the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 06 Jul 2026
Smoke transport and potential impacts from all observed pyroCb events in North America during 2013–2023
Abstract. Pyrocumulonimbus (pyroCb) events provide an efficient pathway for wildfire smoke and combustion products to enter the Upper Troposphere and Lower Stratosphere (UTLS), where they are transported far from their source regions and perturb atmospheric composition, radiation, and dynamics. Although there are individual case studies, their cumulative transport behaviour over decadal time scales remains insufficiently characterized. Here, we investigate the short-, mid-, and long-range transport pathways, seasonal and regional variability, vertical structure, and residence characteristics of smoke from all observed North American pyroCbs during 2013–2023 using the Langley Trajectory Model, evaluated against satellite observations. Results show that smoke transport is influenced by source region, season, injection layer, and the evolving synoptic-scale circulation. Summer accounts for the majority of events, while Canadian pyroCbs occur more frequently than pyroCbs in the United States and dominate much of the UTLS and long-range trans-Atlantic transport. Canadian smoke is often transported along midlatitude pathways toward Europe, whereas U.S. pyroCbs more strongly contribute to lower-latitude short- and mid-range transport and to tropical transport pathways. Short- and mid-range tropospheric transport reveals persistent free- and upper-tropospheric downstream convergence regions, indicating areas where smoke may repeatedly influence clouds, radiation, weather-relevant conditions. In contrast, UTLS and stratospheric transport pathways exhibit faster and longer-range transport, including recurrent midlatitude, tropical, cross-hemispheric pathways. This study provides a decadal transport framework for interpreting how smoke injected by North American pyroCbs is subsequently redistributed. The results help connect pyroCb smoke with observed smoke signatures and atmospheric impacts, while supporting future model evaluation and mission planning.
- Preprint
(2101 KB) - Metadata XML
-
Supplement
(3475 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-3536', Anonymous Referee #1, 29 Jul 2026
- AC1: 'Reply on RC1', Rubel Chandra Das, 14 Aug 2026
-
RC2: 'Comment on egusphere-2026-3536', Anonymous Referee #2, 19 Aug 2026
General impression:
The paper is well written and deals with an important topic: The long-range transport of North American wildfire smoke towards Europe and northern Asia. The manuscript contains new information about smoke transport in the northern hemisphere. It is a good and complementary contribution to all the numerous papers on smoke events since the Canadian pyroCb event in August 2017.
The literature review in section 1 needs to be improved and a better indication is required that self-lofting aspects are not considered .
Minor revisions are needed
Detailed comments:
The literature overview must be improved and should include more European papers.
Line 38: Ansmann et al. (ACP, 2018) showed that the stratospheric impact of the Canadian fire event 2017 was comparable with the Pinatubo impact over Europe in 1992.
Ansmann et al., 2018, https://doi.org/10.5194/acp-18-11831-2018
Line 42: Baars et al. (ACP, 2019) provided continental wide European lidar network observations of the Canadian wildfire smoke event in 2017 and showed that the stratospheric perturbation was visible in the lidar data over Europe even six to nine months after the fires. Ohneiser et al., ACP, (2022) reported long term observation of the Australian fires and showed that the stratospheric smoke remained visible in the lidar observations for more than 2 years.
Baars et al., 2019, https://doi.org/10.5194/acp-19-15183-2019
Ohneiser et al., 2022, https://doi.org/10.5194/acp-22-7417-2022
Line 43: Ansmann et al. (ACP, 2022), Ohneiser et al. (ACP, 2021), and Ohneiser et al. (ACP, 2022) focused on ozone depletion in polar regions. Ozone depletion at high latitudes are much stronger than the ones reported by Bernath et al. (2022) and Solomon et al. (2022) for the stratospheric ozone layer at midlatitudes. Ohneiser et al. (ACP, 2021) was the first paper in which a potential impact of wildfire smoke on ozone depletion was discussed.
Ansmann et al., 2022, https://doi.org/10.5194/acp-22-11701-2022
Ohneiser et al., 2021, https://doi.org/10.5194/acp-21-15783-2021
Ohneiser et al., 2022, https://doi.org/10.5194/acp-22-7417-2022
Line 50: I did not find Zhang et al. (2024) in the references. Why is there no alphabetic order of references?
