the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Jul 2025
Vertical Profiling of Canadian Wildfire Smoke in the Baltimore-Washington Corridor – Interactions with the Planetary Boundary Layer and Impact on Surface Air Quality
Abstract. The 2023 Canadian wildfires yielded record-breaking emissions that were transported long distances over large sections of the mid-Atlantic region, significantly impacting regional surface air quality. In this study, we analyzed the effect of long-distance transported wildfire smoke on the Baltimore-Washington Corridor (BWC), a highly populated and industrialized metropolitan region prone to air quality exceedances. Central to the analysis is the Vaisala CL61 ceilometer in Beltsville (suburban BWC), whose linear depolarization ratio (LDR) profiles provide a continuous, altitude-resolved fingerprint for distinguishing wildfire smoke from locally generated urban aerosols. By combining the LDR-derived with satellite imagery, surface air quality observations, and NOAA HYSPLIT trajectory analysis, we analyzed four discrete smoke events to characterize smoke's vertical distribution and interaction with the planetary boundary layer (PBL). One of the cases showed that the timing of smoke plume descent in relation to synoptic frontal passage was decisive in determining its impact on air quality. In contrast, those events with well-mixed smoke in the PBL during the advection-driven conditions exhibited a clear deterioration in air quality near the surface, with particulate levels exceeding the regulation threshold. The results underscore the importance of accurately representing vertical mixing in smoke forecasts and illustrate the added value of routine ceilometer LDR measurements for real-time identification of lofted smoke plumes – information not attainable from column-integrated satellite products or surface monitors alone.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(10300 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(10300 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
CC1: 'Comment on egusphere-2025-2991', Michael Fromm, 31 Jul 2025
-
AC2: 'Reply on CC1', Nakul Karle, 17 Sep 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2991/egusphere-2025-2991-AC2-supplement.pdf
-
AC2: 'Reply on CC1', Nakul Karle, 17 Sep 2025
-
CC2: 'Comment on egusphere-2025-2991', johan villanueva, 01 Aug 2025
Firstly, I would like to congratulate you on the excellent work you have conducted on studying the dynamics between the planetary boundary layer (PBL) and the smoke generated by wildfires. It is extremely important to understand how the PBL interacts with various gases and atmospheric pollutants, especially how wildfire events directly impact the air quality we breathe, as these episodes are closely linked to numerous respiratory issues and public health concerns.
A notable aspect of this study is the effective integration of multiple atmospheric instruments, which allows rigorous validation of the obtained results and provides a more comprehensive perspective of PBL dynamics. This includes not only interactions with smoke and pollutants but also with diverse meteorological events, such as wet or dry fronts, which vary according to the seasons.
Furthermore, I am highly interested in knowing, considering the clear limitations of current atmospheric models, how feasible and beneficial would it be to integrate advanced artificial intelligence techniques, such as machine learning, to improve the prediction of vertical interactions between wildfire smoke and the PBL, and consequently optimize preventive management of public health risks?
Citation: https://doi.org/10.5194/egusphere-2025-2991-CC2 -
AC1: 'Reply on CC2', Nakul Karle, 15 Aug 2025
Thank you very much, Johan, for the thoughtful read, kind words, and thought-provoking question. We share your motivations: the practical problem is not just “is there smoke in the atmosphere?” but “when and how will a lofted smoke layer couple with the PBL?” We believe this is where ML can help when we incorporate it with the physics and the measurements we already trust. Even though AI/ML integration is beyond the scope of the present manuscript, we view it as a clear next step and intend to pursue it. Highlighted a few recent advances that we came across that suggest a likely way forward.
The integration of advanced AI and ML is both possible and beneficial to improve the prediction of vertical interactions between wildfire smoke and the PBL. On feasibility, there is good evidence that ML can already improve key ingredients of the problem. For example, recent work shows that fusing multiple remote-sensing estimates with ML numerically improved PBL height retrievals and their diurnal evolution (Zhang et al., 2025). Xu et al. in their work underscored a two-step smoke injection scheme that explicitly represented Black Carbon (BC) absorption and convection, reducing the underestimation of elevated plume tops (Xu et al., 2025). At the same time, Siyuan Wang presented an ML emulator trained on Large Eddy Simulation (LES), which provided fast, physically consistent plume-rise estimates suitable for forecast cycling (Wang, 2024). These studies improve our understanding of the initial vertical distribution of smoke and its interaction with the convective mixed layer.
