the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 16 Sep 2024
Characterization of fog microphysics and their relationships with visibility at a mountain site in China
Abstract. Enhancing the understanding of fog microphysical processes is essential for reducing uncertainty in fog forecasts, particularly in predicting fog visibility and duration. To investigate the complex interactions between aerosols and fog microphysics and their impacts on visibility degradation, simultaneous measurements of aerosol and fog microphysical characteristics were conducted from April to May, 2023 at a mountain site (1483 m) in the Yangtze River Delta (YRD) region, China. During this campaign, 8 fog events were explored, revealing significantly higher fog droplet number concentrations (Nd) compared to those observed in clean areas. A strong correlation was found between pre-fog aerosol number concentration (Na) and the peak Nd of each fog event, indicating the substantial influence of pre-existing aerosol levels on fog microphysics. Water vapor supersaturation ratio (SS) within fogs was estimated to 0.07 % ± 0.02 %, slightly higher than previous estimates in urban and suburban areas. The broadening of the droplets size distribution (DSD) at stages of formation, development, and mature were dominantly driven by activation, condensation, collision-coalescence mechanism, respectively. This evolution process often led to a shift from unimodal to trimodal DSD, with peaks observed around 6, 12, and 23 μm. For fog events occurring under high Na background, a notable decrease of temperature during mature stage promoted a secondary activation-dominated process, resulting in the formation of numerous small fog droplets and reducing large droplet size. The evolution of DSD can significantly influence visibility (VIS) in fogs. Detailed comparison of several visibility calculation methods suggests that estimating visibility based on the extinction of fog droplets only led to considerable overprediction when 100 m < VIS ≤ 1000 m. The results highlight the necessity of incorporating both fog droplet and aerosol extinction in fog visibility forecasts, particularly in anthropogenically polluted regions.
- Preprint
(1322 KB) - Metadata XML
-
Supplement
(369 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-2264', Anonymous Referee #1, 19 Sep 2024
Eight fog events are observed and analyzed in this manuscript, with a focus on the characterization of fog microphysics and their relationships with visibility. This is a meaningful study that will likely attract the attention of ACP readers. However, I struggled with the manuscript for the following reasons:
Major comments
1. Analysis of Pre-Fog Aerosols
In Section 3.2, the authors explore the relationship between pre-fog aerosols and fog droplets. Under stable conditions, this relationship is logically sound due to weak wind speed. However, the article reports that wind speed during observation is relatively high (4 to 8 m/s), which suggests that advection plays a significant role in these fog events. The authors also state that "the pre-fog aerosols measured at the observation site may not fully represent the particles that actually activated into fog droplets." This raises the question: Can pre-fog aerosols be reliably replaced by aerosols observed during fog? The rationale behind this needs further explanation. Additionally, how does Section 3.2 lay the foundation for the subsequent content? The logic in Section 3.2 should be clarified.
In Section 3.3, pre-fog aerosols are used in the estimation by the κ-Köhler equation. How can the authors be certain that the pre-fog aerosols and those that activated into fog droplets share similar physical and chemical properties? For instance, fog event E3 had a long lifetime. Are the changes in aerosol physicochemical properties negligible? Observing supersaturation in fog is challenging, and bias is inevitable. The authors should discuss the sources of errors in this algorithm and provide references to support this approach. Wang et al. (2021) can be referenced.
2. Mechanism in Fog Event E3
The authors note that "the main wind speeds ranged from 4 to 8 m/s" in lines 157-158, indicating that advection influences the observations. In lines 256-258, they state, "The enhanced supersaturation facilitated the further activation of smaller particles that were un-activated during the SSQ1 stage, resulting in a secondary activation-dominated process during E3." Does this imply that un-activated aerosols from the SSQ1 stage remained stationary without being affected by advection? This statement is confusing and potentially misleading.
The authors also mention "excess water vapor" in line 258. However, Figure 4 shows an increase in supersaturation from the SSQ1 stage to the SSQ2 stage during E3. Does lower supersaturation correspond to excess water vapor during the SSQ1 stage? Please clarify this analysis.
