the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 28 Jul 2026
Introduction of prognostic aerosols in the P3 microphysics scheme: Simulation of layered Arctic mixed-phase clouds
Abstract. This study presents the implementation of prognostic aerosols within the Predicted Particle Properties (P3) microphysics scheme to improve the physical consistency of aerosol–cloud–precipitation interactions. Two new prognostic number mixing ratios representing "water-friendly" (CCN) and "ice-friendly" (INP) aerosols are added. Aerosols then evolve dynamically and interact with hydrometeors via nucleation, scavenging, evaporation, and sublimation. The impacts of the modified scheme are investigated using convection-permitting (1-km horizontal grid spacing) simulations of an Arctic mixed-phase cloud case observed on 19–20 January 2026 near Inuvik, Canada. Simulations are compared against "in situ" observations, ground-based radar measurements, and retrievals from the EarthCARE satellite. Results show that the inclusion of prognostic aerosols significantly affects cloud microphysical properties and phase partitioning. The prognostic aerosol simulation produces systematically fewer but larger ice particles, leading to increased reflectivity and enhanced riming. It also results in higher liquid water content and lower ice water content compared to the simulation without prognostic aerosols, thereby favoring the persistence of supercooled liquid layers and improving agreement with retrievals. Sensitivity experiments demonstrate that the choice of ice nucleation parameterization plays a dominant role in controlling the cloud phase. Aerosol-dependent formulations produce lower ice crystal concentrations and more persistent liquid water than the original temperature-dependent scheme, thereby enhancing mixed-phase cloud occurrence. The new prognostic aerosol framework improves the representation of aerosol-cloud indirect effects in P3 and provides a pathway toward more accurate simulations of clouds and precipitation, which are critical for radiative balance and weather prediction in polar regions and elsewhere.
- Preprint
(6951 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-4229', Anonymous Referee #1, 27 Aug 2026
-
RC2: 'Comment on egusphere-2026-4229', Anonymous Referee #2, 19 Sep 2026
Overall summary
This manuscript presents the implementation of prognostic water-friendly and ice-friendly aerosol number concentrations within the P3 microphysics scheme and evaluates the new framework using 1-km GEM simulations of an Arctic mixed-phase cloud case observed during a field campaign. The study benefits from an extensive observational dataset, including ground-based measurements and EarthCARE retrievals, and the implementation represents a useful development toward more physically consistent aerosol–cloud interactions in P3. The results show substantial sensitivity of cloud phase partitioning, ice particle properties, reflectivity, and radiative fluxes to the aerosol-aware configuration. I find the model development potentially important and the observational evaluation valuable. However, a major issue is that the primary BASE–AERO comparison simultaneously introduces prognostic aerosols and changes the ice-nucleation parameterization from C86 to DM10. Indeed, the sensitivity experiments indicate that much of the difference in mixed-phase cloud properties is controlled by the ice-nucleation formulation itself. Consequently, the manuscript currently does not cleanly distinguish the impact of prognostic aerosol evolution from that of adopting an aerosol-dependent ice-nucleation scheme. I therefore recommend that the authors substantially reorganize the analysis and conclusions around this attribution issue and provide more quantitative evidence for the incremental benefit of prognostic aerosols.
Major issues
1. The BASE–AERO comparison conflates prognostic aerosols with a change in ice-nucleation parameterization
The manuscript is framed as evaluating the impact of introducing prognostic aerosols into P3. However, the primary comparison simultaneously changes two important things:
BASE: no prognostic aerosols + C86 temperature-dependent ice nucleation
AERO: prognostic aerosols + DM10 aerosol-dependent ice nucleation.The manuscript itself states that the ice-nucleation formulation is changed when prognostic aerosols are enabled. The sensitivity experiments appear to demonstrate that the dominant BASE–AERO differences in cloud phase arise from the ice-nucleation formulation rather than from prognostic aerosol evolution itself. The manuscript notes that simulations using the same nucleation scheme produce similar results. The experimental matrix is actually strong enough to expose this issue: BASE-DM10_Nai1 versus AERO-DM10_Nai1, for example, approximately isolates prognostic aerosol processing while keeping the nucleation treatment comparable.
This creates a mismatch between the paper's framing and what the experiments demonstrate. Statements such as the abstract's claim that “the inclusion of prognostic aerosols significantly affects cloud microphysical properties and phase partitioning” is not matching with the fact that much of that effect may actually be caused by replacing C86 with DM10. The title of the paper does not even include anything related to ice nucleation.
I would ask the authors to reorganize the analysis around attribution. They should explicitly separate:
a) the effect of prognosing/processing aerosols;
b) the effect of changing the ice-nucleation parameterization
Their existing sensitivity simulations may already provide much of what is needed, but the paper currently treats them mainly as a secondary sensitivity analysis rather than using them to establish causality.
2. What is actually gained from prognostic aerosols needs to be demonstrated quantitatively
Related to Issue 1, I am not yet convinced that the manuscript clearly establishes the incremental value of making aerosol concentrations prognostic.
The authors show physically reasonable evolution of droplet and ice particle concentration depletion associated with activation/nucleation/scavenging and replenishment through evaporation and sublimation. The aerosol variables are initialized from coarse, multi-year climatological fields and subsequently evolve through the microphysics.
But the key question for this paper should be:
How much improvement comes from the prognostic evolution of aerosols compared with simply using a reasonable fixed CCN/INP concentration with the same aerosol-aware nucleation parameterizations?
