the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 18 Sep 2024
Sensitivity of predicted ultrafine particle size distributions in Europe to different nucleation rate parameterizations using PMCAMx-UF v2.2
Abstract. The three-dimensional chemical transport model, PMCAMx-UF v2.2, designed to simulate the ultrafine particle size distribution, was used to investigate the impact of varying nucleation mechanisms on the predicted aerosol number concentration in Europe. Two basic case scenarios were examined: the original ternary H2SO4-NH3-H2O parameterization and a biogenic vapor-sulfuric acid parameterization. Using the organic-based parameterization, PMCAMx-UF predicted higher N10 (particle number above 10 nm) concentrations over Europe by 40–60 % on average during the simulated period, which is a relatively small difference given the differences in the two assumed mechanisms. Adjusting the nucleation rate by an order of magnitude for both mechanisms led to an average change of ±30 % in N10 for the ternary ammonia case, and ‒30 to 40 % for the biogenic vapor case. In the biogenic organic nucleation scenario, reducing the fresh nuclei diameter from 1.7 nm to 1 nm resulted in reductions in N10 and N100 by −13 % and −1 %, respectively. Incorporating extremely low-volatility organic compounds (ELVOCs) as the nucleating species resulted in predicted increase in N10 concentration by 10–40 % over continental Europe compared to the ammonia parameterization. Model predictions were evaluated against field measurements from 26 stations across Europe during the summer of 2012. Among the tested scenarios, the measurements showed better agreement with the ternary ammonia and ELVOC-based parameterizations for N10, whereas for N100, all simulated cases appear to agree quite well with the field data.
- Preprint
(2127 KB) - Metadata XML
-
Supplement
(1508 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on gmd-2024-145', Anonymous Referee #1, 12 Oct 2024
This paper compares ultrafine aerosol concentrations in Europe in PM-CAMx simulations with two different new particle formation mechanisms. The base case is a scaled ternary NPF parameteriation by Napari et al (2002) involving H2SO4 and NH3. The adjusted case is an organic-inorganic parameterization by Riccobono et al (2014). The base case and model used was previously evaluated in much the same way over the same domain by Patoulias et al (2018). A key finding is that, on the land surface in Europe, the predicted 100nm-particle concentrations do not depend strongly on the particle formation mechanism, at least provided that they are in good agreement with observations.
While the study does not describe ambitious model developments or innovative evaluations compared to previous work by the same authors, the detailed, dedicated study of the sensitivity to simple new particle formation parameterizations is useful and the well-motivated tests were interesting and in some cases novel, for example, the sensitivity to the reduction of the nucleation diameter. The method of adjustment of the particle formation parameterizations to suit the species the model represents is also novel and interesting. The model is state-of-the-art, and the evaluation is thorough for the measurements used, although these are limited to ground sites.
I think the paper could be suitable for publication in GMD if the comments below can be addressed.
Scientific comments
- A key message of the paper is that the detailed mechanism for particle formation does not strongly affect the simulated particle number concentrations. Once this is understood, it then follows that it is not critical that the particle formation mechanisms discussed in the paper are state of the art. Still, I did not understand why the parameterization of Baranizadeh et al was not also used. Perhaps the code is no longer available, but some explanation would be good (even if the answer is not very exciting). Also, other relatively similar but more up-to-date options for inorganic particle formation are now available, for example the parameterization of Yu et al, who presents lookup tables in a GMD 2020 paper. This parameterization includes the potentially important contribution of ions. Similarly, a parameterization of organic-inorganic NPF is available in Lehtipalo et al, Science Advances (2018) which has some similar limitations to Riccobono et al (2014), such as being valid at only one temperature, but at least it is based on chamber experiments in which ELVOC were detected by a mass spectrometer rather than inferred imprecisely. Some discussion of this more recent work might still be useful in the manuscript.
- It makes sense to adjust the rate constant given in Riccobono et al, which is very uncertain as well as not representative of the VBS species in PMCAMx. However, it took me a little time to understand how Chen et al is being used (simply to bound the range of possible nucleation rates for a given sulfuric acid concentration, such that to a first approximation it doesn’t matter which species are forming particles). In Chen et al (2012), the participation of ELVOCs in particle formation might be very different to the current study. Some more discussion of how this might influence the results would be useful. The caption of Figure S1 would ideally say where exactly the data in the plot come from (I think the PMCAMx simulation, not measurements, right?) I think the plot needs a bit more explanation because (if I understand correctly) it’s reversed compared to almost all data analysis anyone ever sees, in that the “fit line” is used to get the individual data points, rather than the other way around. Also, it seems important enough to the study to put it into the main text.