Downstream of the source of wildfire, assimilating what we measure about the vertical column makes a clear difference. In a recent study by Gao et al., the authors assimilated ceilometer aerosol-extinction profiles and demonstrated persistent improvements in 3-D aerosol structure and PM2.5 forecasts versus surface-only DA. This study demonstrated that dense ceilometer networks can nudge models toward more realistic aerosol profiling (Gao et al., 2025). In parallel, satellite-lidar evaluations show that many reanalysis/forecasts struggle with smoke vertical distribution, precisely the deficiency targeted by profile DA and ML (Shang et al., 2024).
Advances in AI/ML combined with physics-based understanding and high-resolution vertical observations can improve the smoke-PBL interaction predictions. More accurate timing and distinction between regular aerosol and smoke particles can lead to faster and efficient inputs for respective agencies, increasing decision-making and preventive management of public-health concerns during extreme smoke events.
References:
Shang, Xiaoxia, Antti Lipponen, Maria Filioglou, Anu-Maija Sundström, Mark Parrington, Virginie Buchard, Anton S. Darmenov et al. "Monitoring biomass burning aerosol transport using CALIOP observations and reanalysis models: a Canadian wildfire event in 2019." Atmospheric Chemistry and Physics 24, no. 2 (2024): 1329-1344.
Gao, Lina, Lipeng Jiang, Wei Sun, Peng Yan, Bing Qi, Chengli Ji, and Fa Tao. "Data assimilation of ceilometer aerosol extinction coefficient profile contributes to predictions of the three‐dimensional structures of aerosols in East China." Journal of Geophysical Research: Atmospheres 130, no. 9 (2025): e2024JD042408.
Wang, Siyuan. "Emulating Wildfire Plume Injection Using Machine Learning Trained by Large Eddy Simulation (LES)." Environmental Science & Technology 58, no. 50 (2024): 22204-22212.
Xu, Rui, Yan Yu, Xianglei Meng, Huiwen Xue, Chuanfeng Zhao, and Jintai Lin. "Atmospheric convection and aerosol absorption boost wildfire smoke injection." Geophysical Research Letters 52, no. 14 (2025): e2025GL115989.
Zhang, Damao, Jennifer Comstock, Chitra Sivaraman, Kefei Mo, Raghavendra Krishnamurthy, Jingjing Tian, Tianning Su, Zhanqing Li, and Natalia Roldán-Henao. "Best estimate of the planetary boundary layer height from multiple remote sensing measurements." Atmospheric Measurement Techniques 18, no. 14 (2025): 3453-3475.
Citation: https://doi.org/10.5194/egusphere-2025-2991-AC1
-
AC1: 'Reply on CC2', Nakul Karle, 15 Aug 2025
-
RC1: 'Comment on egusphere-2025-2991', Anonymous Referee #1, 17 Aug 2025
This manuscript describes analysis of wildfire smoke events over the Baltimore-Washington region using multiple observations and trajectory model simulations. It brings together these multiple lines of evidence for four different events, and shows how there can be different transport regimes even for similar satellite signals. The paper is generally well-written and presents interesting results; I think it will be acceptable for publication with relatively minor revisions.
Major Comment: The manuscript focuses on 4 episodes, but contains no discussion of how these were chosen or how these 4 episodes fit into the variations for 2023 (or any other year). Line 91 states “analysing four representative case studies for different plume-PBL interaction scenarios”, but I see no discussion of why representative. The Results section starts straight away with the first case study. There needs to be a discussion of how the 4 events were chosen. There also needs to be a discussion of how these episodes compare with the rest of the summer. E.g., Are these the only 4 smoke events (using some criteria)? How does the PM2.5 compare with the rest of the summer? The think plots show a few quantities for each day of the summer would help, both introducing the events and also showing if other events.
Minor Comments.
