Entwined long-range transports of biomass burning aerosols over the South-West Indian Ocean: a case study of aerosol river in September 2017

Gantois, Dominique; Baron, Alexandre; Sicard, Michaël; Duflot, Valentin; Groenen, Niels; Bègue, Nelson; Eckhardt, Sabine; Evangeliou, Nikolaos; De Mazière, Martine; Vigouroux, Corinne; Payen, Guillaume; Marquestaut, Nicolas; Metzger, Jean-Marc; Chazette, Patrick; Flamant, Cyrille

doi:10.5194/egusphere-2026-3070

Preprints

https://doi.org/10.5194/egusphere-2026-3070

Preprints

03 Jul 2026

| 03 Jul 2026

Status: this preprint is open for discussion and under review for Atmospheric Chemistry and Physics (ACP).

Entwined long-range transports of biomass burning aerosols over the South-West Indian Ocean: a case study of aerosol river in September 2017

Dominique Gantois, Alexandre Baron, Michaël Sicard, Valentin Duflot, Niels Groenen, Nelson Bègue, Sabine Eckhardt, Nikolaos Evangeliou, Martine De Mazière, Corinne Vigouroux, Guillaume Payen, Nicolas Marquestaut, Jean-Marc Metzger, Patrick Chazette, and Cyrille Flamant

Abstract. Biomass burning (BB) aerosols emitted over Southern Africa (SAF) and South America (SA) represent a major seasonal perturbation to the usually pristine atmosphere of the South-West Indian Ocean (SWIO) but remain understudied in this remote region. Following a multi-instrumental approach, we characterize BB plumes reaching Reunion Island (21° S, 55° E) during September 2017, combining ground-based measurements (sun-photometer, lidars, Fourier Transform Infrared spectrometer), spaceborne observations, CAMS EAC4 reanalysis, and the Lagrangian transport model FLEXPART. Aerosol optical depth at 550 nm over Reunion reached unusually high values (0.16–0.42) during the second half of September, with organic matter contributing up to 60 %. MODIS imagery revealed two large-scale smoke plumes originating from SAF and SA transported toward the SWIO, and CALIOP profiles showed smoke layers extending from 4 to 9 km of altitude above Madagascar. On September 19^th, a single layer was identified over Reunion between 2.8 and 4.7 km with an Ångström exponent (Å₃₅₅/₅₃₂) of 1.32±0.23, consistent with moderately aged BB particles essentially originating from SAF. On September 25^th, two vertically decoupled layers were identified: a lower layer (3.3–5 km, Å=1.45±0.12) associated with mixed aged aerosols of SAF and SA origin, and an upper, drier layer (5–9 km, Å=1.60±0.06) of potentially fresher SAF smoke, consistent with rapid convective uplift into the mid-troposphere. This study offers new insights concerning the dynamical processes that govern aerosol variability over Reunion Island and highlights its value as a strategic long-term observational site in the SWIO.

Received: 05 Jun 2026 – Discussion started: 03 Jul 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Status: open (until 16 Aug 2026)

Post a comment Subscribe to comment alert

RC1:
'Comment on egusphere-2026-3070', Anonymous Referee #1, 09 Jul 2026 reply
The manuscript presents a detailed case study of long-range transported biomass-burning aerosols over the South-West Indian Ocean during September 2017. By combining AERONET sun-photometer observations, ground-based lidar measurements, FTIR CO retrievals, MODIS and CALIOP observations, CAMS EAC4 reanalysis, and FLEXPART simulations, the authors characterize the horizontal transport, vertical layering, optical properties, water vapor environment, and possible source regions of smoke plumes reaching Reunion Island. The topic is important, especially because the SWIO remains relatively under-observed despite being strongly influenced by seasonal biomass-burning outflow from Southern Africa and, occasionally, South America. Overall, the paper is interesting and potentially valuable, but I think several points should be clarified or strengthened before publication.
The manuscript is framed around the concept of an “aerosol river” or “river of smoke”, but the definition of this term remains somewhat qualitative. The authors identify large-scale smoke transport features from MODIS AOD and wind fields, but it is not entirely clear what objective criteria are used to distinguish an “aerosol river” from a more general long-range transported smoke plume. I suggest that the authors define this concept more explicitly, ideally with reference to spatial continuity, transport pathway, aerosol loading, and/or lifetime. This would make the central framing of the manuscript clearer.

