This paper presents observations of radiation fog at a site in Poland made over the course of three nights. The emphasis is on the vertical properties of the fog layer, with the main novelty being the use of the OPC-N3 instrument to measure the height variation of the droplet spectrum. Radiation fog is optically thin, but may develop into a deeper optically thick fog. Predicting this transition is an important forecasting problem and various criteria have been suggested to diagnose this transition; these are assessed using the data obtained here.

I think the paper is worthy of eventual publication but that major revisions are necessary before it can be accepted.

Major Comments

1.    The radiation calculations. The greater part of the paper is concerned with presenting the observations themselves. It is useful to have more observations of (radiation) fog, especially when they include profiles of microphysical quantities. These, I think, largely stand on their own and make the paper of interest to a broad audience.

A less significant part of the paper consists of a comparison of the observed radiative fluxes with fluxes calculated with a radiation code using the observed fog properties. As a check on closure, as noted on L18-19, this is useful, but I would question what is being checked here. The implication of this statement is that the microphysical measurements are being used to check the radiation ones. Especially given the large adjustments made during calibration in Eqs. 1-3, I would argue that the uncertainties in the microphysical measurements are greater than those in the radiative fluxes, so it is more of a check of the consistency of the microphysical measurements and their processing.

The separation of the discussion of the initialization of the radiation code in section 3.3 and the discussion of the results make it difficult for the reader to follow the argument. Some details are missing; for example, what was the surface albedo and how was the asymmetry parameter of the aerosol determined?

In the case of LW radiation, the variation of the downward flux through the day is reltaively small (approximately 330 - 370 Wm-2), but the temperature profile is obtained from soundings made at a different location at only two times during the day. This leads to uncertainties that are comparable in size to the variations in the fluxes and makes the regressions shown in Eqs. 21 and 22 rather meaningless. I would argue that the absolute values of the LW fluxes are poorly constrained, and that attention should be focused on the differences between the two local observing sites, indicating the impact of the fog layer.

2.    Numerous small corrections need to be made in the text. Whilst many of them are trivial and can be corrected by careful copy-editing, there are a number that should be corrected at source, such as the sentence on L128 that is almost identical to that on L125, or the disagreement between Eq. 11 and the equivalent equation in Fig. 7. In a few cases, the meaning was unclear. I have tried to provide a fairly full list in the detailed comments below, but it is not exhaustive.

Detailed Comments

1.    Title. characteristics instead of characteristic.

2.    L11. Delete "amount of".

3.    L12.  This sentence is ungrammatical and I don't know what it means. I suspect that one of "where" or "were" should be deleted.

4.    L14. The figure of 150 Wm-2 is a particular value of the difference in radiative flux due to the fog. This should be in a separate sentence. As written, it is not clear whether a temporal or a vertical difference is intended. I think the intended meaning is that the difference in the net LW flux across the fog layer at any one time may amount to 150 Wm-2.

5.    L15. This will be true if the fog is optically thin.

6.    L27. More precisely, I think this means that poor visibility is involved in 30-35% of delays; it's not clear what "of the time" refers too, surely not all flights. The sentence needs clarifying.

7.    L 45. Delete the initial "The".

8.    L46. This should be something like "if and when a transition from optically thin to optically thick fog occurs."

9.    L47. This should be LW_N or LW_Net, not Delta LW, because the actual value of the flux is being considered.

10.    L50. A continuous tense is inappropriate here. It should be something more like, "They used temperatures at 25 and 50 m to check this condition."

11.    L54. It's not clear what the inequalities are supposed to represent and they are stated in opposing senses. I think you should say something like "Waersted et al. (2017) proposed that a transition to optically think fog occurs when LWP > 30 gm-2, but Costabloz et al. (2024) suggested a value of 15 gm-2 to match more closely the time when the other criteria are met.

12.    L58 "for example" is redundant after "such as".

13.    L75. gm-2h-1 is not the unit of sensible heat flux. You should say that the sensible heat flux is equivalent to the formation of cloud water at a rate of 30 gm-2h-1.

14.    L79-82. These references must be in brackets.

15.    L96. They instead of He.

16.    L98. Insert "a" before "tethered".

17.    L103. I think this should be "The second fog case, with six soundings, consisted of..."

18.    L110. This is the first point at which OPC-N3 is mentioned, but there is no statement about what it is until L.170. You should say that it is a particle counter.

19.    L117. Please give a reference for Fu and Lou's radiative transfer model.

20.    L120. "at two sites in Strzyzow" or "near Strzyzow".

21.    L122. "Numerical simulation was"

22.    L128. This repeats the previous sentence.

23.    L131. "ceilometer"

24.    L141. "in the valley"

25.    L142. "launching" not "lunching".

26.    L156. "Measurements were made using two meteorological balloons filled with helium. The balloon was tethered using a Vaisala TTW111 winch..."