1. I agree with comments in CC1, especially the length of trajectories shown and the PBL height.
2. Line 196. I am not sure “End of May 2023, the …” is correct grammar.Citation: https://doi.org/10.5194/egusphere-2025-2991-RC1 -
AC3: 'Reply on RC1', Nakul Karle, 17 Sep 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2991/egusphere-2025-2991-AC3-supplement.pdf
-
AC3: 'Reply on RC1', Nakul Karle, 17 Sep 2025
-
RC2: 'Comment on egusphere-2025-2991', Anonymous Referee #2, 22 Aug 2025
The paper deals with the analyses of different smoke events transported to the Washington – Baltimore area. Intensive fires happened in Canada in late spring 2023 and therefore they are the main study cases, where discussion and analyses try to show the impact of these fires in air-quality for the measurement site in the Howard University Beltsville Campus. Most of the analyses are based on measurements by a Vaisala CL61 Ceilometer capable of providing lidar depolarization and attenuated backscattered at one wavelength. Additional measurements of nitrogen oxides (NOₓ), carbon monoxide (CO), volatile organic compounds (VOCs), and particulate matter (PM2.5) are presented to support the impact of transported smoke in air-quality. Main analyses are focused on identifying different transport patterns associated with different meteorological conditions and how they ultimately impact particle types in planetary boundary layer (PBL). This objective is of great importance and has potential for its publication in Atmospheric Measurement Techniques. But although the paper is generally well-written, I have some major concerns:
My first major concern is that authors do not identify appropriately the novelty of this study compared to previous studies in the same region. The Washington – Baltimore area is home of AERONET/Pandora networks, being the Howard University Beltsville Campus less than 15 km from NASA Goddard Space Flight Center where reference instruments of these networks are deployed with long datasets. There are many reference studies using these data (for example, look at Eck et al., 2013). Moreover, the station has been referenced for many multiwavelength lidar Raman developments and there are many studies. For example, Veselovskii et al., (2013) demonstrated how multiwavelength Raman lidar can retrieve aerosol optical and microphysical properties vertically resolved. Fire transport is not new – Veselovskii et al., (2015) study fire transport from California and how it impacts with the PBL. The DISCOVER-AQ field campaign served to study air-quality and how meteorological conditions and hygroscopic growth ultimately affect aerosol optical and microphysical properties (see Perez-Ramirez et al., 2021). I am discouraged because authors ignore all these studies, and I believe that it must be addressed.
My second major concern is related to the comments from the public discussions, which I support. I did not see clearly how the synoptic conditions changed during the entire study period. Backward-trajectories can give an overview of where air-masses come from, but they alone can not serve to interpret meteorological conditions. Authors need to give extensive explanations of the meteorological conditions supported by meteorological maps/charts. I encourage homogenizing the explanations as in the current manuscript there are maps for the first event (although only for 500 hPa with is not enough) while there is nothing for the other events.
My other major concern is about the use of lidar depolarization as the key parameter to study the transport of smoke particles. Apart from that authors do not specify if it is aerosol depolarization or total depolarization, this parameter only gives information of particle non-sphericity. Lidar community has largely used this parameter only for identifying aerosol non-sphericity (see articles from High Spectral Resolution Lidar developed in NASA Langley). I cannot accept references to Vaisala CL61 White paper (page 15, lines 336 and 339). Authors must acknowledge this limitation and base their study on aerosol attenuated backscatter that is related to aerosol load.
My last major concern is that discussion of smoke events is not well-addressed, taking into account the large databases publicly available. Indeed, it is based just in a satellite image. I strongly support the use of satellite aerosol retrievals that can give a quick overview of fires intensity. Also, if authors claim limitations in models they should compare the events with models such as MERRA-2 or CAMS. This has been done in previous studies for the study region (see Veselovskii et al., 2019). The large number of AERONET stations in North America can also serve to have a better picture of these extreme smoke events.
All these major concerns should encourage authors to review their conclusions. Particularly, statements such as ‘These findings advocate for the wider inclusion of
depolarization lidar in regional air‑quality networks and the assimilation of near‑real‑time LDR fields into smoke‑forecast systems (Lines 439-440)’ or ‘Depolarisation lidar proved indispensable for diagnosing smoke layers ' vertical and microphysical character (Lines 448-449).’ must be re-thought because LDR only serve for aerosol typing. I insist that for appropriate vertical aerosol optical and microphysical properties characterization multiwavelength lidar Raman are needed, and even these techniques require of case-dependent optimized-constraints (see Perez-Ramirez et al., 2019).
Minor Concerns
- Lines 44 – 45: Reference needed
- Line 48: Reference needed
- Line 65: Reference needed for the statement of uncertainties of smoke impact on PBL.
- Line 67: Please, specify which past studies.
- 1.1 – Study region: Give some coordinates for the region
- Line 110: Give reference.
- Line 131: LDR can not say anything about aerosol phase state. Please correct.
