Sensitivity of convective cloud and rainfall responses to urban aerosol emission strengths

Keil, Friederike; Quante, Markus; Heinold, Bernd; Matthias, Volker

doi:10.5194/egusphere-2026-2934

Preprints

https://doi.org/10.5194/egusphere-2026-2934

Preprints

19 Jun 2026

| 19 Jun 2026

Status: this preprint is open for discussion and under review for Atmospheric Chemistry and Physics (ACP).

Sensitivity of convective cloud and rainfall responses to urban aerosol emission strengths

Friederike Keil, Markus Quante, Bernd Heinold, and Volker Matthias

Abstract. Urban aerosol emissions can modify convective precipitation through aerosol-cloud interactions, yet the sensitivity of convective events to emission strength and its detectability against meteorological variability remain poorly constrained. We investigate aerosol-cloud-precipitation interactions using the COSMO-DCEP-MUSCAT model system with direct aerosol-cloud coupling, systematically varying urban emission strengths across a wide scaling range for two differing mid-latitude convective cases. Lagrangian trajectory analysis tracks air volume transport from urban sources into convective regions, complemented by ensemble simulations to separate emission-induced changes from natural meteorological variability.

Results reveal a strong dependence on meteorological conditions. Under moderate convective instability, high emission scaling factors trigger ice-phase invigoration through enhanced ice nucleation and latent heat feedback, producing substantial increases in ice content and precipitation intensity accompanied by spatial expansion and temporal shifts. Emission-response analysis reveals nonlinear behavior, with low-to-moderate scaling factors producing no systematic response while high scaling factors trigger substantial microphysical modifications. Under stronger convective instability, even the highest emission scaling factors produce only modest microphysical responses and spatial redistribution rather than intensification, as strong background dynamics buffer aerosol-induced impacts. Ensemble analysis demonstrates that robust signals emerge only for selected variables during specific convective stages, with most responses remaining comparable to natural atmospheric variability.

These results indicate that detectable precipitation responses require strong emission perturbations under favorable meteorological conditions, with effects manifesting as intensity changes and spatial and temporal redistribution. The convective regime critically determines the magnitude and detectability of aerosol effects, with implications for understanding urban impacts on precipitation across different atmospheric environments.

Received: 21 May 2026 – Discussion started: 19 Jun 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 7593 KB)

Supplement (2377 KB)

Download & links

Friederike Keil, Markus Quante, Bernd Heinold, and Volker Matthias

Status: open (until 01 Aug 2026)

Post a comment Subscribe to comment alert

RC1:
'Comment on egusphere-2026-2934', Anonymous Referee #1, 20 Jul 2026 reply
Keil et al. investigates the sensitivity of convective responses to urban aerosol emissions using ensembles of convection-resolving simulations for two meteorological cases and analysis of microphysical changes along Lagrangian trajectories. The main findings of this work were that the microphysical response to urban emissions was highly nonlinear and dependent on convective regime.
Overall, this study has a very thoughtful experimental design; I especially appreciated the use of Lagrangian trajectories for the analysis and the use of ensemble simulations to test whether the aerosol signal rises above meteorological noise. However, I have a few questions (mostly clarifications) regarding the methods and analysis, that I recommend be addressed prior to publication.

Major comments:
The emissions scaling factors range from x0 to x5000. Are these a realistic range? The authors should contextualize these more, because right now it is unclear whether these emissions are outside the range of plausible scenarios. For example, would the URBANx5000 put emissions at the level of e.g., highly polluted megacities, or is it a purely hypothetical extreme case? It would also help to add a value of mean near-surface concentration in Table 1. I appreciate the authors’ reasons for using a scaling factor rather than spatially uniform CCN as is commonly done, but adding this information would help to compare the results of this work to previous studies.

Does the trajectory analysis use the same trajectories across all simulations, or are trajectories run for each specific simulation? Based on line 153 it seems like they are run for each simulation, but this isn’t explicit. Does “ensemble member” refer to the five simulations with the same aerosol emissions, or to all the simulations?

Simulation Design:
The use of saturation adjustment should be highlighted more, even as early as Section 2.2. As the authors acknowledge in the conclusions, this likely has a big impact on the estimated aerosol effects; with saturation adjustment, these may represent more of a lower-bound on potential aerosol effects.

The authors refer to the simulations several times as convection-resolving at 1km. I would argue that at 1km, these simulations are not fully resolving individual convective cells, especially in the more weakly forced convection in Case I. Not suggesting the authors need to rerun all these simulations at finder grid spacing necessarily, but potential impact of grid spacing should at least be acknowledged.

Might have missed this in Section 2.2, but what is the vertical grid spacing of D2?

Figure 9/14 are unclear. Does the ordinate show ratios of each ensemble member relative to the mean of the baseline ensemble? Do BASE and URBANx500 simulations have statistically significant differences only if the error bars do not intersect 1? Also, are the values used here spatial/temporal means for each ensemble member? To show whether the aerosol effect emerges above background meteorological noise, it would be more effective to show the full distributions of both BASE and URBANx500. These ensemble runs are not fully used throughout the rest of the manuscript, which seems like a missed opportunity to strengthen the process analysis. For example, Figures 6-7 could include stippling for panels c/g to show where the ensemble of URBANx500 is systematically different from BASE.

Minor comments:
Do all the simulations have a similar aerosol concentration/profile by the time the air mass trajectories reach the area where storms initiate? Though only the trajectories passing below 1km are included by construction, the aerosol profile may be different in e.g., BASE vs. URBANx5000, which could lead to differences beyond just those driven by magnitude.

lines 196-199 + lines 202-208: Here and in several other places, the statements on spatial extent are quite qualitative — can the authors quantify the area expansion of raining region? Related to major comment #4, do these trends hold across all the ensemble members? This would be especially relevant for Case II, where the results are mostly related to spatial redistribution.

lines 320-322: How are the early vs. later time periods selected? Same with the lower vs. higher altitude.

lines 455-456: What is meant by “the number […] of detectable aerosol-induced modifications”?

Reply
Citation: https://doi.org/10.5194/egusphere-2026-2934-RC1
RC2: 'Comment on egusphere-2026-2934', Anonymous Referee #2, 22 Jul 2026 reply

Please see the attached pdf file.

Reply

Citation: https://doi.org/10.5194/egusphere-2026-2934-RC2

Friederike Keil, Markus Quante, Bernd Heinold, and Volker Matthias

Supplement

https://doi.org/10.5194/egusphere-2026-2934-supplement

Friederike Keil, Markus Quante, Bernd Heinold, and Volker Matthias

Viewed

Total article views: 58 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
38	16	4	58	4	4	4

HTML: 38
PDF: 16
XML: 4
Total: 58
Supplement: 4
BibTeX: 4
EndNote: 4

Views and downloads (calculated since 19 Jun 2026)

Month	HTML	PDF	XML	Total
Jul 2026	38	16	4	58

Cumulative views and downloads (calculated since 19 Jun 2026)

Month	HTML	PDF	XML	Total
Jul 2026	38	16	4	58

Viewed (geographical distribution)

Total article views: 50 (including HTML, PDF, and XML) Thereof 50 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 29 Jul 2026

Short summary

Using simulations of two differing rain events, we investigated how sensitive convective precipitation is to urban emission levels by varying emission strengths across a wide range. Low emissions produce only modest effects, but high levels can trigger systematic changes, including substantial increases in ice content and precipitation intensity accompanied by spatial and temporal shifts. The magnitude of these responses depends critically on the prevailing meteorological conditions.


Total:	0
HTML:	0
PDF:	0
XML:	0