Dark brown carbon from wildfires: a potent snow radiative forcing agent?

Wildfire smoke plumes contain light-absorbing carbonaceous aerosols that, upon dry or wet deposition on snow, reduce snow albedo and enhance melt by absorbing additional solar energy. Pristine snow has very high reflectance in the UV and visible wavelengths; deposition of short-wavelength-absorbing aerosols yields a relatively larger drop in albedo at those wavelengths. Beyond carbonaceous aerosols, post-fire enhanced dust emissions and charred debris can also halve local snow albedo near burned forests, and other non-wildfire factors (volcanic ash, mineral dust, snow algae) further darken snow. Wildfire frequency, extent, and smoke transport over thousands of kilometers have increased, leading to persistent snow darkening impacts lasting years after single events. Field studies across regions such as the Andean glaciers, Tibetan Plateau, Arctic, and Antarctica demonstrate significant snow radiative forcing from carbonaceous aerosols; in the Himalayas, snow darkening has advanced melt by ~20 days and increased snow and air temperatures. Midlatitude snowpacks, near expanding wildfire activity and dense human populations, are particularly vulnerable. Biomass burning emits OC and BC; BC is the strongest absorber per particle, while OC dominates by mass (>95%). Within OC, brown carbon (BrC) absorbs mainly in UV–VNIR, with absorption strengths spanning orders of magnitude; the upper range, termed dark brown carbon (d-BrC), is refractory, water-insoluble, BC-like in some properties, and abundant in wildfire smoke (observed up to four times more abundant than BC). Field observations show high water-insoluble organic carbon (WIOC)/BC ratios in snow, implying substantial insoluble BrC including d-BrC that is less effectively scavenged and thus persists at the surface to reduce albedo. While radiative effects of water-soluble BrC have been studied, the role of insoluble, strongly absorbing d-BrC in snow remains poorly quantified. Prior albedo and forcing studies, and models like SNICAR, have largely focused on BC and weakly absorbing OC, potentially misattributing forcing among species. This study addresses the gap by explicitly representing d-BrC optical properties and abundance in an aerosol–snow radiative transfer framework to quantify its contribution to snow radiative forcing relative to BC.

• Numerous studies link deposition of light-absorbing particles (LAPs) from wildfires and other sources to snow albedo reduction and accelerated melt across global cryospheric regions (Andes, Tibetan Plateau, Arctic, Antarctica).
• Observations in the Indian Himalayas show aerosol-induced snow darkening advancing melt by ~20 days and warming snow and air temperatures.
• Post-fire environments also elevate dust emissions and deposit charred debris, further reducing albedo; snow darkening also results from volcanic ash, mineral dust, and snow algae.
• Carbonaceous aerosol emissions from biomass burning consist of BC and OC; while BC is the strongest absorber, OC dominates by mass. BrC, especially the strongly absorbing, water-insoluble fraction d-BrC, has recently been shown to dominate shortwave absorption in wildfire smoke and to co-emit with BC.
• Field observations report high WIOC/BC ratios in midlatitude glacier snowpacks, highlighting substantial insoluble organic matter that encompasses d-BrC.
• SNICAR and related radiative transfer models have been used extensively to simulate spectral albedo and LAP forcing, but most past work emphasized BC or water-soluble OC. Declining BC deposition since the 1980s due to emission controls further motivates reassessment of other absorbers.
• Few prior studies quantify the OC contribution to snow radiative forcing, and none explicitly include d-BrC despite its observed dominance in wildfire absorption, risking misattribution when translating albedo perturbations to forcing.
• Deposition experiments and remote sensing comparisons (e.g., Beres et al., Hadley & Kirchstetter) provide controlled and observational constraints for modeling LAP impacts in snow, often using SNICAR for interpretation or meta-analysis.

Study design: A coupled aerosol–snow radiative transfer approach was employed to quantify the radiative forcing (RF) of d-BrC relative to BC in snow, using laboratory- and field-constrained optical properties and SNICAR-ADv4 simulations. A representative midlatitude mountain site, Mount Olympus, Washington (47.8152 N, 123.7047 W), was used for irradiance and solar geometry forcing.

