the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 May 2026
Global radiative forcing of stratospheric aerosols injected by the 2020 Australian extreme wildfire event
Abstract. During the last decades, extreme wildfires have injected large amounts of biomass-burning aerosols in the stratosphere. Partly composed of carbon, these aerosols absorb incoming solar radiation, inducing changes in the atmosphere's energy balance, and can self-loft to altitudes higher than 30 km, with an increased residence time of several months. In this study we estimate the radiative forcing of stratospheric aerosols from the Australian New Year Super Outbreak (ANYSO) in 2020. We first model individual self-lofting plumes from the Pacific Northwest Event (PNE) in Canada in 2017 and from ANYSO to constrain the aerosols' optical properties. We use observations to track them and model their heating rates. For the PNE plume a Single Scattering Albedo (SSA) of 0.95 is the best estimate to compute the heating rates, as well as a SSA of 0.90–0.95 for the ANYSO plume. Cloud cover and geometrical thickness of the plume have a crucial impact on these computations. We then compute the direct radiative forcing of Southern Hemisphere aerosols injected in the stratosphere by ANYSO in 2020. Cloud cover has a crucial impact on those forcings, especially at the top of the atmosphere (TOA) where it makes values ranging from negative (clear-sky) to positive (all-sky). The global TOA radiative forcing of stratospheric aerosols from ANYSO over 2020 was evaluated at 0.08 to 0.19 W.m−2 in all-sky, and −0.17 to −0.12 W.m−2 in clear-sky. The surface radiative forcing estimate is −0.04 to −0.06 W.m−2 in all-sky and −0.29 to −0.24 W.m−2 in clear-sky.
- Preprint
(2336 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-2006', Anonymous Referee #1, 10 Jun 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2006/egusphere-2026-2006-RC1-supplement.pdfCitation: https://doi.org/
10.5194/egusphere-2026-2006-RC1 -
AC1: 'Reply on RC1', Raphaël Lebrun, 24 Sep 2026
This article presents a comprehensive analysis of the global radiative forcing of carbonaceous aerosols in the stratosphere in the aftermath of the 2019-2020 Australian wildfires. Overall, the paper is well written and documented. It should be published after the authors address the mostly minor comments below.
Thank you for your work and insights on this review. The suggested references were added in the paper, as well as a better explanation regarding the observational quantification of aerosols’ optical properties, and a deeper explanation regarding the spectrally constant refractive index (but spectrally dependent SSA and asymmetry parameter). Please find our responses to your comments below.
Comments
Line 24. Self-lofting across the tropopause was also postulated during the 2017 PNE pyroCb by Torres et al (2020a).The reference was added in the paper line 24.
Lines 39-41. As a general statement, the sentence ‘The remote sensing methods used to observe these stratospheric aerosols are not yet able to precisely quantify their absorption properties’, is not accurate. If the authors refer to a specific remote sensing method used in their analysis, they should mention it. Near UV aerosol optical depth and single scattering albedo for both the PNE and ANYSO events have been retrieved from satellite near UV observations by the DSCOVR- EPIC (Torres et al.2020a) and the Sentinel5-Precursor TROPOMI sensors (Torres et al., 2020b) respectively.
Thank you for this comment. The precise quantification of aerosol absorption by remote sensing is highly challenging, not only for stratospheric aerosols but also for tropospheric aerosols. This precise quantification challenge was the intended focus of the sentence mentioned. We acknowledge, however, that a more elaborated explanation was needed to describe both the current challenges and the achievements in this field. We have therefore revised the text to highlight the broad range of estimates reported in the literature for aerosol SSA, refractive index, and OC/BC ratio, as well as their spectral variability. For instance, we have included the near-UV estimate of SSA(388 nm) = 0.92–0.95 for PNE (2017) from DSCOVR-EPIC (Torres et al., 2020a) line 40.
Line 68. TROPOMI retrievals of AOD and SSA over both cloud-free areas and AOD over cloudy scenes were used to produce an estimated 546 kt of injected aerosol mass above 12 km (Torres et al., 2020b).