The multi-instrument approach is a major strength of the study, but the logical connection among the different datasets could be made clearer. At present, the paper sometimes reads as a sequence of independent observations: AERONET/CAMS for column AOD, MODIS/CALIOP for regional transport, lidar for vertical structure, FTIR/FLEXPART for CO transport, and WVMR for moisture conditions. I think the manuscript would be stronger if the authors more explicitly explained how each dataset constrains a specific part of the interpretation, and how the combined evidence supports the final attribution of source region, aging state, and vertical decoupling.

The attribution of aerosol origin, especially for the September 25 event, should be presented with more caution. The manuscript suggests that the lower layer is associated with mixed aged aerosols from Southern Africa and South America, while the upper layer is dominated by fresher Southern African smoke. This interpretation is plausible, but it relies heavily on FLEXPART simulations and indirect optical indicators. It would be useful to discuss the uncertainties in source attribution, including emission inventory uncertainties, injection height assumptions, transport pathway uncertainty, and the limited temporal sampling of lidar observations.

The interpretation of aerosol aging based on the Ångström exponent should be strengthened. The authors use Å355/532 values to infer moderately aged, aged, or potentially fresher biomass-burning particles. While this is reasonable, Ångström exponent alone cannot uniquely determine aerosol age because it can also be affected by size distribution, humidification, mixing state, and retrieval assumptions. I suggest that the authors revise the language to avoid over-interpreting Å as a direct aging indicator, or provide additional support from lidar ratio, depolarization, water vapor, transport time, or previous literature.

The treatment of lidar retrieval uncertainties deserves more attention. The authors use a lidar ratio of 75 sr for both days and both wavelengths after deriving a Raman-based value for the September 25 upper plume. This choice is understandable, but it may influence the retrieved extinction profiles and layer AODs. I suggest that the authors briefly assess or discuss the sensitivity of the extinction/AOD results to the assumed lidar ratio, especially for the September 19 case where Raman constraints were limited.

The water vapor analysis is interesting and potentially important for distinguishing vertically decoupled layers. However, the WVMR uncertainties reported for some layers are large, reaching approximately 74% of the mean value. I suggest that the authors be more cautious in interpreting the moisture differences, particularly for the lower layers, and clarify which conclusions are robust despite the large uncertainty. The upper dry layer on September 25 appears convincing, but the quantitative interpretation of moisture-aerosol coupling should be better qualified.

The comparison between CAMS AOD and AERONET is useful, and the reported correlation is high. However, the use of CAMS aerosol speciation to infer that OM dominates the AOD should be discussed more carefully. CAMS aerosol composition is model-dependent and affected by emission inventories, aerosol aging, removal processes, and optical assumptions. I suggest that the authors explicitly state that the OM/BC partitioning from CAMS should be interpreted as model-based evidence rather than a direct observational constraint.

The manuscript states that the proposed multi-instrumental approach helps reduce uncertainties in aerosol radiative forcing assessments. However, no radiative transfer calculations or radiative forcing analyses are presented in this study. I therefore suggest that the authors either moderate this statement or expand the discussion to better explain how the retrieved aerosol vertical structure and transport characteristics could contribute to future radiative forcing studies. In particular, the vertical distribution of absorbing aerosols, aerosol aging, particle size evolution, and smoke-cloud interactions are all known to substantially influence radiative forcing estimates and deserve further discussion. The authors may also consider placing their work within the context of recent studies that have demonstrated how observational constraints on aerosol absorption, particle size, and microphysical properties can substantially improve radiative forcing estimates. Relevant examples include Tiwari et al. (2023), Guan et al. (2026a), Guan et al. (2026b), and Tiwari et al. (2026).

Related to the previous point, the Introduction and Discussion could benefit from a broader connection to recent work on aerosol absorption, emissions, and radiative forcing uncertainty. The authors may consider citing additional studies that discuss biomass-burning emissions, absorbing aerosol optical properties, and observational constraints on aerosol forcing, for example: Wang et al., 2025; Wang et al., 2026(pre-print); Samset et al., 2018; Kahn et al., 2023; Konovalov et al., 2021.

Some of the terminology should be made more consistent. For example, the manuscript uses “biomass burning aerosols”, “smoke”, “elevated smoke”, “continental pollution mixed with smoke”, “organic matter”, and “black carbon” in different places. These terms are not interchangeable. I suggest that the authors clarify when they refer to observed aerosol type, model aerosol species, or inferred source category.

The paper would benefit from a clearer statement of novelty. The study is based on a specific case in September 2017, but the authors should more clearly explain what is new compared with previous studies of Southern African smoke export, AEROCLO-sA, BiBAC, and previous Reunion Island aerosol climatology. Is the novelty mainly the vertical layering over Reunion, the combined FTIR-FLEXPART validation, the water vapor-aerosol coupling, or the identification of mixed South American and Southern African transport? This should be stated more directly.