Aerosol optics and size distributions:

• d-BrC absorption strength parameterized via the brown–black continuum (Saleh et al.), characterized by imaginary refractive index at 550 nm, k550, spanning 0.1–0.25. Field-observed wildfire d-BrC refractive index (k550 ≈ 0.19) was also used.
• Wavelength dependence of k given by k(λ) = (λ/550 nm)^w × k550, with w = 7.6e^(-12.4 k550) (per Saleh parameterization).
• BC refractive indices followed Flanner et al., based on Chang and Charalampopoulos.
• Mie theory (PyMieScatt) applied to compute ensemble optical properties (single-scattering albedo, asymmetry parameter, mass extinction coefficient) for each species across 400–1000 nm (10 nm resolution). Size distributions assumed lognormal: • d-BrC (continuum tests): dg = 80 nm, σg = 1.8, ρp = 1.27 g cm^-3 • d-BrC (wildfire-observed): dg = 170 nm, σg = 1.5, ρp = 1.4 g cm^-3 • BC: same lognormal parameters as SNICAR defaults for comparability.
• Computed optical properties were validated against Flanner et al. for consistency.

Snow properties and configuration:

• SNICAR-ADv4 Adding–Doubling solver used to simulate spectral albedo of optically thick snow (semi-infinite; no ground transmission).
• Snowgrain shapes assumed hexagonal plates; two representative grain sizes for snow age: • Fresh snow: 100 µm • Aged snow: 1000 µm
• Aerosols assumed externally mixed, homogeneously distributed in the surface snow layer.

Aerosol concentrations in snow:

• BC concentration range: 10^-1 to 10^2 ppb (ng g^-1), representing low to high pollution based on global observations.
• d-BrC concentrations considered up to an order of magnitude higher than typical BC due to observed greater abundance in wildfire smoke (4–10× BC by mass/number). Simulations examined d-BrC:BC ratios of 0, 4, and 10 alongside three base BC levels (10, 10^1, 10^2 ppb).

Spectral-to-broadband albedo and irradiance:

• Spectral albedo computed for 400–1000 nm; LAP impact above 1000 nm deemed negligible due to snow’s strong intrinsic absorption; below 400 nm the irradiance fraction is small (~7%).
• Broadband albedo calculated by irradiance-weighted integration over 400–1000 nm using a midlatitude surface irradiance spectrum; fraction of broadband irradiance in VNIR (400–1000 nm) set to 0.69.

Radiative forcing calculations:

• Instantaneous radiative forcing (IRF) at time t computed as IRF = [α_noLAP(SZA,d) − α_LAP(SZA,d,k550,[LAP])] × Fb(SZA) × f_VNIR, where SZA varies with time and location (pvlib python).
• Lookup tables of broadband albedos vs. SZA, grain size, LAP concentration, and k550 enabled efficient time-resolved IRF calculations.
• Temporal averaging: IRFs computed at 30-minute intervals throughout the year and averaged to obtain daily mean, seasonal mean (NH spring: Mar–May; summer: Jun–Aug), and annual mean RFs for each scenario.

Comparisons and validation:

• Modeled spectral and broadband albedos for mixtures compared against field spectroradiometer observations (e.g., Kaspari et al., Mount Olympus; Li et al., Tibetan Plateau) for plausibility across BC concentration ranges.
• Scenario analyses included sensitivity to d-BrC k550, snow aging, aerosol concentration, and d-BrC:BC abundance ratios.

Assumptions and scope:

• External mixing; homogeneous vertical and horizontal aerosol distribution within the optically thick snow layer; planar, level snow surface; temporally invariant concentrations for the base annual cycle at the site; focus on midlatitude solar geometry and irradiance.

• d-BrC substantially enhances snow radiative forcing when coexisting with BC, owing to its higher abundance in wildfire emissions.
• Annual mean RF enhancement due to adding d-BrC at 4× the BC abundance is 0.6–17.9 W m^-2 across scenarios, corresponding to a 1.6–2.1-fold increase compared to BC-only deposition.
• Abundance-weighted comparisons using field-constrained d-BrC optics (k550 ≈ 0.19): • At d-BrC:BC = 4:1, d-BrC annual mean RF is 22–46% greater than co-emitted BC RF (range corresponds to BC concentrations from ~1 to 10 ppb; effect diminishes as BC concentration increases). • At d-BrC:BC = 10:1, d-BrC annual mean RF is 83–187% greater than co-emitted BC RF.
• Concentration scaling examples in aged snow: • BC-only: increasing BC from 10 ppb to 100 ppb raises annual mean RF from ~1.57 to 8.51 W m^-2 (≈4.4×). • Adding 100 ppb d-BrC to 10 ppb BC increases RF to ~5.28 W m^-2 (≈2.36× over 10 ppb BC alone), demonstrating strong enhancement from d-BrC addition.
• Seasonality at Mount Olympus: Seasonal mean RFs exceed annual means by approximately +30% in spring and +70% in summer due to more favorable solar geometry and higher irradiance.
• Snow aging amplifies LAP impacts: Aged snow (1000 µm grains) exhibits larger albedo reductions and higher RF than fresh snow (100 µm).
• Spectral behavior: d-BrC depresses albedo more strongly at shorter visible wavelengths, aligning with higher incident irradiance fractions, thus disproportionately increasing RF.
• Mixture effects are non-additive due to shadowing among particles; RF from combined BC+d-BrC is less than the arithmetic sum of their separate RFs in clean snow.
• Modeled spectral/broadband albedos and RFs overlap with ranges reported from field campaigns and remote sensing across multiple midlatitude regions, supporting plausibility of simulations.