- Line 140: How temperature and relative humidity profiles were obtained ?
- Figure 1: Please, correct axis. It is difficult to read. Same happens in Figures 4, 6 and 8.
- Figure 1: I observe sub-figures with wind vectors. How were they obtained? Do you have wind profiles? Same happens in Figures 4, 6 and 8.
- Line 168: Referenced HYSPLIT appropriately.
- 2.5: Description of AERONET is vague. Which version of data are you using? What is data quality level used? Also, reference AERONET appropriately.
- Figure 3: Why did you use trajectory ensemble? If it provides clear added value. Why did you not use trajectory ensemble for the rest of cases?
- Line 238: Figure #? It looks like a error. Same for line 248
- Line 240: I insist that LDR alone only serves aerosol typing. Be careful.
- Line 245: Why is stabilization of the PBL typical in the evening?
- Lines 257 – 258: Angström exponent can only give information about possible predominance of fine or coarse mode. See studies from AERONET.
- Line 264: Give reference for NAQQS limits.
- Line 281: I do not understand the mention to GOES. Is it a typo?
- Line 301: With your lidar data alone you can not identify particle as fine mode predominance. See my major comments.
- Lines 309 – 310: The statement of the impact of clouds in PBL is vague. From your data, is there differences in PBL when these cirrus clouds are present?
- Case study 3: I can not understand the increase of air-quality related parameters If smoke is transported at altitudes above PBL and is decoupled.
- Lines 354 – 355: I can not understand your statement of pre-frontal winds role in facilitating vertical mixture.
- Lines 358 – 359: AOD is related to aerosol load. For better characterization of smoke properties you need to use AERONET inversion data.
- Lines 360 – 362: I can not see the impact of smoke on ozone suppression
- Lines 391: Please, provide reference.
- Line 401 – 402: You have not showed attenuated backscatter that is related to total aerosol load. LDR serves only for aerosol typing. This is related to my previous major comment
- Discussions and Conclusions: I recommend using a table that summarizes the main findings – i.e. statistics of different air-quality related parameters
References
- Eck et al., 2014: Observations of rapid aerosol optical depth enhancements in the vicinity of polluted cumulus clouds, Chem. Phys., 14, 11633–11656
- Perez-Ramirez et al., 2019: Retrievals of aerosol single scattering albedo by multiwavelength lidar measurements: Evaluations with NASA Langley HSRL-2 during discover-AQ field campaigns, Remote Sensing of Environment, 222, 144-164.
- Perez-Ramirez et al. 2021: Spatiotemporal changes in aerosol properties by hygroscopic growth and impacts on radiative forcing and heating rates during DISCOVER-AQ 2011, Chem. Phys., 21, 12021–12048, 2021
- Veselovskii et al., 2013: Retrieval of spatio-temporal distributions of particle parameters from multiwavelength lidar measurements using the linear estimation technique and comparison with AERONET, Meas. Tech., 6, 2671–2682,
- Veslovskii et al., 2015: Characterization of forest fire smoke event nearWashington, DC in summer 2013 with multi-wavelength lidar, Chem. Phys., 15, 1647–1660.
Citation: https://doi.org/10.5194/egusphere-2025-2991-RC2 -
AC4: 'Reply on RC2', Nakul Karle, 17 Sep 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2991/egusphere-2025-2991-AC4-supplement.pdf
Interactive discussion
Status: closed
-
CC1: 'Comment on egusphere-2025-2991', Michael Fromm, 31 Jul 2025
This is a very interesting data set in Beltsville, MD, USA. It appears to be a valuable, specialized resource for studies like this. It will be great to learn more about it and for the science community to have eventual access to these data.
Much of the paper deals with “descent” of smoke layers as manifested in lidar time-series data. I would caution that sloping aerosol and cloud features in such data representations may not be attributable to meteorological forces or sedimentation. A lidar time series simply shows what is blowing overhead at different times. There’s no way to know a particle’s vertical history or future from such a rolling snapshot of a particulate layer. Please see a comment posted to an ACP paper published back in 2010: https://acp.copernicus.org/articles/10/11921/2010/acp-10-11921-2010-discussion.html
For some reason, all the HYSPLIT trajectories are shorter than the specified 72 hours. Note the time series below each one. I suspect that this is an artifact of the HRRR data choice. The HRRR data are not global. It can be seen that some of the trajectories end up at about the same latitude in Canada. Maybe that’s the edge of the HRRR data grid? I tested one of the scenarios, using GFS global data, and the results came out as 72 hours long. Regardless of the reason, the 72-hr premise is not borne out in the data.