In line 261, the authors discuss the "evaporation of liquid water from previously formed large fog droplets." Both large and small droplets are affected by evaporation, but small droplets are more susceptible to dry air because of a larger surface area concentration. The authors only mention large droplets in this context. Moreover, under the influence of advection, even if previous large droplets evaporate, they may not affect current observations. Is this correct? I suggest revising the analysis to clarify the mechanism.
Minor Comments
1. There is a formatting issue. When there is no space before a paragraph, a blank line should be inserted between consecutive paragraphs (e.g., a blank line is needed between lines 42 and 43). Alternatively, please refer to the formatting style of articles already published in ACP.
2. In line 37, the article focuses on mountain fog; there is no need to mention maritime fog in the introduction.
3. Distinction Between Clean and Polluted Backgrounds
In lines 159-163, the authors differentiate between clean and polluted backgrounds based on fog microphysical properties. However, the distinction between clean and polluted backgrounds should be based on aerosol concentration, as fog microphysics are also influenced by meteorological conditions. The concentration of cloud condensation nuclei (CCN) at the same supersaturation level would be more appropriate for this distinction. Numerous studies, such as Figure 2 in Wang et al. (2024), provide CCN concentration data under different background conditions.
4. In Section 2.1, the authors mention that the observation site is far from Hangzhou but claim that the site is generally near the top of the planetary boundary layer (PBL) around midday based on the PBL height of Hangzhou. This is unreliable because the boundary layer height varies by location.
5. The installation of instruments is important for observation results. Could you provide photos of the observation setup in the supplement? This would help readers better understand the instrument installation.
6. In line 145, the threshold involved in the definition of fog requires a reference for support.
7. The information in the figures should be clearly explained. For instance, there is a lack of explanation for Dp in Figure 1; Q1 and Q2 are not explained in the title of Figure 6. Please check other figures.
8. In line 158, there is an "s" at the end of "speeds." Is speed a countable noun?
9. Water Vapor Consumption in Line 218
The hygroscopic growth of aerosols affects the water vapor mixing ratio, but temperature directly influences the saturated water vapor mixing ratio, not water vapor itself. The authors mention only water vapor consumption in line 218. Please reorganize the explanation to clarify the mechanism behind the relatively high supersaturation.
10. Definition of Activation Ratio in Line 243
The authors define the Activation Ratio (AR) as "the CCN number concentration at a supersaturation setting of 0.2% relative to the total particle concentration." Why was 0.2% chosen? Please provide a reference to justify this choice.
11. In line 270, why was 880 nm used in this study? Please provide a reference or explanation.
12. In lines 296-299, the “≤” symbol is not in Times New Roman font.
13. Introduction
In line 68, the authors focus on polluted regions. The criterion for distinguishing between polluted and clean backgrounds is aerosol mass concentration, but the authors do not use this threshold to determine whether the observation site is polluted or clean. Describing the background as having high or low aerosol loading would be more accurate. If the authors wish to continue using the terms "polluted" and "clean," they should provide criteria to support these distinctions.
In lines 67-68, The authors emphasize the impact of interactions between aerosols and fog microphysics on visibility (“their impacts on visibility degradation”). However, only the effect of aerosols on visibility is highlighted. What about the influence of interactions between aerosols and fog on visibility? Additionally, while the effect of aerosols on fog microphysics is analyzed in the manuscript, the effect of fog on aerosols is not addressed (Qian et al., 2023). The interactions between aerosol and fog should be more prominently discussed.
14. There are large uncertainties in the aerosol–cloud interactions (ACIs) (Fan et al., 2016). If the conclusion provides novel insights into ACIs based on the findings related to interactions between aerosols and fog, it could significantly enhance the manuscript's appeal and attract more attention.
References
Fan, J., Wang, Y., Rosenfeld, D., and Liu, X.: Review of Aerosol–Cloud Interactions: Mechanisms, Significance, and Challenges, J. Atmos. Sci., 73, 4221-4252, https://doi.org/10.1175/jas-d-16-0037.1, 2016.
Qian, J., Liu, D., Yan, S., Cheng, M., Liao, R., Niu, S., Yan, W., Zha, S., Wang, L., and Chen, X.: Fog scavenging of particulate matters in air pollution events: Observation and simulation in the Yangtze River Delta, China, Sci. Total Environ., 876, 162728, https://doi.org/10.1016/j.scitotenv.2023.162728, 2023.
