the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 16 Dec 2024
Adiabatic and radiative cooling are both important causes of aerosol activation in simulated fog events in Europe
Abstract. Aerosol-fog interactions affect the visibility in, and life cycle of, fog and are difficult to represent in weather and climate models. Here we explore processes that impact the simulation of fog droplet number concentrations (Nd) at sub-kilometer scale horizontal grid resolutions in the UK Met Office Unified Model. We modify the parameterization of aerosol activation to include droplet activation by radiative cooling in addition to adiabatic cooling, and determine the relative importance of the two cooling mechanisms. We further test the sensitivity of simulated Nd to: (a) interception of droplets by trees and buildings, (b) overestimation of updrafts in temperature inversions, which leads to artificially high Nd values; and (c) potential mechanisms for droplet deactivation due to downward fluctuations in supersaturation.
We evaluate our model against observations from the ParisFog and LANFEX field campaigns, building on evaluation described in the companion paper. Including radiative cooling in the activation mechanism improves how accurately we represent the liquid water path and the vertical structure of the fog in our LANFEX case study. However, with radiative cooling the Nd are overestimated for most of the ParisFog cases, and the LANFEX case. The time-averaged overestimate exceeds a factor of four (the normalized mean bias factor exceeds 3.0) in four out of eleven ParisFog cases. Our sensitivity studies demonstrate how these overestimates can be mitigated. Assuming the overestimate affects both radiative and adiabatic cooling, we find that although radiative cooling is more often the dominant source, both cooling sources can sometimes dominate activation.
- Preprint
(1954 KB) - Metadata XML
-
Supplement
(1000 KB) - BibTeX
- EndNote
Status: open (until 27 Jan 2025)
-
RC1: 'Comment on egusphere-2024-3397', Anonymous Referee #3, 31 Dec 2024
reply
This paper is the 2nd in a pair of papers aimed at exploring the factors modulating the concentration of fog droplets (Nd) in an NWP model with prognostic aerosol microphysics. While the 1st paper focused on aerosol activation leading to fog formation via adiabatic cooling, this manuscript examined the role of radiative cooling and several additional factors which can potentially modify supersaturation and aerosol activation. They found some sensitivity to all processes that were tested, with radiative cooling sometimes being greater than adiabatic cooling. Deposition processes were also found to be important. In addition to fog droplet concentration, the authors compared liquid water content, liquid water path, and fog top height from the model to observations from the LANFEX and ParisFog field campaigns.
It is known that fog is a challenging forecast problem. The observations used here have a large degree of variability, and it is difficult to conclude that a particular model configuration performs better for the majority of cases. Nevertheless, the relative behavior of the experiments is informative and the publication of this study adds to the understanding of the model processes required to accurately simulate fog. Recommendation: Accept with minor revisions.
General Comments/Questions:
It is not surprising that radiative cooling is important in these simulations, as the fog type simulated is “radiation fog”. It also seems unsurprising that the AD-RAD experiments overpredicts Nd, if the code used for AD was already modified to give improved values of Nd without radiative cooling. Why not add the radiative cooling to Def-ARG?
The sensitivity of the various cloud properties to changes in the various cooling sources is shown, but is there is any significant change in the spatial coverage of the fog between experiments?
The radiative effects include interaction with both the land surface and the fog, correct? It looks like your analyses are made only during foggy periods, and the droplet activation due to radiative cooling is relatively constant with time in most cases. I was curious if you were able to detect an increase in radiative effects as the fog became optically thicker? – or perhaps then the surface radiative effects would decrease?
Specific Comments:
Line 107: “We use observations from including…” There seems to be something missing between “from” and “including”?
Top of page 10: Should the 2nd equation be subtracting the 2nd term rather than adding? If not, please clarify how “New Nd” can decrease.
How is cloud fraction computed?
Fig. 14: Are there any other characteristics that would help categorize when radiative cooling is dominant?
Line 561: Add a space in “paperfocused”.
Citation: https://doi.org/10.5194/egusphere-2024-3397-RC1 -
RC2: 'Comment on egusphere-2024-3397', Anonymous Referee #1, 06 Jan 2025
reply
Review of “Adiabatic and radiative cooling are both important causes of aerosol activation in simulated fog events in Europe” by Ghosh et al.
This study adds to a growing body of work that suggests radiative cooling is an important source of supersaturation for activation of droplets in fog. As with the companion paper, I felt that some sections could be shortened, but overall it is another straightforward comparison of simulations and observations. I have only a few comments.
- References should be included for the first two sentences of the introduction.
- A lot of time is spent showing results for Def-ARG. Since the difference between Def-ARG and AD was already extensively covered in the companion paper, I would ask the authors to consider whether these simulation results could be mostly or entirely removed. Since not all simulations are shown in every analysis due to clutter problems if all are shown, removing Def-ARG would allow the authors to show another simulation, perhaps AD-RAD-SED.
- Figure 5 – the legend uses “Mod-Kappa” from the companion paper rather than AD.
- The authors make a few comments about how the AD and AD-RAD simulations match the observed fog top well. I have a hard time seeing this. It seems like the simulated fog top is 20-50m higher than observed. Perhaps that is not a lot, but it seems substantial when the fog top is only at 50m.
Citation: https://doi.org/10.5194/egusphere-2024-3397-RC2
Data sets
Adiabatic and radiative cooling are both important causes of aerosol activation in simulated fog events in Europe Pratapaditya Ghosh, Ian Boutle, Paul Field, Adrian Hill, Marie Mazoyer, Katherine J. Evans, Salil Mahjan, Hyun-Gyu Kang, Min Xu, Wei Zhang, and Hamish Gordon https://doi.org/10.5281/zenodo.14005081
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 162 | 25 | 8 | 195 | 24 | 3 | 3 |
- HTML: 162
- PDF: 25
- XML: 8
- Total: 195
- Supplement: 24
- BibTeX: 3
- EndNote: 3
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1