and partitioning scheme into GEOS-Chem, a widely used chemical transport model. This addition represents a major change to the calculation of modeled OA mass, especially under high humidity conditions, incorporating previously neglected mechanisms into the overall calculation of particulate concentration and composition. On the whole this is a very well composed manuscript. Text is clear and strong, figures are well-composed, and organization is thoughtfully planned. I have only a few major questions and concerns that I would like to see clarified or addressed before recommending publication.
- First, after multiple readings I am still not completely clear on the details of the “RH = 0” dry case comparison. As described in the supplementary information, most figures compare the modified simulation to itself, using RH = 0 conditions as a baseline: “in this work, we compared the water-sensitive OA scheme at RH > 0 % with the water-sensitive OA scheme at RH = 0 %”. Does this mean that the updated version of GEOS-Chem is run again, but with RH in all BAT calculations pinned to 0? Some other model modification? Or is this a filtering process for comparison of actual modeled conditions at low vs. high humidity? Details here should be clarified, considering their importance for the interpretation of all difference plots. If there are multiple simulations being performed (base, modified, modified with RH=0, etc) they should be clearly listed, named, and described in a main manuscript table.
- It appears to me that one major uncertainty in this new implementation of BAT is that of assigned O:C ratios for the binned and simplified GEOS-Chem species. I understand the necessity of the approach taken here (as described in SI), but I have to wonder at the sensitivity of final results to variability in these assumed properties. A series of sensitivity tests using boundary values for reasonable O:C ranges would help to quantify how sensitive model results actually are to this uncertainty and simplification.
- While the modeling results on their own are very interesting and helpful, there is a notable absence in this manuscript of comparison to observations. Of course modeled OA in general has many areas needing improvement, making the comparison a tricky one, but there is value in noting how these mechanism improvements and changes translate to real world comparisons. I think some form of comparison with meaningful observations is a reasonable expectation here, no matter how good or bad the impact on agreement may be.
- It’s not clear to me why the model domain cuts off the west coast of the United States. This should be addressed or more clearly explained.
- The supplementary information seems unusually extensive to me, and includes some figures and descriptions that I think are crucial to the overall manuscript narrative. I recommend looking over this content carefully and considering whether or not some of it should be moved to the main manuscript.
- A 50% increase in wall clock run time is pretty massive, and probably not acceptable for most modelers. The manuscript text mentions possibilities for computational efficiency improvements. Is there any sense of how much these might mitigate the computational cost of incorporating these improvements?
Citation: https://doi.org/10.5194/egusphere-2024-2712-RC1 -
RC2: 'Comment on egusphere-2024-2712', Anonymous Referee #2, 26 Nov 2024
The authors expand the treatment of organic aerosol formation in GEOS-Chem to include consideration of water associated with organics using the BAT approach previously developed in the group. The manuscript is well written. One major comment is that the paper does not have evaluation with observations. The OA evaluation would likely be complicated given many observations are from filter samples that are exposed to low RH during measurement. If an OA evaluation cannot be easily performed, a box model sensitivity simulation demonstrating the effect of measurement artifacts and or box model simulations demonstrating the sensitivity to other model aspects could be useful. After accounting for measurement drying, what would be the "measured" OA? What are the most sensitive parts of the implementation/where should effort be put into improving the base, dry formulation? For example, are ketones a reasonable assumption for functionality? What about the assumption of externally mixed inorganic particles? What about the lumping and binning of a VBS or Odum 2-product (the enthalpy used with those is often much lower than a pure species of the same C*)?
Minor comments:
- Equation 1: How is the molar mass of water included? Line 157 indicates it may be in the sum over k? It seems like the text may need to be better synced with the equation.
- Figure 1e—the outflow from NYC and northeast US seems notably enhanced. What drives the enhancement?
- Line 277 indicates the inorganic electrolytes are phase separated for O:C<0.5. Figure 2 shows O:C is generally high. Does this suggest the electrolytes should be mixed with OA?
- Figure 4: Consider shifting the time axis to run ~midnight to midnight. It took a minute to realize it was another time. Where are 6, 7pm?
- Consider a rename of section 3.2 as model performance is often used in reference to how predictions perform relative to observations.
Citation: https://doi.org/10.5194/egusphere-2024-2712-RC2
Data sets
GEOS-Chem organic aerosol mass concentrations output data, Zenodo Camilo Serrano Damha https://doi.org/10.5281/zenodo.13352426
Model code and software
Binary Activity Thermodynamics-volatility basis set (BAT-VBS) model (Fortran) Camilo Serrano Damha, Andreas Zuend, and Kyle Gorkowski https://doi.org/10.5281/zenodo.8270272
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 160 | 55 | 9 | 224 | 47 | 3 | 3 |
- HTML: 160
- PDF: 55
- XML: 9
- Total: 224
- Supplement: 47
- BibTeX: 3
- EndNote: 3
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1