Speaking of trajectories, it might not be the case that they show an “origin” any more than a possible path through smoke. There is no information inherent in the trajectories indicating a polluting origin point. The fact that the trajectories illustrated in the paper are all shorter than stated adds to the uncertainty of their interpretation. But even when that is corrected, the trajectory paths and endpoints by themselves do not identify a smoke-initiation point.
The PBL is central to this manuscript. I could not tell from the illustrations where the variable PBL was. Plotting the PBL throughout the lidar time series curtains would be a wonderful addition.
Citation: https://doi.org/10.5194/egusphere-2025-2991-CC1 -
AC2: 'Reply on CC1', Nakul Karle, 17 Sep 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2991/egusphere-2025-2991-AC2-supplement.pdf
-
AC2: 'Reply on CC1', Nakul Karle, 17 Sep 2025
-
CC2: 'Comment on egusphere-2025-2991', johan villanueva, 01 Aug 2025
Firstly, I would like to congratulate you on the excellent work you have conducted on studying the dynamics between the planetary boundary layer (PBL) and the smoke generated by wildfires. It is extremely important to understand how the PBL interacts with various gases and atmospheric pollutants, especially how wildfire events directly impact the air quality we breathe, as these episodes are closely linked to numerous respiratory issues and public health concerns.
A notable aspect of this study is the effective integration of multiple atmospheric instruments, which allows rigorous validation of the obtained results and provides a more comprehensive perspective of PBL dynamics. This includes not only interactions with smoke and pollutants but also with diverse meteorological events, such as wet or dry fronts, which vary according to the seasons.
Furthermore, I am highly interested in knowing, considering the clear limitations of current atmospheric models, how feasible and beneficial would it be to integrate advanced artificial intelligence techniques, such as machine learning, to improve the prediction of vertical interactions between wildfire smoke and the PBL, and consequently optimize preventive management of public health risks?
Citation: https://doi.org/10.5194/egusphere-2025-2991-CC2 -
AC1: 'Reply on CC2', Nakul Karle, 15 Aug 2025
Thank you very much, Johan, for the thoughtful read, kind words, and thought-provoking question. We share your motivations: the practical problem is not just “is there smoke in the atmosphere?” but “when and how will a lofted smoke layer couple with the PBL?” We believe this is where ML can help when we incorporate it with the physics and the measurements we already trust. Even though AI/ML integration is beyond the scope of the present manuscript, we view it as a clear next step and intend to pursue it. Highlighted a few recent advances that we came across that suggest a likely way forward.
The integration of advanced AI and ML is both possible and beneficial to improve the prediction of vertical interactions between wildfire smoke and the PBL. On feasibility, there is good evidence that ML can already improve key ingredients of the problem. For example, recent work shows that fusing multiple remote-sensing estimates with ML numerically improved PBL height retrievals and their diurnal evolution (Zhang et al., 2025). Xu et al. in their work underscored a two-step smoke injection scheme that explicitly represented Black Carbon (BC) absorption and convection, reducing the underestimation of elevated plume tops (Xu et al., 2025). At the same time, Siyuan Wang presented an ML emulator trained on Large Eddy Simulation (LES), which provided fast, physically consistent plume-rise estimates suitable for forecast cycling (Wang, 2024). These studies improve our understanding of the initial vertical distribution of smoke and its interaction with the convective mixed layer.
Downstream of the source of wildfire, assimilating what we measure about the vertical column makes a clear difference. In a recent study by Gao et al., the authors assimilated ceilometer aerosol-extinction profiles and demonstrated persistent improvements in 3-D aerosol structure and PM2.5 forecasts versus surface-only DA. This study demonstrated that dense ceilometer networks can nudge models toward more realistic aerosol profiling (Gao et al., 2025). In parallel, satellite-lidar evaluations show that many reanalysis/forecasts struggle with smoke vertical distribution, precisely the deficiency targeted by profile DA and ML (Shang et al., 2024).
Advances in AI/ML combined with physics-based understanding and high-resolution vertical observations can improve the smoke-PBL interaction predictions. More accurate timing and distinction between regular aerosol and smoke particles can lead to faster and efficient inputs for respective agencies, increasing decision-making and preventive management of public-health concerns during extreme smoke events.