The research question asked whether dark brown carbon from wildfires is a significant snow radiative forcing agent relative to BC. The simulations demonstrate that although BC is more absorbing per particle, d-BrC’s much higher abundance in wildfire smoke makes it a potent snow forcer. When realistic d-BrC:BC ratios are included, RF increases by factors of about 1.6–2.1 over BC-only scenarios, and the annual mean RF attributable to abundance-weighted d-BrC alone can exceed that of BC. This directly addresses the prior omission of insoluble, strongly absorbing d-BrC in snow RF assessments and shows that models focusing solely on BC likely underestimate wildfire-induced snowmelt forcing. The spectral analyses connect stronger short-wavelength absorption by d-BrC with larger albedo perturbations where solar irradiance is high, explaining its disproportionate role. The seasonality results highlight particularly large enhancements in spring and summer, coinciding with critical melt periods for water resources. Agreement in magnitude with literature-reported RFs across diverse midlatitude sites strengthens confidence in the approach. Overall, incorporating d-BrC into snow radiative transfer calculations refines attribution of LAP-induced snowmelt and implies that regional and global models need to include d-BrC to avoid low bias in cryospheric forcing estimates.

This study quantifies the radiative impact of wildfire-emitted dark brown carbon in snow and shows that d-BrC, despite weaker per-particle absorption than BC, is a major snow radiative forcer due to its higher abundance. Including d-BrC with BC increases annual mean radiative forcing by 0.6–17.9 W m^-2 and by a factor of 1.6–2.1 over BC-only deposition, with particularly strong effects in aged snow and during spring and summer. The framework integrates laboratory- and field-constrained optics into SNICAR-based radiative transfer to produce time-resolved and seasonal/annual RF estimates at a representative midlatitude site, with results consistent with observed ranges. These findings suggest current climate models likely underestimate LAP-induced snowmelt by neglecting d-BrC. Future research should: (1) extend analyses to high-latitude and multilayer snow–firn–ice systems; (2) incorporate internal mixing, surface slope effects, and cryoconite features; (3) evaluate temporal variability of LAP concentrations tied to wildfire seasonality; and (4) assess coatings and lensing effects that may further enhance absorption. Improved representation of d-BrC will refine projections of glacier mass loss, snowmelt timing, and associated water resource and climate feedbacks.

• Aerosol–snow mixing was assumed external and homogeneous; internal mixing and vertical heterogeneity, which can enhance absorption, were not explicitly modeled.
• The snow surface was treated as optically thick, planar, and level; effects of inclined terrain and sub-surface features like cryoconite holes were not included.
• Temporal variability of LAP concentrations was not represented; concentrations were held constant over the annual cycle despite real-world wildfire seasonality and deposition pulses.
• The analysis focused on 400–1000 nm; contributions outside this band were neglected due to low irradiance below 400 nm and strong intrinsic snow absorption above 1000 nm.
• d-BrC has not yet been directly measured in snow; its presence and concentrations were inferred from WIOC measurements and co-emission with BC in smoke, introducing uncertainty in snow concentrations.
• Only midlatitude irradiance and solar geometry (Mount Olympus) were considered; while results are extendable to other midlatitudes, polar regions may differ more substantially.
• External coatings (e.g., sulfates) and lensing effects that can enhance absorption were not included.
• Other LAPs (mineral dust, volcanic ash, algae) can dominate at some sites; the framework is most applicable where wildfire aerosol deposition is the primary darkening agent.