Wang, Y., Niu, S., Lu, C., Lv, J., Zhang, J., Zhang, H., Zhang, S., Shao, N., Sun, W., Jin, Y., and Song, Q.: Observational study of the physical and chemical characteristics of the winter radiation fog in the tropical rainforest in Xishuangbanna, China, Sci. China, Ser. D Earth Sci., 64, 1982-1995, https://doi.org/10.1007/s11430-020-9766-4, 2021.
Wang, Y., Li, J., Fang, F., Zhang, P., He, J., Pöhlker, M. L., Henning, S., Tang, C., Jia, H., Wang, Y., Jian, B., Shi, J., and Huang, J.: In-situ observations reveal weak hygroscopicity in the Southern Tibetan Plateau: implications for aerosol activation and indirect effects, npj Clim. Atmos. Sci., 7, https://doi.org/10.1038/s41612-024-00629-x, 2024.
-
RC2: 'Comment on egusphere-2024-2264', Anonymous Referee #2, 28 Sep 2024
General
This manuscript presents an observational study of fog microphysics using measurements collected at a mountain site and tests several visibility estimation parameterizations based on in situ data. The results are clearly presented and could contribute meaningfully to short-term visibility forecasting during fog events. I believe the topic is appropriate for ACP. However, I have the following concerns that should be addressed before considering this work for publication.
Major
- Paper structure
By the end of the Introduction section, you should introduction the structure of the remaining of the manuscript.
Figures 4a-4c are methodology while the panel 4d is a result. You may consider to split this figure and move panels 4a-4c up to the method section.
Section 3.5.1: this section presents previous parameterizations of VIS. Part of the text should be moved to Introduction part and part of it should be moved to methodology. This part can also serve as your motivation of testing the parameterizations using measurements from the mountain site. The results and relevant discussion should remain in this section.
- Introduction:
This section needs more work. For example, there is no mentioning of aerosol extinction in the intro part until the very end. 'aerosol extinction' appears abruptly without any information on how it is related to VIS or microphysics. Second, the motivation of the study presented in this manuscript does not seem clear to me. You listed quite several past studies on fog microphysics and VIS, what are their disadvantages or limitations? What are the values of your work will add to the current understanding or parameterization in terms of VIS forecast? Why this work is necessary given the abundant of work have been done in the past?
- Incomplete descriptions of the presented figures and lack of discussions:
You seem to only described Figure 1a in Section 3.1, while there are ample information shown in Figures 1b-1e that should be described and discussed.
From my reading, only Figure 4, Figure 5 and Figure 6 are described in detail (while lacking specific reference to the panels in the main text). The rest of the figures deserve more detailed discussions.
- Grammar errors
I found many grammar errors in the abstract. I tried to capture some of them in my minor comments, but they are by no means a complete list. I did not list any grammar errors in the main text. The authors should do a thorough proof reading before resubmitting.
Minor
Line 17: Clarify whether the elevation of 1483 m is above ground level or mean sea level.
Line 18: Consider rephrasing to, "In this study, eight fog events were investigated during the campaign, ..."
Line 23: Add "and collision-coalescence mechanisms."
Line 24: Rephrase as, "Peaks were observed at around ..."
Citation Format: When citing a reference at the beginning of a sentence (e.g., "Song et al. (2019) found that ..."), you do not need to cite it again in parentheses at the end of the sentence.
Line 164-167, and Equation 1-3: The linear relationship between LWC and Nd within a specific D_eff bin seems expected based on equations 1–3. Since D_eff is the ratio of the third to second moments, it can be treated as particle size, meaning that LWC should increase with higher Nd. Can you clarify or further discuss this?
Lines 183-184: Could the difference in findings between this study and the previous one be due to the different elevations of the measurement sites?
Line 198: When you mention the second approach, do you mean Nd is equivalent to NCCN? Please clarify.
Lines 213-215: Are the studies you compare your results to all focused on fog events, or do any deal with clouds, such as the Gong et al. paper?