Lines 51-60: You need to include the self-lofting aspect into the discussion (Ohneiser et al., ACP, 2023) right in the beginning of this paper!
Ohneiser et al., 2023, https://doi.org/10.5194/acp-23-2901-2023
Line 74: Ozone depletion was of the order of 5% to 10% at midlatitudes (Bernath et al, 2022, Solomon et al., 2022), caused by the Australian wildfire smoke, but 20%-25% were reported for Antarctic ozone depletion by Australian wildfire smoke in 2020 and 2021 (Ansmann et al., 2022). The Bernath et al. (2022) and the Solomon et al. (2022) papers were published more than one year after the paper of Ohneiser et al. (2021).
Line 77: Again, state clearly that you mean the self-lofting aspect here!
Section 3: Langley Trajectory Model (LaTM)
LaTM is not able to simulate self lofting. This is not a strong drawback, but must be mentioned. The accumulation of smoke around the tropopause, visible in ground-based and spaceborne lidar observations, is an indication that self lofting takes place, as described by Ohneiser et al. (2023). Even optically thin layers produce some warming. Self lofting is active even in cases of AODs below 0.5 or below 0.1 and works against sedimentation effects.
Lines 142-164: long description, long list of application.
Lines 161-165: Here the question comes up: How can you use satellite observations to support LaTM modeling with focus on North American wildfires when at the same time strong, year-by-year Siberian wildfires contribute to the overall northern hemispheric aerosol burden in the upper troposphere and lower stratosphere, and thus to AI in Figure 7a (OMPS observations)? This point should at least be discussed.
Line 293: A rigorous sensitivity analysis must include smoke scenarios with AOD >1.0 , 2.0, 3.0… Khaykin et al. (GRL, 2018) reported strong self-lofting effects after the Canadian PyroCb event in August 2017. During the first three days after the event, the smoke layers ascended by about 3 km per day in the stratosphere! To that time the smoke AOD was >2 at 500 nm wavelength.
Line 307-308: the same question as mentioned above: How can you compare observations with simulations without taking Siberian wildfire smoke contributions into account? Please comment on that.
Figure 2: panels need labels (a), (b), (c), ….., (s), (t)
Figure 2: I roughly estimated the smoke AOD in Figure 2 (CALIOP observations). AOD values are not given in the manuscript. For AODs of about a 0.01 to 0.2 at 532 nm in Fig. 2, the lofting effect is quite small so that all the comparisons with CALIOP will indicate: good consistency! However, the AOD can easily be >1.0 shortly after the injection event.
Section 5 is a strong and unique section!
Figure 3: panels need labels (a), (b), (c).
Figure 3: mean altitude? Please provide a more precise explanation. Do you mean the geometrical center height? Or simply the base height of the smoke layer?
Figure 4: panels need labels (a), (b), (c), (d), (e).
Lines 392-396: Now you introduce self-lofting. That is by far too late!
Figure 5: panels need labels (a), (b), (c).
Figure 6 shows model results. I am surprised that pyroCb activity can push the smoke up to 6 km above the tropopause. When CALIOP shows such features I always assume that this is combined effect of pyroCb convection and self lofting. But LaTM does not consider self-lofting! Any comment on that?Is that related to the fact that MERRA meteorological fields are indirectly influenced by self-lofting and air mass warming effects? Please comment on that!
Section 6: Multi-sensor corroboration of transport pathways
Again, the basic question I have: How can you compare observations (influenced by all fires in the northern hemisphere, i.e., including Siberian fire smoke) with transport modelling exclusively applied to simulate transport features of North American fire smoke?
Figure 8: panels need labels (a), (b), (c), (d).
Sections 8 and 9 (together) cover 6 pages. This is a very long discussion, much too long to my opinion. Facts and a brief outlook are ok. Readers are usually not very much interested in such a long discussion at the end of a long paper.