References:
Shang, Xiaoxia, Antti Lipponen, Maria Filioglou, Anu-Maija Sundström, Mark Parrington, Virginie Buchard, Anton S. Darmenov et al. "Monitoring biomass burning aerosol transport using CALIOP observations and reanalysis models: a Canadian wildfire event in 2019." Atmospheric Chemistry and Physics 24, no. 2 (2024): 1329-1344.
Gao, Lina, Lipeng Jiang, Wei Sun, Peng Yan, Bing Qi, Chengli Ji, and Fa Tao. "Data assimilation of ceilometer aerosol extinction coefficient profile contributes to predictions of the three‐dimensional structures of aerosols in East China." Journal of Geophysical Research: Atmospheres 130, no. 9 (2025): e2024JD042408.
Wang, Siyuan. "Emulating Wildfire Plume Injection Using Machine Learning Trained by Large Eddy Simulation (LES)." Environmental Science & Technology 58, no. 50 (2024): 22204-22212.
Xu, Rui, Yan Yu, Xianglei Meng, Huiwen Xue, Chuanfeng Zhao, and Jintai Lin. "Atmospheric convection and aerosol absorption boost wildfire smoke injection." Geophysical Research Letters 52, no. 14 (2025): e2025GL115989.
Zhang, Damao, Jennifer Comstock, Chitra Sivaraman, Kefei Mo, Raghavendra Krishnamurthy, Jingjing Tian, Tianning Su, Zhanqing Li, and Natalia Roldán-Henao. "Best estimate of the planetary boundary layer height from multiple remote sensing measurements." Atmospheric Measurement Techniques 18, no. 14 (2025): 3453-3475.
Citation: https://doi.org/10.5194/egusphere-2025-2991-AC1
-
AC1: 'Reply on CC2', Nakul Karle, 15 Aug 2025
-
RC1: 'Comment on egusphere-2025-2991', Anonymous Referee #1, 17 Aug 2025
This manuscript describes analysis of wildfire smoke events over the Baltimore-Washington region using multiple observations and trajectory model simulations. It brings together these multiple lines of evidence for four different events, and shows how there can be different transport regimes even for similar satellite signals. The paper is generally well-written and presents interesting results; I think it will be acceptable for publication with relatively minor revisions.
Major Comment: The manuscript focuses on 4 episodes, but contains no discussion of how these were chosen or how these 4 episodes fit into the variations for 2023 (or any other year). Line 91 states “analysing four representative case studies for different plume-PBL interaction scenarios”, but I see no discussion of why representative. The Results section starts straight away with the first case study. There needs to be a discussion of how the 4 events were chosen. There also needs to be a discussion of how these episodes compare with the rest of the summer. E.g., Are these the only 4 smoke events (using some criteria)? How does the PM2.5 compare with the rest of the summer? The think plots show a few quantities for each day of the summer would help, both introducing the events and also showing if other events.
Minor Comments.
1. I agree with comments in CC1, especially the length of trajectories shown and the PBL height.
2. Line 196. I am not sure “End of May 2023, the …” is correct grammar.Citation: https://doi.org/10.5194/egusphere-2025-2991-RC1 -
AC3: 'Reply on RC1', Nakul Karle, 17 Sep 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2991/egusphere-2025-2991-AC3-supplement.pdf
-
AC3: 'Reply on RC1', Nakul Karle, 17 Sep 2025
-
RC2: 'Comment on egusphere-2025-2991', Anonymous Referee #2, 22 Aug 2025
The paper deals with the analyses of different smoke events transported to the Washington – Baltimore area. Intensive fires happened in Canada in late spring 2023 and therefore they are the main study cases, where discussion and analyses try to show the impact of these fires in air-quality for the measurement site in the Howard University Beltsville Campus. Most of the analyses are based on measurements by a Vaisala CL61 Ceilometer capable of providing lidar depolarization and attenuated backscattered at one wavelength. Additional measurements of nitrogen oxides (NOₓ), carbon monoxide (CO), volatile organic compounds (VOCs), and particulate matter (PM2.5) are presented to support the impact of transported smoke in air-quality. Main analyses are focused on identifying different transport patterns associated with different meteorological conditions and how they ultimately impact particle types in planetary boundary layer (PBL). This objective is of great importance and has potential for its publication in Atmospheric Measurement Techniques. But although the paper is generally well-written, I have some major concerns:
My first major concern is that authors do not identify appropriately the novelty of this study compared to previous studies in the same region. The Washington – Baltimore area is home of AERONET/Pandora networks, being the Howard University Beltsville Campus less than 15 km from NASA Goddard Space Flight Center where reference instruments of these networks are deployed with long datasets. There are many reference studies using these data (for example, look at Eck et al., 2013). Moreover, the station has been referenced for many multiwavelength lidar Raman developments and there are many studies. For example, Veselovskii et al., (2013) demonstrated how multiwavelength Raman lidar can retrieve aerosol optical and microphysical properties vertically resolved. Fire transport is not new – Veselovskii et al., (2015) study fire transport from California and how it impacts with the PBL. The DISCOVER-AQ field campaign served to study air-quality and how meteorological conditions and hygroscopic growth ultimately affect aerosol optical and microphysical properties (see Perez-Ramirez et al., 2021). I am discouraged because authors ignore all these studies, and I believe that it must be addressed.
