Remote sensing of exceptional winter aerosol pollution events and representativeness of the surface – column relationship over Paris metropolitan area

Abstract. In this study an optical parameter derived from lidar measurements is found to be relevant to monitor the evolution of near-surface particulate concentrations. This highlights the opportunities offered by future spaceborne lidar missions in air quality assessment on a global scale. This work is carried out following a dedicated field campaign in the Paris area (France) during winter 2016–2017, from 1st November to 31st January. Two of the most intense winter aerosol pollution events occurring over the last decade were sampled using a ground-based N2-Raman. The lidar operated continuously at the wavelength of 355 nm, favourable to the measurement of submicron aerosols mainly linked to traffic emissions. The data analysis uses the synergy between ground-based and spaceborne lidar observations, and data from the air quality monitoring network Airparif. The first severe aerosol pollution event occurred on 1st December 2016; it concerned a circular area of 250 km in diameter around Paris with maximum PM10 (PMx is the mass concentration of particles with an aerodynamic diameter smaller than x µm) values of 121 ± 63 µg m-3. The second event took place from 21st to 22nd January which covered all of Western Europe, with maxima of PM10 (156 ± 33 µg m-3) and aerosol extinction coefficient (AEC) between 0.6 and 1 km-1, within the winter atmospheric boundary layer. These two major aerosol pollution events share very low boundary layer height, down to 300 m above ground level. However, they did not take place in the same weather condition; moreover, they are associated with significantly different lidar ratios: 72 ± 15 sr and 56 ± 15 sr, respectively in December and January. Such results are consistent with available spaceborne lidar data (70 ± 25 sr) and values found in the literature. During these two events, the continuous temporal evolution of the aerosol extinction coefficient allows us to investigate the representativeness of optical parameters found in the planetary boundary layer to assess surface aerosol concentration. No one-to-one relationship between the aerosol optical thickness (AOT) and PM2.5 values stands out within our study. In contrast, the maximum lidar-derived aerosol extinction coefficient found within the planetary boundary layer is identified as a consistent variable to assess the evolution of ground aerosol concentration.


I suggest that prior to considering this paper for publication, therefore, the words quoted above should be removed from the title and the abstract, which would then start with "This work is carried out following a dedicated field campaign in the Paris area (France) during winter 2016-2017", clarifying from the onset what is the main content of the paper. The incidental study on the representativeness is mentioned towards the end of the abstract and this is ok ("During these two events [...] allows us to investigate the representativeness of optical parameters found in the planetary boundary layer to assess surface aerosol concentration."). What is in my opinion should be avoided is to start with a very broad promise, and then not be able to satisfy the reader's curiosity.
In summary, the dataset is too limited to be suitable for a general study on the abovementioned representativeness (few observations and specific meteorological conditions). This is was already addressed in the "quick review" report, but it seems that I have not persuaded the authors. I suggest that in the present form of the paper, the authors have not presented sufficient evidence to be able to state that the observations are generally representative of pollution events over Paris in the winter and therefore that the results can hint a relationship between surface and columnar properties, but cannot be considered to be general. I will also try and produce a detailed review, but please consider the above to be the major point to be addressed (in my opinion).

Best regards.
During this first stage of the review process, our manuscript has been quite widely revised following your comments, without losing our scientific objective, mainly dedicated to the interest of lidar measurements to assess the impact of winter particulate pollution.
Still, you consider that the relationship between surface and remote sensing observations is not sufficiently addressed to be the topic of our study. Your comments were overwhelmingly constructive and helpful, yet we partly disagree with you in this last comment. Regarding the finite nature of the dataset used in the paper, we show in Section 3 that even though the measurements are sampled in a given period of time the dataset are representative of pollution events occurring over the Paris area during winter.
It is difficult to do without the detailed description of the case studies used to establish the relationship between the surface and the PBL. It is indeed important to define them well in order to correctly set the boundary conditions of the study, and this last point answers your objection. Moreover, as we highlight, the two observed events of particulate pollution find equivalents in previous years, which shows that the statistical approach presented in our article is robust and may be generalized to other similar situations that are the majority in winter in the Paris region.
To follow your advice, be more concise and show that our study is mainly based on two major winter pollution events, we have revised the title and the abstract.

RC3 -Received and published: 25 September 2019
Please refer also to my previous comment on the fact that "surface -column relationship is touched in the paper but not the main topic". I believe that the data collected are very valuable, but my major objection is with the scope of the paper, as explained therein. The paper as it stands raises the reader's expectations a bit too much.

Detailed suggestions follow below:
 MAJOR COMMENTS: 1) I suggest to modify the title and abstract as explained in my previous referee comment (21 August). This is the major objection I have to the paper as it stands now.
The title of the paper has been shortened and modified to ensure that there is no ambiguity as to the scope of its results: "Remote sensing of exceptional winter aerosol pollution events and representativeness of groundbased measurements" The two first sentence of the abstract have also been removed.
2) Another statement that I think could be reviewed on the basis of the above is on page 3, lines 6-7.
Changes have been made in the text: "Hence, the main purpose of this paper is to describe the meteorological conditions that underlie the establishment of significant winter APEs, characterize observed APEs using in situ and remote sensing data and finally investigate the link between ground-based aerosol measurements and particles trapped within the winter PBL."

