the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Sep 2026
Observation of dark brown carbon in urban aerosols and its contribution to surface dimming
Abstract. Dark brown carbon (d-BrC) aerosols are commonly associated with biomass burning and wildfire emissions, yet their occurrence and climatic impacts in urban environments remain elusive. Southeast Texas – a hub for petrochemical industries at the intersection of continental and marine air masses – is strongly influenced by aerosol-induced surface dimming and extreme meteorological events. Here, the in-situ photoacoustic measurements reveal that over 80 % of aerosol light absorption at blue and near-infrared wavelengths is due to non-black carbon (BC) aerosols. The computer-controlled scanning electron microscopy and particle-scale electron energy loss spectroscopy demonstrates the presence of refractory d-BrC in the coastal urban atmosphere as an absorption contributor. The diurnal pattern of BC and non-BC absorption suggests association of the non-BC aerosols such as d-BrC particles with local non-biomass burning combustion sources, potentially including traffic, flares, and industrial activities. Observationally constrained radiative-transfer calculations show that d-BrC contributes 40 % of daytime-mean surface dimming and 50 % of the total top-of-atmosphere radiative forcing attributable to light-absorbing carbon. Such radiative changes that influence local meteorology may also impact ~25 % of the global population residing in coastal urban regions like Houston. Our findings demonstrate the presence and climatic importance of d-BrC in a coastal urban environment in addition to wildfire plumes, highlighting the need for incorporation of d-BrC in aerosol-climate models.
- Preprint
(2255 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 22 Oct 2026)
- RC1: 'Comment on egusphere-2026-4745', Anonymous Referee #1, 30 Sep 2026 reply
-
RC2: 'Comment on egusphere-2026-4745', Anonymous Referee #2, 05 Oct 2026
reply
The manuscript presents an interesting and potentially important investigation of strongly absorbing organic aerosol in a coastal urban environment. The combination of in-situ photoacoustic measurements, SP2 measurements, solvent extraction, thermal-optical carbon analysis, electron microscopy/EELS, and radiative-transfer calculations is ambitious. In particular, the particle-resolved STEM-EELS observations provide valuable evidence for the presence of refractory, strongly absorbing organic particles with optical properties consistent with dark brown carbon (d-BrC) in the Houston coastal urban atmosphere.
After reading the manuscript carefully, I feel it is generally well written and logically organized, and I consider the topic is relevant to ACP. However, several assumptions underlying the quantitative attribution of aerosol absorption and the subsequent estimate of the climatic importance of d-BrC require substantially stronger justification. My first concern is the subtraction approach used to derive non-BC absorption in Sect. 2.3. A second concern is the way the abundance of d-BrC is inferred for the radiative-transfer calculations. Fortunately, these two issues are not closely connected: the RF/dimming calculations do not appear to use the non-BC absorption residual from Sect. 2.3 directly, but instead rely on EELS-derived optical properties and a d-BrC/BC abundance inferred from thermal-optical carbon fractions. Thus, the uncertainties in the subtraction method do not mathematically propagate directly into the RF calculations; nevertheless, the radiative calculations introduce another major uncertainty through the assumed correspondence between OC4/ELVOC and d-BrC. I therefore recommend reconsideration after major revisions. The principal issues are as follows.
Major comments
1. My primary concern is the robustness of the non-BC absorption derived by subtraction. The central result in Sect. 3.1—that approximately 80% and 88% of total absorption at 405 and 1047 nm, respectively, is due to non-BC—is obtained from babs,non-BC =babs,TOT-babs,BC, where babs,TOT is directly measured by the IPNs, whereas babs,BC is inferred from SP2 rBC mass using an assumed MAC, an absorption-enhancement factor, and an assumed spectral dependence. The residual therefore combines uncertainties from different instruments with several microphysical assumptions. The authors currently show uncertainties associated with selected values of the enhancement factor and AAE, but this does not constitute a complete uncertainty propagation or an observational closure of the subtraction approach. The authors should quantify the combined uncertainty in babs,non-BC, including at minimum SP2 mass uncertainty and detectable-size range, IPN measurement uncertainty and cross-wavelength comparability, MAC variability, absorption-enhancement variability, and the wavelength dependence of coated BC absorption. Recent work has demonstrated that BC size, morphology, mixing state, and coating properties can produce substantial and nonlinear changes in both the magnitude and spectral dependence of BC absorption (e.g., Beeler and Chakrabarty, 2022; Guan et al., 2026; Tiwari et al., 2026; Shen et al., 2026). Therefore, a fixed enhancement factor combined with a power-law AAE should not be treated as a uniquely determined representation of ambient Houston BC. A sensitivity/bounding analysis using physically plausible BC microphysical states would be much more convincing. At the same time, the very large reported residual should be considered quantitatively. If the reported mean non-BC fractions of 80% and 88% are taken at face value, eliminating the residual through BC absorption alone would require BC absorption to be approximately 5 and 8.3 times the current central estimates at 405 and 1047 nm, respectively. Thus, it may be unlikely that normal BC microphysical variability alone explains the entire residual. This makes a rigorous bounding analysis particularly valuable: the authors should demonstrate whether a large non-BC contribution remains under the full physically plausible uncertainty space.