Line 225: The VIS during the development stage of the 04/12 event does not appear to decrease at a slower rate compared to the formation stage. Did you apply specific thresholds for the rate of change of VIS to define these stages? If so, please justify how these thresholds were determined.
Lines 230-244: It would be helpful to include specific figure and panel numbers after each discussion sentence, particularly when referring to DSD descriptions, to make it easier for readers to follow along. This is especially important in the case of the E3 event.
Citation: https://doi.org/10.5194/egusphere-2024-2264-RC2 -
CC1: 'Comment on egusphere-2024-2264', Almuth Neuberger, 25 Oct 2024
Comment on the manuscript: Characterization of fog microphysics and their relationships with visibility at a mountain site in China by Liu et al.
Thank you for your interesting research manuscript! We discussed your work within our research group since we are doing similar research and got interested in your findings. We have some remarks and questions concerning the experimental set-up. Many technical details (e.g. on the sampling efficiency of the GCVI inlet) are currently missing and should be added to allow a reliable assessment of the presented results. Moreover, the reasoning behind many of the key findings are often not clear to the reader and some more clarifications (incl. add the right references) would clearly help here.
Below, we have listed a few questions and remarks.
- Our most important comment is the lack of describing the sampling efficiency of the GCVI system. Has it been determined? How well do you sample larger droplets? Have zero-measurements been performed? This is an important task and will have a substantial impact on most of the results and interpretation presented here. (see e.g. Figure S4 and others in Karlsson et al, 2021). At the moment, it is not clear if any particle loss corrections (for the aerosol instrumentation behind the inlets and for the fog monitor) have been done.
- A schematic of the set-up which includes instrument names, inlets, piping, flow rates (or a reference to it) would be very useful to the reader.
- l. 23 and 240: How do you know that it was indeed collision-coalescence? Just because another peak in the size distribution appeared? Could it be that it is just condensational growth? Please elaborate.
- One of the key findings is that secondary activation was observed after additional cooling. However, this is not really clear from the figures and key parameters like wind-direction and speed are not shown.
- Introduction: Please refer to e.g. Elias et al. (2009) and Hammer et al. (2014) who also discussed the contribution of hydrated aerosol to light extinction.
- l. 68: You mention that particle number size distributions were measured but the actual findings/curves (mean distributions and timelines) are never shown. However, it would be useful to add these graphs to the manuscript or SI to better interpret the findings.
- l. 83: Do you have supporting data showing that the site is usually near the top of the PBL?
- l. 132: The CCN counters are usually kept longer at one fixed temperature in order to achieve a stable supersaturation. Have you checked that 1 min is a long enough period? Especially, when switching from 0.7% down to 0.1% we doubt that this will be sufficient.
- l. 149: Consider including the interstitial aerosol number concentration in Tab. 1. Have you performed a closure study to see if Nd and the number of fog residuals agree?
- l. 176: What is the p-value if you talk about significance but only have a few data points? Have you also looked at the size distribution? Has that also changed in the different pre-fog Ntotal conditions? If you have so much more particles than droplets, why would you expect the Ntotal to be correlated with Nd and not just the N100 or even higher? How does the size distribution behind the CVI look like?
- l. 180 and Fig 3: It is not really clear why certain points were excluded. Please explain and reason why the data points after rain events should be excluded. Do you then sample artifacts? In addition, in Figure 3, please state which kind of linear regression has been applied. Since both x- and y-values are prone to errors, you should use an orthogonal regression, which does not seem to be used here.
- l. 184: What does it indicate that your slope values are higher than those measured by Duplessis et al. (2021)? What are the consequences? It would help to elaborate more here.
- l. 214: Please also include the SS values from those four publications you are referring to.
- l. 225: To us, the classification into fog stages seems to only have worked semi-well, especially in E3. How exactly did you divide the fog events and why did you choose this definition?
- l. 228: What concentrations do you consider to be “high” or “low”? It would improve the interpretation to explicitly state the measured concentrations here.
- l. 242ff: Why should the activation ratio of the residuals show that there was scavenging? Please explain in more detail as a correlation doesn’t mean causality. What is the reasoning to define the AR via the CCNC measurement and not via the CPC/SMPS measurements?