- Some discussion of the size dependence of condensational growth rates, and coagulation rates, in the 1-1.7nm range and its effect would be interesting and welcome in section 4.3.2.
- The NMB results quoted in the conclusion, at lines 424-425, look too good to be true, at first glance. It seems like a different metric might give a more realistic summary of the results (NME?) Looking at the conclusion left me wondering what kind of overfitting or tuning had been done to get NMB values so low given (for example) the large uncertainties in ammonia, BVOC and SO2 emissions that the particle formation and growth rates depend on, and the sparsity of measurements of their concentrations. Then Figures 4 and 5 reassured me that the biases were still more or less as expected from this kind of work at individual stations. As the model underpredicts at some stations and overpredicts at others, it is presumably partly luck that the stations are distributed such that the biases end up cancelling out in the average. Some more discussion of this, and whether or not the results are right for the right reasons, might be useful. Even though a clear message of the paper is that the fine details of the new particle formation mechanism don’t matter very much to the particle number concentrations.
- It would be great to investigate whether the key result, the lack of sensitivity to particle formation mechanism, generalizes to higher altitudes. Even if no evaluation can be done, the particle concentration vertical profiles might still be interesting. In the conclusion (and abstract), should there also be a caveat that the lack of sensitivity to particle formation mechanism may not generalize outside Europe, e.g. to tropical rainforests, deserts, pristine oceans, or extensive ice sheets?
Minor technical or editorial comments
Was Riccobono et al added to the Napari parameterization or simply used in its place? I assume the latter, from later context, but it would be good to be explicit around line 140.
What was the time frequency of simulation output at the EUCAARI sites? Is it instantaneous output or means?
Typo in title of section 4.3.2
Abstract line 20: “Among the tested scenarios, the measurements showed better agreement with the ternary ammonia and ELVOC-based parameterizations for N10”—compared to what? A scenario with no NPF?
Citation: https://doi.org/10.5194/gmd-2024-145-RC1 -
RC2: 'Comment on gmd-2024-145', Anonymous Referee #2, 01 Nov 2024
The manuscript presents an extensive analysis of the sensitivity of two nucleation parameterizations. According to the authors, despite differences in details of the simulations and results, the bottom line is that their results suggest "low sensitivity of ... larger particles (a proxy for CCN) to the details of the nucleation mechanism ...". I think this is reassuring result for the modeling community: Increased complexity may not be as rewarding as anticipated.
Overall, while not flashy, I think this is an excellent manuscript and I fully recommend it for publication as is. The manuscript is well written, well presented, and well argued. Given Reviewer 1 covers most of my other comments, I will make mine brief. I encourage the authors to reflect on the second paragraph of their Conclusions section (which Reviewer 1 and I found interesting). It would be interesting to read more of your thoughts on the implications of this low sensitivity and where we go from here.
Below I list a few minor comments that the reviewers are welcome to respond to and address in the revised manuscript, but these are very minor, and it is okay to ignore them. They are mainly about the vertical coordinate in the used model.
S3: Model Application.
- To be sure, when you say 36x36 km grid, these are perfect squares on a structured grid, right?
- Could you describe the 14 vertical layers in more detail? Are they uniform? Is the vertical coordinate terrain following or not?
- Clarify what you meant by “A rotated polar stereographic map projection was used for the simulations by PMCAMx-UF”
- How were the meteorological inputs used to derive the simulation? They prescribe the fields at each time step? What’s the time step?
- How were the vertical layers of WRF and PMCAMx-UF were unified? Interpolation? Or did you choose the model coordinate to be the same for both?
S4: Results:
- Maybe I missed it, but how did you define ground level? Grid-mean in the first layer? Or at the bottom interface of the first layer? What’s the height of the first layer?
Citation: https://doi.org/10.5194/gmd-2024-145-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 157 | 55 | 149 | 361 | 16 | 3 | 5 |
- HTML: 157
- PDF: 55
- XML: 149
- Total: 361
- Supplement: 16
- BibTeX: 3
- EndNote: 5
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1