2. Sect. 2.3 requires clarification of the definition of MACBC. The manuscript adopts a MAC of 11.25 m² g⁻¹ at 550 nm and subsequently multiplies this value by an enhancement factor f=1.5±0.48. However, 11.25 m² g⁻¹ is commonly obtained by applying an enhancement factor of 1.5 to a fresh/uncoated BC MAC of approximately 7.5 m² g⁻¹. For example, the closely related treatment used by Chakrabarty et al. (2023) explicitly describes 11.25 m² g⁻¹ as the MAC after accounting for a mean absorption enhancement of 1.5. The authors should therefore state unambiguously whether 11.25 m² g⁻¹ in Eq. (2) represents an intrinsic/uncoated BC MAC or a MAC that already includes coating enhancement. If the latter is intended, multiplication by f=1.5 in Eq. (2) would double-count the coating effect. There is also an apparent incorrect reference. The cited Beeler and Chakrabarty (2021), DOI 10.1088/1748-9326/ac29ea, concerns PM2.5 exposure and SARS-CoV-2 transmission and does not provide the BC optical basis related here. The authors should identify and cite the appropriate optical study, potentially including Beeler and Chakrabarty (2022), and carefully verify all literature support for the adopted MAC and enhancement factor. Furthermore, Wu et al. (2018) reported Eabs,550=1.50±0.48 for a particular observational dataset in the Pearl River Delta, with substantial temporal variability. It should not automatically be considered a universal enhancement factor applicable to Houston.
3. Equation (1) defines only babs,non-BC. However, several figure captions and parts of the discussion subsequently refer to this residual as “BrC absorption”. This attribution is not established by subtraction alone. The STEM-EELS results independently demonstrate the presence of particles having properties consistent with d-BrC, which is an important strength of this study. Nevertheless, observing d-BrC particles does not establish that the entire ensemble-scale non-BC residual is caused by d-BrC, nor does it quantitatively determine the d-BrC contribution to that residual. Other absorbing non-BC components may contribute. The terminology throughout Figs. 2 and 5 and the corresponding discussion should therefore remain “non-BC absorption” unless a quantitative attribution to BrC/d-BrC can be demonstrated. The authors should also compare the wavelength dependence of the ensemble-scale residual with the particle-scale d-BrC optical properties derived by EELS. This is particularly important because the inferred non-BC contribution remains extremely large at 1047 nm. The acknowledgement of this limitation in the Conclusions is useful, but a caveat introduced at the end of the paper does not replace quantitative treatment of the uncertainty in the central results.
4. The radiative-transfer calculation appears to be independent of the non-BC residual discussed above. Instead, d-BrC optical properties are derived from STEM-EELS, while the d-BrC/BC abundance is inferred from thermal-optical carbon measurements by assuming that OC4 represents ELVOC/d-BrC and EC represents soot BC. This assumption requires much stronger justification because it directly controls the reported d-BrC/BC number ratio and hence the calculated 40% surface-dimming and 50% TOA-RF contributions. In particular, OC4 is an operational thermal fraction, not a chemical or optical identification of d-BrC. Ma et al. (2016) showed an empirical relationship between thermal carbon fractions and volatility, but thermal-optical analysis itself does not provide molecular or chromophore information. More importantly, the volatility mapping in Ma et al. was developed using an IMPROVE-type thermal protocol, whereas the present manuscript states that NIOSH 5040 was used. The temperature definitions of OC1–OC4 differ substantially between these protocols. The authors therefore need to demonstrate that the cited OC4-to-volatility relationship is transferable to their specific NIOSH protocol. Even if OC4 could be reasonably associated with very-low-volatility organic material, ELVOC is not synonymous with d-BrC. Indeed, the manuscript's own Particle 3 illustrates that low-volatility organic particles need not have d-BrC-level absorptivity. I therefore recommend that the authors either provide an independent quantitative constraint on d-BrC abundance or perform an explicit sensitivity analysis in which only a variable fraction of OC4 is assigned to d-BrC. Without such an analysis, the reported 40% and 50% contributions should be presented as model scenarios or potential contributions based on an assumed upper-level d-BrC abundance, rather than as observationally established contributions.