- l. 259ff: The statement about the evaporation of large droplets due to the formation of smaller droplets is not really clear. Could you elaborate more here and also provide some more references for this effect?
- l. 270: Why are you using 880nm? Is it because of the visibility sensor that comes with the GCVI? In that case, it should probably also be 3 and not 3.912 in eq. 5 because of how the visibility sensor is calibrated (see manual of the Belfort visibility sensor).
- I would suggest moving the first part of chapter 3.5.1 to the methods section. This is not really results-
- l. 289: Were all data points included when performing the linear regression? It would be helpful to add the result to the figure. Is the slope similar if you only include values below e.g. 1km?
- l. 290ff: Adding a new parameter (here Nd) gives more information and therefore improves the parametrization. Please clarify the last two sentences of this paragraph.
- l. 300: Mie theory should be a good prediction for observed visibility. Make your explanation more detailed.
- l. 254: Please also mark these quasi-equilibrium states in the temporal evolution shown in Fig. 5.
Figures:
- Fig. 1: We would recommend you to choose different colorbars which have a more intuitive and uniform distribution of colors, e.g. ‘Blues’. Having white in the middle of the colorspectra is very misleading. https://journals.ametsoc.org/view/journals/bams/96/2/bams-d-13-00155.1.xml
- Fig. 2: As you calculate LWC by using D and Nd, isn’t the outcome of this figure trivial? Maybe move it to the supplement?
- Fig. 4: please plot dN/dlogD as commonly used. The x-axis should probably be ‘nm’ not ‘um’. Please write somewhere that this is E3 as you later on talk a lot about this specific event.
- Fig. 5: For better comparability, we would suggest to use the same colorbar and axis limits for all events (also Fig. 1 and S5) and use the same axis for Dp and Deff in subplots (c).
- Fig. 6: Please explain in the figure caption what SS Q1 and SS Q2 means. Have you considered plotting one subplot where the 4 average size distributions are plotted on top of each other so that one can more easily see the differences? Why did you choose a linear scale for the diameter (x-axis)? Typos: ‘Development’, ‘Dissipation’
- Fig. 7: How come that you still measure so many fog residuals even though the effective diameter is smaller than the cut-off of the CVI?
- Fig. 8a and eq. 6 don’t match. Which one is the correct value for a?
- Fig. 8a: isn’t the interesting regime the small visibilities when LWC>0? Why did you choose a linear axis and not like in Fig. 8b a log-axis?
- Fig. 8: I would recommend to stick to the notation that you introduced earlier: VIS_K, VIS_KN, VIS_G, VIS_GN
- Fig. S1: typo in the colorbar title: ‘Terrain’.
- Fig. S6 isn’t mentioned in the text.
- Fig. S8: please use dN/dlogD and plot the xaxis on a log-scale. The labels of the 2 curves have been switched: black is dry, blue ambient
- Please be consistent with your statistic parameters and linear regressions in all plots (r/r^2, p)
- Please also be consistent with the time on the x-axis in the different figures and do not shrink and stretch the time. It makes it hard to compare the different fog events.
Elias, T. et al., 2009: Particulate contribution to extinction of visible radiation: Pollution, haze, and fog. Atmospheric Research, 92 (4), 443–454, https://www.sciencedirect.com/science/article/abs/pii/S0169809509000192
Hammer, E. et al., 2014: Size-dependent particle activation properties in fog during the ParisFog 2012/13 field campaign. Atmospheric Chemistry and Physics, 14 (19), 10 517–10 533 https://doi.org/10.5194/acp-14-10517-2014, URL https://acp.copernicus.org/articles/14/10517/2014/
Karlsson, L., Krejci, R., Koike, M., Ebell, K., Zieger, P., 2021. A long-term study of cloud residuals from low-level Arctic clouds. Atmos. Chem. Phys. 21, 8933-8959.
Citation: https://doi.org/10.5194/egusphere-2024-2264-CC1
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 158 | 55 | 68 | 281 | 16 | 4 | 6 |
- HTML: 158
- PDF: 55
- XML: 68
- Total: 281
- Supplement: 16
- BibTeX: 4
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1