My second major concern is related to the comments from the public discussions, which I support. I did not see clearly how the synoptic conditions changed during the entire study period. Backward-trajectories can give an overview of where air-masses come from, but they alone can not serve to interpret meteorological conditions. Authors need to give extensive explanations of the meteorological conditions supported by meteorological maps/charts. I encourage homogenizing the explanations as in the current manuscript there are maps for the first event (although only for 500 hPa with is not enough) while there is nothing for the other events.
My other major concern is about the use of lidar depolarization as the key parameter to study the transport of smoke particles. Apart from that authors do not specify if it is aerosol depolarization or total depolarization, this parameter only gives information of particle non-sphericity. Lidar community has largely used this parameter only for identifying aerosol non-sphericity (see articles from High Spectral Resolution Lidar developed in NASA Langley). I cannot accept references to Vaisala CL61 White paper (page 15, lines 336 and 339). Authors must acknowledge this limitation and base their study on aerosol attenuated backscatter that is related to aerosol load.
My last major concern is that discussion of smoke events is not well-addressed, taking into account the large databases publicly available. Indeed, it is based just in a satellite image. I strongly support the use of satellite aerosol retrievals that can give a quick overview of fires intensity. Also, if authors claim limitations in models they should compare the events with models such as MERRA-2 or CAMS. This has been done in previous studies for the study region (see Veselovskii et al., 2019). The large number of AERONET stations in North America can also serve to have a better picture of these extreme smoke events.
All these major concerns should encourage authors to review their conclusions. Particularly, statements such as ‘These findings advocate for the wider inclusion of
depolarization lidar in regional air‑quality networks and the assimilation of near‑real‑time LDR fields into smoke‑forecast systems (Lines 439-440)’ or ‘Depolarisation lidar proved indispensable for diagnosing smoke layers ' vertical and microphysical character (Lines 448-449).’ must be re-thought because LDR only serve for aerosol typing. I insist that for appropriate vertical aerosol optical and microphysical properties characterization multiwavelength lidar Raman are needed, and even these techniques require of case-dependent optimized-constraints (see Perez-Ramirez et al., 2019).
Minor Concerns
- Lines 44 – 45: Reference needed
- Line 48: Reference needed
- Line 65: Reference needed for the statement of uncertainties of smoke impact on PBL.
- Line 67: Please, specify which past studies.
- 1.1 – Study region: Give some coordinates for the region
- Line 110: Give reference.
- Line 131: LDR can not say anything about aerosol phase state. Please correct.
- Line 140: How temperature and relative humidity profiles were obtained ?
- Figure 1: Please, correct axis. It is difficult to read. Same happens in Figures 4, 6 and 8.
- Figure 1: I observe sub-figures with wind vectors. How were they obtained? Do you have wind profiles? Same happens in Figures 4, 6 and 8.
- Line 168: Referenced HYSPLIT appropriately.
- 2.5: Description of AERONET is vague. Which version of data are you using? What is data quality level used? Also, reference AERONET appropriately.
- Figure 3: Why did you use trajectory ensemble? If it provides clear added value. Why did you not use trajectory ensemble for the rest of cases?
- Line 238: Figure #? It looks like a error. Same for line 248
- Line 240: I insist that LDR alone only serves aerosol typing. Be careful.
- Line 245: Why is stabilization of the PBL typical in the evening?
- Lines 257 – 258: Angström exponent can only give information about possible predominance of fine or coarse mode. See studies from AERONET.