27.    Section 3.1 repeats material from section 2.3

28.    L177. "...measurements of radiation fog were made in Strzyzow..."

29.    Figure 3. The caption should be "The figure shows when the soundings were made..."

30.    L190. "breaks" instead of "brakes".

31.    L194. "During this campaign, the ShadowGraph...firstly to calibrate OPC-N3 and secondly to monitor the..."

32.    L200. Why are the differences so large? https://amt.copernicus.org/articles/16/2415/2023/ suggests a factor of 2. It would be useful to include a brief explanation of the origin of the calibration factor for readers who do not consult Nurowska et al. (2023).

33.    L205. This repeats L188.

34.    L221. A comma is needed after alpha.

35.    L245. Are not these forms of alpha scalings rather than deviations?

36.    L256. "short waves" should be "shortwave bands". Similarly for "long waves".

37.    L259. I wondered how you chose the resolution for the radiation calculations. In practice, so long as you are interested in the fluxes at the top and bottom of the fog layer, rather than heating rates within the fog, this is probably not too crucial.

38.    L265. It is not clear how you have chosen the constant value. Whilst the thermal wavelengths are a bit too large for geometrical optics to apply, you could calculate a mean value of the effective radius by using the result from geometric optics, tau= 3 LWP/(2 rho_w r_e), so LWP/mean r_e should be equal to the integral of LWC/r_e through the cloud.

39.    L288. "where" should be "were". Also I don't think "oscillating" is quite the right word: there is no physical mechanism with a restoring force.

40.    L297. "valley" instead of "volley".

41.    L301. "formed" instead of "was starting".

42.    L302. "in the valley"

43.    L303. "disappeared" instead of "was disappearing".

44.    L317. This paragraph is somewhat unclear. I think the argument is that in all these cases the fog is thin, so firstly, this is not inconsistent with the threshold of 30 gm-2 for thick fog proposed by Waersted because in there cases the LWP never exceeded that value; secondly, however, the LWP frequently exceeded the threshold of 15 gm-2 proposed by Costabloz, so this value is too low to be used as a criteria to distinguish cases of thin and thick fog. The paragraph should be rewritten.

The more general question is when you would regard a fog as thick (c.f. L541).

45.    L322. "until it dissipated" instead of "till disappearing".

46.    L327. I believe the region you are referring to is that below 3 m that is omitted from the figure, because the influence of the ground is not very visible in Fig. 7.

47.    L341. "its maximum"

48.    L346. "The fog was deeper with a top at 102 m." instead of "The fog deepen, it top was at 102 m." 

49.    L357. "dropped" instead of "drooped". Similarly on L360.

50.    Figure 7. The meaning of the dashed black line on the plot of LWC is not explained.

51.    Equation 11. This is not consistent with the version on the figure.

52.    Equation 12. Why is this written as "3.00 h 10^-2", rather than simply as "0.03 h".

53.    L401. "on the night" and "looks different".

54.    L445. The word "Apart" is a bit confusing here. I would just say that the normalized vDSD is shown in panel b.

55.    L479. Figures 13 and 14 are introduced before Figure 12. It would be helpful to renumber the figures. Please say exactly which flux you are considering, at least at its first occurrence, rather than just referring to "the SW flux".

56.    L511. "the agreement".

57.    L512. "RMSE does not exceed ..."

58.    L517. "It became positive after sunrise." should be a separate sentence.

59.    L520. "became" for "become".

Nurowska et al have utilized their prior development and validation of the OPC-N3 monitor for suspended particle sizes for use on sounding balloons. This has allowed them to investigate profiles of radiation fog on three nights in Poland and compare LWdown data just above the city of Strzyzow (at 260 m) with an AERONET site at 445 m. They have investigated particle sizes, densities and derived liquid water contents and related these indices to the radiation balance during fog events. These are interesting and useful data and should motivate others to perform similar investigations. A limitation of the study is the restriction to just three consecutive nights so it is not possible to assess the influences of changing aerosol type and abundance or seasonality on their conclusions. Nevertheless, I recommend acceptance of the paper with changes. I have indicated lots of minor changes intended to help with the written English. Note that there are some generic issues listed before the line by line notes (e.g. grater > greater).

Issues:

A) The regression in Fig. 12h looks totally unconvincing even though an equation is provided (equation 22). In fact all equations in the text and in figures should be provided with both the value for Pearson’s r and P. Note calculation of the P value takes into account the number of values used. Hence a reasonable value for r (e.g. >0.5) is not necessarily associated with a relationship that can be distinguished from regressing random numbers (which can be inferred if e.g. P < 0.05).

B) Fog phases v fog stages: Line 191 refers to ‘fog phases’, but Lines 334 onwards describe ‘fog stages’. Choose the terminology ‘phase’ or ‘stage’ and use this throughout.

C) Normally visibility measured at 2 m above the ground is used to define the onset of fog with fog defined as visibility <1 km. The authors subdivided fog events into phases/stages (Line 191), but they don’t seem to have used their own visibility measurements. How do the fog development phases relate to visibility at 2 m (which would have been measured at the time of balloon launch by the TFMini instrument whenever the OPC-N3 was used – according to lines 181 and 184). I suggest a figure is added to the Appendices showing 2 m visibility data for the three fog events in relation to the fog stages - or add the visibility observations to figure 5.