Citation: https://doi.org/10.5194/egusphere-2026-3536-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 227 | 64 | 16 | 307 | 32 | 13 | 14 |
- HTML: 227
- PDF: 64
- XML: 16
- Total: 307
- Supplement: 32
- BibTeX: 13
- EndNote: 14
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript presents a trajectory-based analysis of the transport of smoke from documented North American pyroCb events during 2013–2023. The compilation of a large number of events and the attempt to characterize their seasonal, regional and vertical transport pathways are valuable. The applications to satellite interpretation and campaign planning are also of interest.
However, several important methodological limitations are not adequately reflected in the interpretation of the results. I therefore recommend major revision. The manuscript could become a useful contribution, should the central claims be better aligned with what passive, unweighted, seven-day trajectories can actually demonstrate.
General remarks
1. Scope and claimed impacts
The title includes “potential impacts”, but the study does not directly investigate radiative, chemical, cloud or air-quality impacts. These are mostly hypothesized from the transport patterns or discussed based on previous studies. The title and conclusions should therefore focus more clearly on transport and possible implications.
2. Self-lofting and SWIRLs
The trajectory calculations do not account for radiative self-lofting, which can substantially modify both the altitude and horizontal pathway of intense smoke plumes as they ascend through vertical wind shear, whereas the climb rate of the intense plumes during the first days can be several hundred meters per day. Although self-lofting is acknowledged, its effects on the results shown in Figs. 3–6 are neither evaluated nor adequately discussed.
Similarly, the SWIRLs formed from intense stratospheric plumes are closely linked to self-lofting and can strongly affect residence time and meridional transport. The 2017 PNE, which produced several such vortices (Lesterlin et al., 2021), would provide a useful test case for comparing passive trajectories with observed plume evolution.
3. Meaning of the trajectory PDFs
The trajectory PDFs represent the frequency of idealized parcel occurrences, not aerosol concentration or smoke loading. Each event appears to contribute similarly, irrespective of its emitted mass or intensity, and all events are initialized using the same five-kilometre-deep vertical layer. Therefore, terms such as “concentration” and “smoke loading” should be used more cautiously.
3. Source distribution versus transport susceptibility
The apparent hotspots are determined first by where pyroCbs occurred during 2013–2023 and only subsequently by atmospheric transport. Therefore, the maps conflate source frequency with downstream transport.
The diagnostics cannot, by themselves, identify regions intrinsically “more susceptible” to smoke pollution or predict future smoke risk. Future source locations will depend on fire weather, fuel availability from previous burning, among other factors. For example, one would not necessarily expect the same frequency of severe pyroCb-generating fires in areas where the available fuel has recently burned. The authors should explicitly state that the results are conditional on the observed 2013–2023 source distribution. Claims about future susceptibility, aviation risk and broader wildfire pollution should be considerably qualified.
4. UTLS definition and interpretation of vertical transport
UTLS is not used consistently. In some passages it includes the upper troposphere and lower stratosphere, whereas elsewhere it is treated as a separate third layer. The tropopause and UTLS definitions should be clearly stated and applied consistently. The language should be revised consistently. Statements such as transport “from the troposphere into the UTLS” are awkward when the upper troposphere is itself part of the UTLS. Similarly, a convective overshoot normally refers to penetration above the tropopause rather than merely into the broadly defined UTLS.
5. Tropical transport and apparent ascent
The conclusion that smoke “ascends through the tropical UTLS” is not supported by the analysis. In the latitude–altitude composites, air transported equatorward along rising isentropic surfaces will appear at progressively higher absolute altitudes even without substantial cross-isentropic ascent.
Changes in parcel altitude, potential temperature and tropopause-relative altitude as a function of age would be more informative than the present aggregated cross-sections. The possible role of cloud processing and scavenging for smoke entering the tropical tropopause layer should also be discussed.
The role of the North American Monsoon Anticyclone is particularly underdeveloped, although it is likely important for smoke originating over the western US. NAMA should be considered before invoking interaction with the ASMA farther downstream.
6. Model validation
Section 4 is lengthy and repetitive, yet the validation remains qualitative and is restricted to the first seven days after one event. Visual similarity between the trajectory PDFs and CALIOP or OMPS features does not establish model accuracy in a quantitative sense.
The section should be shortened and possibly strengthened using objective diagnostics, for example the plume-altitude or spatial bias.