3) Angstrom exponent, page 4, lines 25-28. A few points should be clarified in my opinion: (a) when you say "constant" do you mean constant with height or with time? (b) is the Angstrom exponent an instantaneous value or a daily average? (c) the last sentence is unclear (what assumption and what has the horizontal advection to do with it?).
(a) Here, "constant" means constant with height, the sentence has been modified in this way: (b) The value of Ångström exponent taken is a mean value over the studied period, e.g. a mean value of Å from the 20 th to the 23 rd January. Indeed, during such event this value does not present significant A sentence has been added following the paragraph discussed to mention this caveat: "Even though the network of ground-based stations is designed to be the most representative of the regional air quality, the spatial resolution remains coarse and the average could not be representative of all areas of the Paris region."  MINOR COMMENTS: 5) Abstract, line 12: replace "continuously" with "during two 5-day periods" as the lidar was not operated continuosly from 1 November to 31 January (see text).
The sentence has been removed because it was not necessary in the abstract. 6) Abstract, line 13: delete "submicron" (at this wavelength the lidar is also sensitive to supermicron particles) and add "thought to be" before "mainly" (you have no direct measurement of aerosol type/origin).
The sentence has been removed because it was not necessary in the abstract. 7) Abstract, line 15: explain the method used to determine the circular area and measure its diameter.
We used maps of ensemble reanalysis of chemical transport model to assess the dispersion of the APE. In the first case it shows a concentrated zone around the Paris region and its surroundings. In the second event of January the APE is spread in England, Northern France, Benelux and Germany. Such maps are available at https://atmosphere.copernicus.eu/.
Changes have been made in the text: "it concerned a circular area of ~250 km in diameter around Paris as shown by ensemble reanalyses of chemical transport models"

8) Abstract, line 17: explain what other information you have to say that the event covered all of Western Europe.
Same response as for the point 7). The +/-sign is related to standard deviation derived from the spatial averaged.
Changes have been made in the text: "The maximum PM10 (PMx is the mass concentration of particles with an aerodynamic diameter smaller than x µm) was 121±63µg m -3 (spatial average ± standard deviation) and the aerosol extinction coefficient (AEC) ranged from 0.2 to 1 km -1 . The second event took place from 20 th to 23 rd January which covered all of North-western Europe, with maxima of PM10 around 156±33 µg m -3 and AEC between 0.6 and 1 km -1 , within the winter atmospheric boundary layer." 10) Abstract, line 20: the sentence about weather conditions is vague, I suggest to be more specific and describe which type of weather conditions you are referring to.
Changes have been made in the text: "However, they did not take place under identical anticyclonic weather conditions" 11) Page 3, lines 8-9: "the most severe winter APEs above the Paris area": specify over which period of time they are the most severe (e.g. "from year Y to nowadays").

Changes have been made in the text:
"This study is based on a specific field campaign performed during the most severe winter APEs that occurred in the Paris area since 2009."

12) Page 4, line 20: "downgraded" -> "integrated"
Agreed. Changes have been made in the text: "To obtain a sufficient signal to noise ratio (SNR, SNR > 10 (Royer et al., 2011a)) from the N2-Raman channel during daytime, the vertical resolution is set to 15 m" 13) Page 5, line 6: "sources of uncertainties" -> "uncertainties for our lidar system" (it helps to know that Royer et al is not a generic paper but one that details the uncertainties for this specific lidar). Agreed. The sentence has been removed.

17) Page 8, line 24: expand better on the link between pollution levels and anticyclonic conditions.
Three sentences have been added in the text: "Indeed, despite a general trend in emissions to decline in the Paris region, there are still noteworthy episodes of pollution. When a strong high pressure system sets in over a long period of time, it prevents air mass advection, blocking the weather situation. Thus, the pollution still emitted, even if it is less than in the past, remains blocked by the high pressure system and ends up exceeding the health thresholds." 18) Page 8, lines 29-31: make dates consistent with dates in the abstract, please.
In the abstract we were referring to the most polluted days of each event. We agree that was not clear, dates in the abstract have been modified.

19) Page 9, line 1: next to meteorological patterns add "see section 3.2"
Changes have been made in the text: "According to ERA5 reanalyses, the meteorological patterns (see section 3.2) are similar over the 8 days" The table header has been changed following your suggestion, and the caption too:

20)
"The 8 most severely polluted days of the past decade in winter. For each day we give both PM2.5 and PM10 measured at ground level (Airparif network) in the format Max/Mean/Min where: Max and Min are the hourly maximum and minimum value measured at a given background station during the day and Mean is the daily average over all background stations." 21) Figure 3, caption: wind velocity and direction at which altitude level? surface?
Changes have been made in the caption: "The geopotential altitude (white lines) and the wind direction and velocity (black arrow) are given at a 975-hPa level." 22) Page 12, line 7: add "single" before "grid point" and give lat/lon of the grid point centre.
Changes have been made in the text: "We consider the single grid point of 0.25° x 0.25° which includes central Paris (48.875°N, 2.375°E)." Fig. 4, x- Changes have been made in the text:

23)
"the sky is rather aerosol-free" 25) Page 13, line 19: state in the paper that you have chosen to keep the data associated with the middle and high altitude clouds and why. Even better, they could be displayed in a different colour for easy identification.
Changes have been made in the text: "These discrepancies are mainly due to the presence of middle and high-altitude clouds identified on lidar vertical profiles, which may bias the AERONET operational products (Chew et al., 2011). As far as lidar data are not disturbed by high clouds these profiles are kept in the figure."