The estimation given by this reference was added in the paper, line 74.
Line 79. Add the Taha et al. (2021) reference after extinction profiles.
The reference was added in the paper line 83.
Line 103. Specify the spectral range over which refractive indices are assumed constant.The refractive index is assumed constant both in the longwave [4, 500µm] and shortwave [0.2,4 µm] (added line 122).
Line 105. Discuss the effect of (or clarify) the apparent inconsistency between the assumed constant refractive indices and the spectrally dependent SSA and asymmetry parameter.
The spectral dependence of the SSA and the asymmetry parameter arises not only from variations in the refractive index, but also from the particle size distribution. Even for a spectrally flat complex refractive index, both SSA and the asymmetry parameter will vary spectrally due to the different effect of wavelength to particle size ratio. A sentence was added to clarify this point line 123.Citation: https://doi.org/10.5194/egusphere-2026-2006-AC1
-
AC1: 'Reply on RC1', Raphaël Lebrun, 24 Sep 2026
-
RC2: 'Comment on egusphere-2026-2006', Reed Espinosa, 29 Jul 2026
This manuscript estimates the global direct radiative forcing (RF) of stratospheric smoke injected by the 2019/2020 Australian New Year Super Outbreak (ANYSO). The authors first constrain the aerosol single scattering albedo (SSA) by reproducing the observed self-lofting rates of smoke-charged vortices (SCVs) from the 2017 Pacific Northwest Event (PNE) and ANYSO with the ARTDECO/DISORT radiative transfer model, driven by CALIOP/OMPS-LP plume tracking, ERA5 profiles, and MODIS cloud and surface albedo products. The constrained optical properties are then applied to OMPS-LP extinction profiles to compute Southern Hemisphere forcings over 2020, yielding a global top of atmosphere (TOA) RF in both all-sky and clear-sky conditions.
The paper addresses a real gap, namely the wide spread of published ANYSO forcing estimates. The demonstration that clouds reverse the sign of the TOA forcing is clearly presented, and the literature synthesis in Table 2 is a useful contribution in its own right. However, several aspects of the methodology are incompletely described and some conclusions rest on assumptions that need firmer justification. My recommendation is that the following key points, as well as the specific comments below, be addressed before publication:
1. Matching the observed heating rates cannot uniquely constrain SSA because SSA is degenerate with the assumed geometrical thickness DZ. The plume heating rate scales approximately as SAOD x (1-SSA)/DZ; with SAOD measured, the observed ascent rate constrains only this ratio, and a more absorbing aerosol in a thicker plume fits the observations identically to a brighter aerosol in a thinner one. The authors' own Figure 3 shows halving DZ nearly doubles the heating rates, comparable to the full effect of varying SSA from 0.95 to 0.83 in Figure 4. Yet the DZ values actually used in the Section 3.2 computations are never stated, nor is it explained how they were derived from the CALIOP/OMPS-LP tracking. As presented, the "best estimate" SSAs of 0.95 (PNE) and 0.90-0.95 (ANYSO), and therefore the headline forcing ranges, are conditional on an undocumented DZ. This needs to be documented and the joint SSA/DZ uncertainty at least discussed semi-quantitatively.
2. The cloud representation is not described, despite clouds being central to the paper's main conclusion. Section 2.3 names the MCD06COSP product but says nothing about how monthly 1-degree cloud information enters the 1D radiative transfer: Is cloud fraction handled by weighting the fluxes of separate clear and overcast columns (independent column approximation)? What cloud optical depth and cloud top height are prescribed? Note also that a monthly product cannot represent the diurnal cycle of cloudiness, and MODIS samples only the Terra/Aqua overpass times; the resulting bias in the hourly all-sky computations should be acknowledged.