- Line 264: Give reference for NAQQS limits.
- Line 281: I do not understand the mention to GOES. Is it a typo?
- Line 301: With your lidar data alone you can not identify particle as fine mode predominance. See my major comments.
- Lines 309 – 310: The statement of the impact of clouds in PBL is vague. From your data, is there differences in PBL when these cirrus clouds are present?
- Case study 3: I can not understand the increase of air-quality related parameters If smoke is transported at altitudes above PBL and is decoupled.
- Lines 354 – 355: I can not understand your statement of pre-frontal winds role in facilitating vertical mixture.
- Lines 358 – 359: AOD is related to aerosol load. For better characterization of smoke properties you need to use AERONET inversion data.
- Lines 360 – 362: I can not see the impact of smoke on ozone suppression
- Lines 391: Please, provide reference.
- Line 401 – 402: You have not showed attenuated backscatter that is related to total aerosol load. LDR serves only for aerosol typing. This is related to my previous major comment
- Discussions and Conclusions: I recommend using a table that summarizes the main findings – i.e. statistics of different air-quality related parameters
References
- Eck et al., 2014: Observations of rapid aerosol optical depth enhancements in the vicinity of polluted cumulus clouds, Chem. Phys., 14, 11633–11656
- Perez-Ramirez et al., 2019: Retrievals of aerosol single scattering albedo by multiwavelength lidar measurements: Evaluations with NASA Langley HSRL-2 during discover-AQ field campaigns, Remote Sensing of Environment, 222, 144-164.
- Perez-Ramirez et al. 2021: Spatiotemporal changes in aerosol properties by hygroscopic growth and impacts on radiative forcing and heating rates during DISCOVER-AQ 2011, Chem. Phys., 21, 12021–12048, 2021
- Veselovskii et al., 2013: Retrieval of spatio-temporal distributions of particle parameters from multiwavelength lidar measurements using the linear estimation technique and comparison with AERONET, Meas. Tech., 6, 2671–2682,
- Veslovskii et al., 2015: Characterization of forest fire smoke event nearWashington, DC in summer 2013 with multi-wavelength lidar, Chem. Phys., 15, 1647–1660.
Citation: https://doi.org/10.5194/egusphere-2025-2991-RC2 -
AC4: 'Reply on RC2', Nakul Karle, 17 Sep 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2991/egusphere-2025-2991-AC4-supplement.pdf
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 757 | 98 | 37 | 892 | 20 | 26 |
- HTML: 757
- PDF: 98
- XML: 37
- Total: 892
- BibTeX: 20
- EndNote: 26
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Ricardo K. Sakai
Rocio D. Rossi
Adrian Flores
Sen Chiao
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(10300 KB) - Metadata XML
This is a very interesting data set in Beltsville, MD, USA. It appears to be a valuable, specialized resource for studies like this. It will be great to learn more about it and for the science community to have eventual access to these data.
Much of the paper deals with “descent” of smoke layers as manifested in lidar time-series data. I would caution that sloping aerosol and cloud features in such data representations may not be attributable to meteorological forces or sedimentation. A lidar time series simply shows what is blowing overhead at different times. There’s no way to know a particle’s vertical history or future from such a rolling snapshot of a particulate layer. Please see a comment posted to an ACP paper published back in 2010: https://acp.copernicus.org/articles/10/11921/2010/acp-10-11921-2010-discussion.html
For some reason, all the HYSPLIT trajectories are shorter than the specified 72 hours. Note the time series below each one. I suspect that this is an artifact of the HRRR data choice. The HRRR data are not global. It can be seen that some of the trajectories end up at about the same latitude in Canada. Maybe that’s the edge of the HRRR data grid? I tested one of the scenarios, using GFS global data, and the results came out as 72 hours long. Regardless of the reason, the 72-hr premise is not borne out in the data.
Speaking of trajectories, it might not be the case that they show an “origin” any more than a possible path through smoke. There is no information inherent in the trajectories indicating a polluting origin point. The fact that the trajectories illustrated in the paper are all shorter than stated adds to the uncertainty of their interpretation. But even when that is corrected, the trajectory paths and endpoints by themselves do not identify a smoke-initiation point.
The PBL is central to this manuscript. I could not tell from the illustrations where the variable PBL was. Plotting the PBL throughout the lidar time series curtains would be a wonderful addition.