Overall, I find the manuscript scientifically interesting and suitable for publication after revision. The multi-instrumental dataset is valuable, and the study provides useful evidence for vertically structured biomass-burning aerosol transport over the SWIO. However, the authors should clarify the definition of “aerosol river”, better discuss uncertainties in source attribution and lidar retrievals, avoid over-interpreting Ångström exponent and CAMS speciation, and soften or better support statements related to radiative forcing.
Reference
Guan, L., Cohen, J. B., Wang, S., Tiwari, P., Liu, Z., & Qin, K. (2026). Improving aerosol absorption estimates via size-resolved constraints based on AERONET and in situ measurements. Geophysical Research Letters, 53, e2025GL117418.
Guan, L., Cohen, J. B., Wang, S., Tiwari, P., Liu, Z., Li, Z., & Qin, K. (2026). In-tandem multi-waveband particulate absorption and size observations yield substantial changes in radiative forcing over industrial Central China. Atmospheric Chemistry and Physics, 26, 3107–3123.
Tiwari, P., Cohen, J.B., Wang, X. et al. Radiative forcing bias calculation based on COSMO (Core-Shell Mie model Optimization) and AERONET data. npj Clim Atmos Sci 6, 193 (2023). https://doi.org/10.1038/s41612-023-00520-1
Tiwari, P., Cohen, J. B., Gao, H., Lu, L., Wang, J., Dubovik, O., and Qin, K.: Microphysical evolution and column loading drive nonlinear regional contrast in black carbon top-of-atmosphere forcing, Atmos. Chem. Phys., 26, 9149–9180, https://doi.org/10.5194/acp-26-9149-2026, 2026.
Wang, S., Cohen, J. B., Guan, L., Lu, L., Tiwari, P., & Qin, K. (2025). Observationally constrained global NOx and CO emissions variability reveals sources which contribute significantly to CO2 emissions. npj Climate and Atmospheric Science, 8, 87.
Wang, S., Guan, L., Cohen, J. B., et al. Observationally constrained aerosol absorption and radiative effects based on multi-platform observations (SSRN preprint, 2026). SSRN preprint
Samset, B. H., Stjern, C. W., Andrews, E., et al. (2018). Aerosol absorption: Progress towards global and regional constraints. Current Climate Change Reports, 4, 65–83.
Kahn, R. A., et al. (2023). Reducing aerosol forcing uncertainty by combining models with satellite and suborbital observations. Reviews of Geophysics, 61, e2022RG000785.
Konovalov, I. B., Golovushkin, N. A., Beekmann, M., Panchenko, M. V., & Andreae, M. O. (2021). Inferring the absorption properties of organic aerosol in Siberian biomass burning plumes from remote optical observations. Atmospheric Measurement Techniques, 14(10), 6647-6673.

Reply
Citation: https://doi.org/10.5194/egusphere-2026-3070-RC1
RC2: 'Comment on egusphere-2026-3070', Anonymous Referee #2, 22 Jul 2026 reply

Please find attached my comments on a separate file.