The existing fixed-ice nucleation experiments are very useful for answering part of this question, but the comparison needs to become a central quantitative result.
For example, I would like to see domain/track-averaged metrics for LWC, IWC, ice number, liquid-layer occurrence, reflectivity, effective radius, etc., comparing AERO versus the best corresponding fixed-aerosol experiment. At present, much of the argument is qualitative.
3. The aerosol initialization and representation are probably the largest physical limitation and deserve deeper discussion
The two aerosol populations are initialized using monthly climatological fields derived from 2001–2007 global simulations at 0.5° × 1.25° resolution. Surface emissions are implemented through relaxation toward climatology.
For a 1-km Arctic case study in 2026, this is a substantial limitation. There is no apparent observational evaluation of the simulated aerosol/CCN/INP state during the event. Thus, although the aerosol concentrations are “prognostic” after initialization, their absolute state may not represent the actual aerosol environment encountered by these clouds.
I suggest the authors more clearly distinguish between prognostic aerosol processing and realistic prediction of the aerosol state. At present, “prognostic aerosols” could give readers the impression that the simulation has a physically constrained, event-specific aerosol field.
I would also like to see sensitivity to initial aerosol concentrations, particularly since Arctic INP concentrations are central to the paper's conclusions.
4. Evaluation should be more quantitative
The observational dataset is one of the strongest aspects of this manuscript, but the evaluation remains surprisingly qualitative.
The authors frequently use terms such as “reasonably well represented,” “better reproduced,” “improving agreement,” etc. For example, the simulations are compared against EarthCARE-derived IWC/CWC and effective radius and against radar reflectivity, but the paper relies heavily on visual comparison.
Since the manuscript claims that AERO improves agreement with observations/retrievals, I recommend adding quantitative metrics. Even relatively simple statistics—bias, MAE/RMSE, and percentage changes —would substantially strengthen the conclusions.
5. Single-case nature of the study limits the broader claims
The paper uses one 24-h Arctic event. The authors recognize this limitation, which is appropriate. But some statements—particularly the final suggestion of improved simulations of clouds/precipitation “in polar regions and elsewhere”—extend beyond what this experiment demonstrates. The abstract currently ends with quite a broad implication and needs to be constrained.
Also, the manuscript has preliminary multi-case results, but they are “not shown.” If those simulations are already available, even a compact statistical analysis across several PONEX cases would add enormous value. It need not reproduce the observations. A simple multi-case assessment of cloud fraction, LWP/IWP, reflectivity, precipitation, or radiative fluxes could establish whether the sign and magnitude of the effects seen here are robust.
6. CCN effects seem underdeveloped compared with INP effects
The manuscript introduces both water-friendly and ice-friendly prognostic aerosols, so the reader initially expects investigation of both aerosol–liquid and aerosol–ice pathways. Yet the results are overwhelmingly controlled by INP/ice nucleation.
The authors themselves report that the original and aerosol-dependent cloud-droplet nucleation formulations produce similar results for this case because vertical velocities are weak. This does not mean the prognostic CCN is not important. It changes the interpretation of the paper. This case primarily demonstrates the importance of INP-dependent ice nucleation in Arctic mixed-phase clouds, rather than the full prognostic aerosol framework.
I suggest making this distinction much clearer. Either the manuscript should narrow its claims accordingly, or it should provide stronger evidence of what the prognostic CCN treatment contributes.
7. Radiative impacts need more physical interpretation
AERO changes outgoing longwave radiation by roughly 10–45 W m⁻² and cloud-top temperatures by several degrees. These are quite substantial responses for what is presented as a microphysical modification.
The paper attributes them largely to changes in cloud phase, cloud-top height, and particle properties, which is plausible, but the causal chain deserves more analysis.
I would encourage the authors to connect:
INP → ice number → particle size/sedimentation → liquid/ice partitioning → cloud vertical structure/optical properties → radiation.
This would turn the paper from primarily a model-development/evaluation study into a stronger process-oriented contribution.
8. Literature review for similar treatment of the prognostic aerosol treatments such as two-moment aerosol approach with prescribed initial aerosol SD, or emission rate (without full chemistry/aerosols) is missing. The review work Fan et al. (2016) listed a few. There are also more recent ones like Lin et al. (2022).
Lin, Y., Fan, J., Li, P., Leung, L.-R., DeMott, P. J., Goldberger, L., Comstock, J., Liu, Y., Jeong, J.-H., and Tomlinson (2022), J. “Modeling impacts of ice-nucleating particles from marine aerosols on mixed-phase orographic clouds during 2015 ACAPEX field campaign,” Atmos. Chem. Phys., 22, 6749–6771. https://doi.org/10.5194/acp-22-6749-2022
9. DeMott et al. (2010) parameterization was updated to be DeMott et al. (2015) parameterization so the updated version should be applied (see Lin et al, 2022 and the reference therein provided above). Also, it is developed for immersion freezing, which means each formed ice particle should freeze an existing droplet. Does the implementation follow this?
Citation: https://doi.org/10.5194/egusphere-2026-4229-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 270 | 137 | 36 | 443 | 30 | 27 |
- HTML: 270
- PDF: 137
- XML: 36
- Total: 443
- BibTeX: 30
- EndNote: 27
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Please find the detailed comments in the attached file.