D) The summaries of fog development for each night of observations (in Sections 4.2.2, 4.2.3 and 4.2.4) are very difficult to follow using the profiles shown in Figures 7, 8 and 9. This is because in the figures the same colour is used for every profile of the same variable. Different variables are shown with different colours. Instead it will be far easier for the readers to follow the written summaries of fog development if, for every variable there is a set colour for each stage of fog development (e.g. red for development, grey for mature, blue for disappearing).

E) Lines 566-567 ‘At the bottom of the fog the smallest droplets evaporate.’ What is the evidence for this? Evaporation of small droplets is feasible after sunrise from the top of fog layers but not the bottom. Instead Weedon et al. (2024, QJRMS, https://doi.org/10.1002/qj.4702) argued that inception of radiation fog is determined by creation of suspended droplets that is faster than their removal by occult deposition (direct deposition onto vegetation). Couldn’t the small droplets at the bottom of the fog be removed progressively by occult deposition rather than evaporation?

Figure changes:

Figure 3 caption: (the same figure as on 5). > (the same data as figure 5).

Figure 4: Indicate the height of Max(LWC) for calculating γ fit and α fit

Figure 5: indicate measured visibility at 2 m when available and indicate phases of fog (development, mature, disappearing).

Figure 6 Caption: Explain orange diamonds. Indicate fog stage.

Figure 7 Indicate fog stage.

Figures 12, 13 and 14: For the axes, key labels and equations change: I > SW and IR > LW.

Figures 13 and 14: Indicate the times of fog.

Figure A1: Indicate the times of fog.

Figure A6: Use much larger fonts for the altitude axes and drop size axes and their values. To save space just show one key to the vDSD colour values.

Minor changes throughout manuscript:

1) For end-sentence citations put authors names inside brackets with year e.g.: L25 Capobianco and Lee (2001). > (Capobianco and Lee, 2001). e.g.: L39 Mason (1982); Price (2011). > (Mason, 1982; Price, 2011).

2) Add ‘the’ before ‘SOFIG3D experiment’, e.g.: L45 during SOFOG3D experiment. > during the SOFOG3D experiment.

3)    droop > drop, drooped > dropped, drooping > dropping

4)    grater > greater

Minor changes (L = manuscript line number):

L44 thick layer > thick fog

L46 if when > if or when

L54 so it more matches > so it more accurately matches

L55 met in SOFOG3D experiment closely in time within around 1 hour. > met in the SOFOG3D experiment within about 1 hour.

L66-67 (the more time before sunrise, the better) > (the earlier before sunrise the more likely)

L103 Second fog case > A second fog case

L103 consist of > consisted of

L104 LWC had > The LWC had

L107 however were > however they were

L113 allow for determining the > allow determination of the

L121 region of Stryzowskie Foothills > the region of the Stryzowskie Foothills

L122 numerical simulation were used > numerical simulations were used

L128 (collaborates with University of Warsaw) is located > (which collaborates with University of Warsaw) located [ADD ‘which’ and DELETE ‘is’]

L156 For measurements were used two meteorological balloons filled with helium. > The measurements used two meteorological balloons filled with helium.

L156 Balloon was > The balloons were 

L161 realtive > relative

L163 however here > here

L189 for few seconds > for a few seconds

L190 15 minutes brakes > 15 minute breaks

L194 monitor situation > monitor the situation

L195 basis of a considerable power laser with invisible to the human eye light, > basis of a laser with considerable power using a wavelength invisible to humans,

L196-197 used measure > used to measure

L205 was no more going upward > ceased going upward

L209 allows to calculate > allows calculation

L219 it is adiabatic condensation rate > it is the adiabatic condensation rate

L235 was taken as the same as in > was taken from

L236 To calculate what αeq is, just > To calculate αeq, just

L245 In later part > In the latter part

L254 Fu-Liou code > The Fu-Liou code

L256 Model covers > The model covers

L256 waves > wavelengths [two occurrences]

L257 Fu-Liou model provide > The Fu-Liou model provides Aerosol Angstrom exponents used in calculations. L276 ‘for 8-10 Sep, we assumed AE = 0.5 and for 11 Sep. AE = 1.0.’ Explain why these values of AE were assumed [Nb later lines 287 to 289 indicates these values were obtained from CIMEL at SolarAOTupper].

L303 5 UTC 5. > 5 UTC.

L343 35 m. Than decreasing > 35 m, then decreasing

L344 at CTH. > at the CTH.

L345 There were done 13 soundings > There were 13 soundings done

L346 as in previous stage > as in the previous stage

L346 The fog deepen, it top was at 102 m. > The fog deepened, its top was at 102 m.

L349 two areas. > two sections.

L374 There were done two profiles > There were two profiles

L391 where remaining > were remaining

L396 summaries > summarizes

L401 look different > looks different

L405 where > were

L405 diapered > disappeared

L419 deepen >deepened

L545 exanimate > examined

L548 nuclei than with > nuclei rather than with

L576 ‘begging stage’ – should this be ‘beginning stage’?