5. The radiative calculations use the refractive index derived from particle-scale EELS measurements. However, only a very limited number of representative particles are presented, and Particle 2 appears to play a central role in defining the optical properties of TRACER d-BrC. The authors should report how many particles were analyzed successfully by low-loss EELS, how particles were selected, and the distributions of k, n, and AAE among the analyzed organic particles. The variability should be propagated into the radiative calculations. Using the optical properties of one strongly absorbing particle as representative of the ambient d-BrC population would otherwise substantially bias the inferred RF and dimming. This issue is particularly important because Particle 3 in the same dataset is considerably less absorbing despite its low volatility. The current dataset itself therefore demonstrates that the optical properties of refractory organic particles are heterogeneous.
6. Sect. 2.7 states that “the absorption contribution of this dimming can be calculated by setting the surface albedo to zero”, and Sect. 3.4 subsequently reports that absorption accounts for 91% and 85% of the total surface dimming for the two d-BrC cases. I do not understand how setting surface albedo to zero isolates aerosol absorption from aerosol scattering. Even over a perfectly absorbing surface, aerosol scattering still modifies the downward direct and diffuse solar irradiance at the surface. Therefore, setting α=0 does not, by itself, constitute an absorption-versus-scattering decomposition. The authors should provide a derivation and validation of this procedure. If the goal is to isolate absorption-induced dimming, an appropriate radiative-transfer counterfactual should be used—for example, calculations in which aerosol absorption is selectively removed while extinction/scattering properties are otherwise controlled. The reported 91% and 85% values should not be retained without demonstrating that the decomposition is physically valid.
7. The model assumes externally mixed d-BrC and BC and treats their forcings as additive. This assumption deserves sensitivity testing in the Houston environment. The manuscript itself later discusses evidence from TRACER for internally mixed carbonaceous particles and emphasizes that secondary coatings modify BC absorption. The statement that external mixing is “likely true” because independent d-BrC and BC particles have been observed in wildfire emissions does not establish the mixing state of the Houston aerosol. The authors should distinguish (i) mixing between d-BrC particles and BC particles and (ii) coatings surrounding individual BC cores, since these are different concepts. At minimum, the impact of plausible mixing states, particle size distributions, PBL heights, and hygroscopic growth on Fig. 6 should be quantified.
8. The synchronized diurnal variations of BC and non-BC absorption are suggestive of common/local combustion influences, but they do not uniquely identify traffic, flaring, or industrial emissions, and they cannot exclude transported biomass-burning aerosol. Accordingly, statements such as “direct observation of anthropogenic d-BrC” should be tempered unless supported by source-specific chemical tracers or transport analysis. Similarly, the suggestion that TRACER d-BrC “explains” or “bridges” the factor-of-two modeled-versus-observed dimming discrepancy reported during the earlier TexAQS campaign is too strong. These are different observational periods, and the current calculations are not a closure analysis of the historical TexAQS radiation budget. The result may indicate a plausible missing absorber, but this should be stated as a hypothesis rather than a demonstrated explanation.
Minor and technical comments
- In the Introduction, describing carbonaceous aerosols including all BrC as having “strong light absorption” is too broad. BC is a strong broadband absorber, whereas BrC spans orders of magnitude in absorptivity.
- Please provide more detail on the wavelength-specific IPN configuration and calibration. Were three separate IPNs used for the three wavelengths? How was inter-instrument comparability established? The described filtered-air autozero procedure addresses baseline drift but does not by itself document absolute absorption calibration.
- In Fig. 1a, “Abundance of fossil fuel emitted organic carbon” is misleading because the plotted quantity is a fossil-fuel-to-biomass-burning OC ratio. The caption and axis description should reflect the actual quantity.