A second example addressing intercontinental transport is necessary because trans-Atlantic pathways are one of the main results. An event involving substantial self-lofting or vortex confinement would be particularly informative.
The statement that LaTM appears to perform better than NAME based on visual comparison should be removed unless a like-for-like comparison is performed.
7. Satellite and balloon comparisons
The climatological OMPS AI and MLS CO fields may contain unrelated signals from African and Asian dust, African and Amazonian fires, Asian pollution and ASMA circulation. Their broad spatial similarity with the trajectory PDFs should therefore be described as contextual corroboration rather than source-specific validation.
The balloon measurements are interesting but weakly integrated into the study. They are discussed at disproportionate length while being presented mainly in the supplement. If they provide direct validation, the source attribution and comparison should be shown more clearly in the main text. Otherwise, the discussion should be shortened and retained as an example of possible operational application. A separate study may be more suitable for the detailed balloon results.
Specific remarks
L. 90–91: Define OMPS-AI and MLS-CO at first use.
L. 128–131: The proposed interpretation is unclear. A lower tropopause does not necessarily constitute a barrier to the absolute vertical development of a pyroCb. Moreover, if it limits vertical development, why are the tropopause-relative injection heights apparently greater at higher latitudes? Please reconsider or better explain this interpretation.
L. 174: Please provide more details on the MLS CO product, particularly the pressure level, used.
L. 179–183: Why are trajectories initialized only at and below the reported maximum injection altitude? Given the uncertainty in the maximum height, a sensitivity test including levels above it would be useful. More importantly, the uniform five-kilometre layer cannot be assumed to represent “the entire mass” of each plume.
L. 193: Which tropopause definition from the repository is used? Please specify whether this is a lapse-rate, cold-point or dynamical tropopause.
Figure 2: The OMPS AI panels are difficult to read. Please adjust the colour scale and panel size.
L. 292–305: The sensitivity analysis is described as “rigorous”, but the supplement mainly presents qualitative comparisons. Please provide quantitative results or use more cautious terminology.
L. 339–341: The higher altitude towards the tropics may mainly reflect isentropic equatorward transport rather than diabatic ascent.
Figure 3: The comment concerning “higher altitude in the tropics” seems to refer to panel c, not panel b. The colour scale should be adjusted so that features other than the broad meridional altitude gradient can be distinguished.
L. 421: “Ascend through the tropical UTLS” is an overinterpretation of the aggregated latitude–altitude distribution. The equatorward isentropic transport should be distinguished from actual cross-isentropic ascent.
L. 426–427: Tropospheric smoke includes upper-tropospheric smoke, so transport “from the troposphere into the UTLS” is ambiguous. Convective overshoot normally refers to penetration above the tropopause.
L. 430: Please distinguish radiative self-lofting from dynamically driven ascent in warm conveyor belts. Evidence for radiative lofting from the free troposphere across the tropopause is not demonstrated here, nor am I aware of any solid observational evidence for such process.
L. 443–445: Please use “parcel age” rather than “residence time”.
L. 460–462: UV AI depends primarily on absolute aerosol-layer altitude and aerosol properties, not directly on tropopause-relative altitude.
L. 465: Replace “above 70° N” with “north of 70° N”.
L. 470–475: I do not see clear evidence of a coherent “tail” moving from Asia towards North America. The MLS field likely combines CO from several unrelated sources, including central African biomass burning, Asian pollution and monsoon transport, as well as Amazonian and North American fire emissions. The OMPS AI field can additionally contain Saharan dust. The role of NAMA is absent from this discussion, and ASMA is not defined at its first occurrence.
Figure 8: Add panel labels (a), (b), etc.
L. 557: Please clarify what is meant by “standard zonal flow”.
L. 567: Define NAMA at first use.
L. 567–575 and L. 594–598: The role of the North American Monsoon Anticyclone in transporting western North American smoke towards the subtropical Pacific and lower latitudes deserves a more complete discussion.
L. 633–637: The statement that the PDF is not mass-conserving is important and should appear earlier in the Methods.
L. 693: “Ascends through the tropical UTLS” remains too strong and should be reformulated.