the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Mar 2024
Effect of Secondary Ice Production Processes on the Simulation of ice pellets using the Predicted Particle Properties microphysics scheme
Abstract. Ice pellets can form when supercooled raindrops collide with small ice particles that can be generated through secondary ice production processes. The use of atmospheric models that neglect these collisions can lead to an overestimation of freezing rain. The objective of this study is therefore to understand the impacts of collisional freezing and secondary ice production on simulations of ice pellets and freezing rain. We studied the properties of precipitation simulated with the microphysical scheme Predicted Particle Properties (P3) for two distinct secondary ice production processes. Possible improvements to the representation of ice pellets and ice crystals in P3 were analyzed by simulating an ice pellet storm that occurred over eastern Canada in January 2020. Those simulations showed that adding secondary ice production processes increased the accumulation of ice pellets but led to unrealistic size distributions of precipitation particles. Realistic size distributions of ice pellets were obtained by modifying the collection of rain by small ice particles and the merging criteria of ice categories in P3.
- Preprint
(6358 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 15 May 2024)
-
RC1: 'Comment on egusphere-2024-594', Anonymous Referee #1, 28 Mar 2024
reply
Review of “Effect of Secondary Ice Production Processes on the Simulation of ice pellets using the Predicted Particle Properties microphysics scheme” by Lachapelle et al.
The authors of this manuscript examined the effect of secondary ice production (SIP) processes on the simulation of ice pellets using an NWP model with a double moment bulk microphysical scheme (P3). Both Hallett-Mossop (HM) and fragmentation of freezing drops (FFD) processes are examined. It is found that adding HM or FFD would significantly improve the simulation of ice pellets for this specific case (both full simulation and 1D idealized simulation). I enjoyed reading this manuscript which is well written, and the results are clearly presented. I made several suggestions below that may help.
Main comments:
L134-136: I believe that the modification of the default HM parameterization in P3 should not be a major obstacle to examine the combined effect of HM and FFD. The interaction between HM and FFD could be quite complex and sensitive to how both HM and FFD are parameterized, which could be a continued study on it alone. Nevertheless, I think it would be worth mentioning in the manuscript about this complexity which warrants further studies.
L137-138/L428-429: the default HM parameterization in P3 scheme with a threshold of 4000 µm for the mean-mass D of ice particles seems extremely large, e.g. some graupels could be much smaller than 4000 µm. Some of the previous studies have disregarded this threshold or using a smaller one. Would it be possible to test different thresholds which might have significant impacts on the results?
L432-433: the maximum number allowed (2×106 m-3) for ice number concentration seems quite small. In situ data suggests that much larger values are possible even without counting those ice particles smaller than ~50 µm. As SIP will produce a large amount of tiny ice splinters, the number concentration might peak locally at a high value. Although the exact maximum value is arguable, 2×106 m-3 seems definitively too low. This means some large Ni will be automatically clipped at this lower value and the total Ni is therefore reduced. I’m wondering if the author tested other thresholds and whether the results are significantly different.
L169: could the authors describe more about the simulation results of using more than 2 ice categories? My understanding is that with more ice categories, the different sizes of ice particles should be better represented. Although for many reasons, such as our limited knowledge of SIP, etc. a better physical model might not produce better prediction results. I believe more discussion on this would be helpful.
Figure 3-5: it seems the best results from nCat2_FFD_MOD still overestimated the period of freezing rain compared to the observation, particularly for UQAM-PK. Might this suggest that current SIP rate in this study is not fast enough to convert liquid into ice?
Fig. B1: nice results from the 1D simulation which illustrates well the impact of modifying the FFD process (section 2.3), e.g. rime ice with similar size to raindrops + much smaller ice crystals!
Other comments:
Fig. B2d: the ice cat 2 (orange line) is missing.
Citation: https://doi.org/10.5194/egusphere-2024-594-RC1 -
RC2: 'Comment on egusphere-2024-594', Anonymous Referee #2, 22 Apr 2024
reply
I agree with the review posted by Reviewer #1. My only additional comment would be that Figure 2 and its related discussion would be aided with some ground truth observations of precipitation type within the shown domain. For example, all snow observation locations would be plotted in the "snow" column of subplots. These observations would ideally come from ASOS, LSR, or crowdsourced mPING reports. Without these observations, it is difficult to determine whether the simulations are improved with the SIP inclusion across the domain (and not just at the small domain of subsequent analyses).