- The A(700) baseline correction in the solvent-extraction analysis should be justified. Please state the assumption regarding BrC absorption at 700 nm and explain how negative corrected absorbance values, if any, were treated at wavelengths above 700 nm.
- Line 312 contains “coupled with coupled with”.
- Please define operationally how “d-BrC” is distinguished from “strongly absorbing BrC (S-BrC)”. If the terms are used interchangeably in this study, the optical thresholds and supporting literature should be clearly stated.
- Please check the sign convention for surface dimming throughout the text and figures. The Methods define dimming as a negative perturbation, whereas some reported magnitudes are presented without a minus sign.
Citation: https://doi.org/10.5194/egusphere-2026-4745-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 325 | 161 | 68 | 554 | 63 | 64 |
- HTML: 325
- PDF: 161
- XML: 68
- Total: 554
- BibTeX: 63
- EndNote: 64
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General comment
This manuscript combines ensemble optical measurements and particle-resolved characterization to investigate dark brown carbon (d-BrC) in a coastal urban environment. The identification of particles with optical properties consistent with d-BrC is potentially important. However, the evidence for their presence is stronger than the quantitative constraints on their contribution to ambient absorption and radiative effects. In particular, the inferred non-BC absorption depends on the assumed BC optical properties, whereas the radiative estimates additionally rely on using OC4 as a proxy for d-BrC abundance. These two inference pathways are not sufficiently connected or validated. I recommend major revision, focusing on the robustness of the BC subtraction, the representativeness of the particle measurements, and the uncertainty in the reported radiative contributions.
Specific comments
1. In Section 2.3, please clarify whether the adopted MAC of 11.25 m2/g represents bare BC or already includes absorption enhancement, and therefore whether multiplying it by an additional factor of 1.5 is justified. The cited Beeler and Chakrabarty (2021) reference also appears unrelated to BC optical properties and should be corrected. Beyond the specified ranges of enhancement factor and AAE, the uncertainty analysis should include MAC uncertainty, SP2 calibration and undetected mass, IPN measurement uncertainty, and the effects of detection-limit screening. Please provide uncertainty intervals for the residual absorption at both wavelengths.
2. In sections 2.3 and 3.1, a wavelength-independent enhancement factor and a single power-law AAE may not adequately represent the absorption spectrum of aged, coated BC. Please test whether physically plausible BC size distributions and mixing states can reproduce the observed absorption at both 405 and 1047 nm simultaneously, preferably using available campaign constraints. A useful diagnostic is the MAC required for SP2-measured rBC to explain all measured absorption. Importantly, the reported residuals are large, particularly at 1047 nm; their explanation by BC variability should be quantitatively tested rather than either assumed or dismissed.
3. In Section 3.2, the STEM-EELS results support the presence of d-BrC-like particles, but do not by themselves establish their contribution to ensemble absorption or exclude substantial bias in the BC subtraction. Please report the number of particles optically characterized, their selection criteria, and the variability and uncertainty of their retrieved optical properties. The manuscript should clearly distinguish this population from the much larger CCSEM dataset. Please also explain how collection, heating, and electron-beam exposure may affect the representativeness of the characterized particles.
4. In sections 2.7 and 3.4, the radiative estimates rely on assigning OC4 to d-BrC, although low volatility does not uniquely imply strong absorption, as illustrated by Particle 3. Please evaluate how the results change when only a fraction of OC4 is assigned to d-BrC. In addition, compare filter-based EC with contemporaneous SP2 rBC, since different BC mass measurements are used in the subtraction and radiative calculations. Most importantly, please forward-calculate absorption from the adopted BC and d-BrC populations and compare it with IPN observations. This absorption-closure test is needed to connect the particle-based radiative calculations with the ensemble residual.
5. In Section 3.4, abstract and conclusions, the reported 40% contribution to surface dimming and 50% contribution to TOA radiative effects should be presented as conditional estimates for the modeled BC+d-BrC system, with uncertainties reflecting the abundance proxy and optical assumptions. Please also revisit the claim that setting surface albedo to zero isolates absorption-driven dimming, because aerosol scattering remains active under this condition. Finally, clarify which scenario and averaging period produce the reported 1.46 W/m2 boundary-layer heating and demonstrate consistency with the stated surface and TOA energy budgets. The main conclusions should distinguish robust evidence of d-BrC occurrence from less-constrained estimates of its climatic importance.