3. The forcing is defined relative to 2019 extinction profiles, with the full 2020-minus-2019 difference attributed to ANYSO and assigned smoke optical properties. The 2019 SH lower stratosphere was itself perturbed (e.g., the June 2019 Ulawun eruption), so part of the difference field is not smoke. Please discuss this attribution uncertainty, justify the choice of a single anomalous year over a multi-year background climatology, and clarify how grid cells with negative extinction differences are treated.
----------------------------
Specific Comments
----------------------------
ln 42: "OC/BC=2-3%" seems inverted; an organic-to-black-carbon ratio of 2-3% would imply nearly pure BC. Presumably BC/OC is intended.ln 63/68: "occured" -> "occurred"
ln 87: "New-Zeland" -> "New Zealand"
ln 91: "initialy" -> "initially"
Figure 1: How is the in-plume SAOD at 550 nm obtained? CALIOP measures at 532 nm (with a lidar ratio assumption) and OMPS-LP retrievals are at other wavelengths with coarse limb-viewing resolution. The wavelength conversion and, more importantly, the potential low bias for a compact SCV smeared over the OMPS-LP averaging volume should be discussed, since any SAOD bias maps directly onto the retrieved SSA. Also, the x-axis here uses DOY while other figures use month and day. It may be clearer to use a single convention in the manuscript.
ln 103: The refractive index is assumed spectrally constant, but brown carbon absorption increases sharply toward blue/UV wavelengths. Some comment on the sign and rough magnitude of the resulting bias in SW absorption is warranted.
ln 105: "dependant" -> "dependent"
Table 1 vs ln 101: The text states Im(refind) = [0.01, 0.04] is used for both cases, but Table 1 lists 0.01-0.03 for ANYSO, while Figure 4 shows ANYSO curves up to Im(refind) = 0.04. Please reconcile.
Table 1: "Assymetry" -> "Asymmetry"
ln 121: "where dT/dz is the net radiative heating rate" — I think this should read dF/dz (the flux divergence). Please also state the sign convention for F.
ln 132: The base case DZ = 10 km appears thick relative to reported SCV depths of a few km (e.g., Khaykin et al. 2020). Please justify, and state the DZ used for each plume in Section 3.2 (see key point 1).
Figure 4: The observed ANYSO heating rate is drawn as a constant 5 K/day, but the theta trajectory in Figure 1 is not obviously linear early on. Similarly, how was the PNE "observed" curve in this figure derived? Was this a fit, and over what period? To my eye, the ANYSO temperature curve in Figure 1 is actually stepper in the first month, which could potentially better agree with the trend of the corresponding computed heating rates.
ln 160: Aerosol growth is offered as the sole explanation for the trend discrepancy. In addition to my prior comment, a discussion on other plausible explanations (e.g., evolution of DZ, chemistry that could impact composition/RI, etc.) should be added.
ln 175: Please state which OMPS-LP wavelength underlies the 550 nm extinction and how "stratospheric" is defined (tropopause product?).
Figure 5: Please note the cause of the white regions (polar night?)
ln 185: Given the background is fixed to 2019, how are locations/months where 2020 extinction falls below 2019 handled (negative k_ANYSO)?
ln 193: "local SH annual mean" is confusing terminology for a hemispheric average; consider rewording.
ln 211: "solild" -> "solid"
ln 229/231: For the local SCV forcing, over what area is "local" defined (plume core, tracked extent)?
ln 250: "a values" -> "values"
Table 2: Please double check the TOA ERF value in the last row corresponding to Yu et al. (2023).
ln 275: "the the" -> "the"
References: Ohneiser et al. (2020a) and (2020b) appear to be the same paper listed twice, and the 2020b entry contains a " " artifact.
Citation: https://doi.org/10.5194/egusphere-2026-2006-RC2 -
AC2: 'Reply on RC2', Raphaël Lebrun, 24 Sep 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2006/egusphere-2026-2006-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Raphaël Lebrun, 24 Sep 2026
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 334 | 126 | 41 | 501 | 30 | 29 |
- HTML: 334
- PDF: 126
- XML: 41
- Total: 501
- BibTeX: 30
- EndNote: 29
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