Citation: https://doi.org/10.5194/egusphere-2024-594-RC2 -
CC1: 'Comment on egusphere-2024-594', Heather Reeves, 25 Apr 2024
reply
Review of “Effect of secondary ice production processes on the simulation of ice pellets using the predicted particle properties microphysics scheme” by Lachapelle et al.
Recommendation: Accept with minor edits
Summary: This paper addresses advances to an NWP microphysics scheme that may allow for better prediction of ice pellets (PL). Specifically, it shows that secondary ice production (SIP) appears to have been pivotal for transitioning falling hydrometeors from all liquid to all ice, thus resulting in PL. Two processes that enhance the conversion from liquid to ice are parameterized in this paper (fragmentation of freezing drops FFD and Hallet-Mossep HM). Additional modifications were made to the FFD code to yield more representative results. This is a strong paper. It’s clear and concise and the science is compelling. I have only minor thoughts below.
- Line 134: It says the two SIP processes are studied independently (as opposed to simultaneously including both FFD and HM in the same experiment) because that requires the default implementation of HM to be modified. I don’t understand why HM would have to be modified. It is described as being a different physical process than FFD, so up to this point I thought these were 2 separate processes. Can the authors clarify?
- Line 138: The paper describes the number of ice splinters produced per unit of rain at a certain temperature range and how it changes outward from there. Is there a citation for this is or is this something the authors of this paper prescribed? If the latter, how sensitive are the results to the number of ice splinters?
- I’m confused by the Appendices. It’s not clear to me why they’re included in the paper as appendices. I think Appendix A could be moved into Section 2. I struggled with Appendix B since its first referenced on line 155, before we know anything about the case study or experiments. I think the sensitivity tests in appendix B give some broader context to the results of this paper that merit putting this in the main body of the paper. And, like above, I think the content in Appendix C could just be put in the main part of the paper as well. Both Appendices B and C include good content, but having them as appendices confuses me as a reader. Moving that content into the main body of the paper will strengthen the story line and give greater import to the creative work presented in these parts of the paper.
- Paragraph starting at line 220: I like the air parcel trajectory approach, but I’d like to know to what degree that trajectory bobs up and down in the vertical. I think a simple way to address this is to add an inset to Fig. 1 that shows a vertical cross section along the trajectory that shows the position of the parcels from each experiment as a function of time/location. That way the reader can assess whether the changes made to the microphysics scheme impact the rate at which the parcels are advected and whether their vertical ascent/descent differs.
- I find Figs. 3,4,5 difficult to read. It’s a lot of skinny lines and some colors are difficult to distinguish and some of the lines overlay each other enough to make it hard for the reader. I wonder if the authors would consider converting this to a “chicklet plot” for the ptype forecasts. This would be a lot easier for the reader to interpret. It would require putting the rates in a separate panel, but I think it’s worth the extra real estate to make a clearer graphic.
- It’s interesting to me that in Fig. 6 the precip rate varies between the experiments (this is also evident in Figs. 3-5). Can the authors add some thoughts to the paper on why this is?
- I see there’s MMR data that shows the vertical level at which the transition from FZRA to PL occurs (line 291). Out of curiosity, have the authors tried to reproduce synthetic MMR data from the simulation to see if the transition from FZRA to PL in the vertical is accurately handled by the FFD experiments?
Citation: https://doi.org/10.5194/egusphere-2024-594-CC1
Data sets
Observation data for four ice pellet events Mathieu Lachapelle, Julie M. Thériault, and Hadleigh D. Thompson https://borealisdata.ca/dataset.xhtml?persistentId=doi:10.5683/SP3/TGS5HU
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 250 | 52 | 12 | 314 | 5 | 7 |
- HTML: 250
- PDF: 52
- XML: 12
- Total: 314
- BibTeX: 5
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1