Reply

Citation: https://doi.org/10.5194/egusphere-2026-3070-RC2
RC3:
'Comment on egusphere-2026-3070', Anonymous Referee #3, 10 Aug 2026 reply
This manuscript presents a case study examining vertical distributions of biomass burning aerosol observed at a site on Réunion Island, originating from South America and South Africa. The authors combine ground-based measurements, satellite observations, and numerical modeling/reanalysis with two stated objectives: (1) to identify the vertical structure, optical properties, and origins of BB aerosol layers reaching Réunion Island, and (2) to assess whether a multi-instrumental approach can help reduce uncertainties in aerosol radiative forcing assessments in the SWIO, an understudied region. The study compiles a suite observational and model datasets, including AERONET AOD, MODIS FRP and AOD, CALIOP aerosol extinction and typing, ground-based lidar and FTIR profiles of aerosol, water vapor, and CO, CAMS total and speciated AOD, and FLEXPART simulated CO transport.
While the data coverage is comprehensive and the objectives are well sound, the manuscript does not clearly articulate the novel scientific findings that emerge from this case. Additionally, the representativeness of an event at one location for the broader SWIO region during the BB season remains uncertain. It also remains unclear how the multi-instrument/model framework leads to reduced uncertainties in aerosol radiative forcing estimates, particularly given that no uncertainties of BB aerosol radiative forcing estimated by prior studies are discussed and no radiative forcing calculations based on this study are presented.
General comments:
Observational gaps for quantifying BB aerosol radiative forcing: South America and South Africa are major biomass burning regions in the world with burning taking place every year in the burning season. Long-range transport of smoke aerosols and CO to downwind regions, including the SWIO, has been well documented by observations from satellites and field campaigns in the past three decades. The manuscript would benefit from a clearer exposition of existing observational gaps that currently limit quantification of aerosol radiative forcing in the region, and a more explicit description of how this case study addresses those gaps. Without such context, and without any radiative forcing analysis, the stated objectives appear only partially addressed.
Aerosol and water vapor: For the case study, the discussion of aerosol and water vapor vertical profiles requires more clarity. Water vapor concentration does not directly represent aerosol hygroscopicity, and the physical meaning of “decoupled” aerosol and water vapor concentrations needs further explanation. If the intent is to infer aerosol hygroscopic growth or microphysical impacts, additional supporting analysis is needed.
Source attribution: Source attribution using FLEXPART CO provides useful insights of airmass origins, but the manuscript overlooks contributions from Madagascar, a known and geographically closer BB source region to Réunion. Both MODIS FRP and CALIOP data presented in this paper show burning activity over Madagascar; its exclusion could bias interpretations of source contributions. I suggest add a “rest of the world” CO tracer in FLEXPART to better interpret the mixture in observed CO profiles. Furthermore, CO has a significantly longer lifetime than aerosols, so the contributions of CO from South America and South Africa source regions are expected to be different from that of BB aerosols. The analysis would be more straightforward by incorporating aerosol-specific tracers.
CAMS aerosol bias attribution: The authors attribute the high bias in total AOD from CAMS to either an overestimation of sea salt or to the spatial averaging of CAMS AOD from its original 0.75° resolution to 1°. However, this explanation appears subjective and is not supported by evidence within the manuscript. It would be helpful for the authors to clarify why such regridding is necessary and to provide quantitative analysis or references to support this claim. Additionally, CAMS omits nitrate aerosol, which is a significant component in many regions. Therefore, the statement that CAMS “accurately reproduces AOD across various aerosol types” is an overstatement and should be revised to reflect this limitation. Furthermore, since CAMS total AOD is a reanalysis product that assimilates satellite AOD, it should consider whether the high bias is due to satellite AOD being systematically higher than AERONET AOD. The manuscript should discuss this possibility and, if feasible, provide a comparison between satellite and ground-based AOD measurements to clarify the source of the bias.
Overall, the case study is informative and demonstrates the value of multi-platform observations, yet the manuscript would be considerably strengthened by extending the analysis beyond a single event. Incorporating longer-term records (at least covering one BB season), combined with model simulations, would enable a more robust and generalizable assessment of BB aerosol transport and radiative impacts in the SWIO. As it stands, the manuscript provides useful descriptive analysis but falls short of substantiating the broader claims regarding radiative forcing uncertainties and regional representativeness.
My recommendation and suggestions:
The topic of the manuscript is interesting and suitable for publication, but it requires major revision. My suggestions for improvements include:
Clearly articulate the new scientific insights gained from this case study. Specify what distinguishes these results from previous work on biomass burning aerosol transport to the southwestern Indian Ocean (SWIO).

Address how observations at Réunion Island relate to the broader SWIO region and the entire biomass burning season.

Demonstrate how combining ground-based, satellite, and modeling data reduces uncertainties in aerosol radiative forcing assessments. If possible, include quantitative comparisons or uncertainty estimates.

Expand Source Attribution to include CO from other regions/sources.

Extend Analysis beyond a single case by incorporating longer-term records or additional case studies, leveraging the available measurement and modeling datasets to provide a more robust assessment of BB aerosol transport and radiative impacts.

Reply
Citation: https://doi.org/10.5194/egusphere-2026-3070-RC3

Viewed

Total article views: 144 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
89	34	21	144	19	17

HTML: 89
PDF: 34
XML: 21
Total: 144
BibTeX: 19
EndNote: 17

Views and downloads (calculated since 03 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	76	27	11	114
Aug 2026	13	7	10	30

Cumulative views and downloads (calculated since 03 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	76	27	11	114
Aug 2026	13	7	10	30

Viewed (geographical distribution)

Total article views: 94 (including HTML, PDF, and XML) Thereof 94 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 13 Aug 2026

Short summary

Smoke from Southern African and South American wildfires can travel for thousands of kilometers before reaching the Indian Ocean and Reunion Island. In this study, we use instruments measuring from the ground, from satellites, as well as atmospheric models, to track these smoke plumes and identify their origins and properties. Our findings should help future research focusing on the impact of these plumes on the atmospheric temperature.


Total:	0
HTML:	0
PDF:	0
XML:	0