Attribution of aerosol light absorption to black carbon, brown carbon, and dust in China – interpretations of atmospheric measurements during EAST-AIRE

There's a small black box that atmospheric scientists carry into the field. It pulls in air, shines light through whatever particles are suspended in it, and tells you how much of that light gets absorbed. Near Beijing, during the spring of 2005, one of these instruments was running continuously, logging absorption numbers every two minutes. The box is called an aethalometer, and it is very good at its job. What it cannot do is tell you who is responsible. Black carbon from diesel engines, brown carbon from coal fires, and mineral dust blown in from the Gobi — they all absorb light, and to most instruments they look identical. That's the problem Yang and colleagues set out to solve. The reason this matters goes straight to climate. When aerosol particles scatter sunlight, they tend to cool the surface. When they absorb it, they warm the atmosphere. The net effect depends entirely on which particles are doing the absorbing — and those particles have different sources, different atmospheric lifetimes, and very different policy levers. Black carbon calls for targeting diesel combustion or open biomass burning. Brown carbon points to coal. Dust requires different thinking entirely. If you cannot separate them, you cannot tell your climate model or your regulator what is actually happening. The diagnostic Yang and colleagues relied on is called the Ångström exponent — the rate at which absorption changes with wavelength. You can express it as a power law: absorption is proportional to wavelength raised to a negative power, where that power is the absorption Ångström exponent. Fresh soot-like black carbon has an exponent close to one, meaning absorption falls off relatively gently as you move toward longer wavelengths. Brown carbon has a much larger exponent — typically around three to four — meaning it absorbs far more strongly in the ultraviolet. Dust has a similarly large exponent driven by iron oxides. The trick is that most monitoring instruments measure absorption at only one or two wavelengths, throwing away the very spectral information that would let you read those fingerprints. To capture that information, Yang and colleagues set up at Xianghe, a site roughly seventy kilometers east of Beijing, during the East Asian Study of Tropospheric Aerosols: an International Regional Experiment — EAST-AIRE — from March 2 through 26, 2005. The site was surrounded by more than twenty visible smokestacks in every direction, plus residential coal burning, vehicular traffic, and periodic dust plumes from the northwest. Their instrument suite paired a nephelometer measuring scattering at three wavelengths with a Particulate Soot Absorption Photometer, or PSAP, at 567 nanometers, and critically, a Magee AE31 aethalometer logging absorption at seven wavelengths spanning 370 to 950 nanometers. Chemical composition came from a continuous carbon analyzer for elemental and organic carbon, plus size-segregated filter samples. After correcting the aethalometer for filter loading effects and multiple scattering in the filter matrix, the corrected aethalometer absorption agreed with the PSAP to within five percent. The wavelength-resolved picture they needed was now in hand. With that data, they approached air mass classification like forensic work. Each air mass type left a specific combination of fingerprints: single scatter albedo, which is the fraction of total light extinction due to scattering rather than absorption; the Ångström exponents; particle size distributions; and chemical tracers. Dust periods had coarse particle fractions and calcium concentrations more than one standard deviation above the campaign medians. Fresh industrial chimney plumes were flagged by elevated nitrogen oxide to carbon monoxide ratios combined with low winds. Coal-derived pollution showed sulfur dioxide to carbon monoxide ratios two standard deviations above the median. Biomass burning correlated with elevated single scatter albedo above 0.87 near observed fires. Background air required scattering below fifty inverse megameters and wind speeds above eight meters per second. These classifications mapped onto strikingly different optical signatures. Fresh chimney plumes had the lowest single scatter albedo — 0.83 at 550 nanometers — and the flattest absorption spectrum, with an Ångström exponent near 1.35, the closest to pure soot behavior in the dataset. Coal pollution was even darker, with albedo dropping to 0.80, but its absorption Ångström exponent was higher, around 1.46, consistent with additional brown carbon contributing at shorter wavelengths. Dust events showed the highest albedo at 0.90 and an absorption Ångström exponent of 1.82, reflecting the wavelength-hungry iron oxides in mineral particles. Biomass burning sat in between, with albedo near 0.89. Each combination told a different story about what was in the air. Now came the decomposition. Yang and colleagues used 950 nanometers as an anchor: at that wavelength, brown carbon and dust absorption are negligible, so the signal is essentially pure black carbon. They extrapolated black carbon absorption from 950 nanometers to shorter wavelengths using an Ångström exponent of one. Subtracting this black carbon contribution from the measured total gave a residual absorption spectrum with an exponent near 3.5 — strongly wavelength-dependent and correlated with organic carbon at a correlation coefficient of 0.95. That residual is brown carbon. Dust absorption was estimated separately using particle size distributions measured by an aerodynamic particle sizer, run through Mie scattering theory with a fitted complex refractive index of 1.53 minus 0.0023i at 550 nanometers. The mass absorption efficiencies — relating how much light each absorber blocks per gram — came out as follows at 550 nanometers. Black carbon: 9.5 square meters per gram. Brown carbon: 0.5 square meters per gram, explicitly a lower limit because the normalization used total organic carbon, not just the absorbing fraction. Dust: 0.03 square meters per gram. Black carbon is roughly nineteen times more efficient per gram at absorbing mid-visible light than brown carbon, and more than three hundred times more efficient than dust. That hierarchy confirms what the field has long suspected — but the wavelength dependence adds something new. At 370 nanometers, in the near-ultraviolet, brown carbon could account for roughly thirty percent of total absorption. At mid-visible wavelengths, it still contributed more than ten percent. Dust was generally small across the campaign average, below about five percent at 370 nanometers, but during actual dust events it rivaled black carbon. This matters for models, which typically treat absorbing aerosol as essentially all black carbon. That approach artificially inflates the apparent black carbon mass absorption efficiency. If you divide total absorption by elemental carbon without removing brown carbon and dust, you get 11.3 square meters per gram — roughly twenty percent higher than the true value of 9.5. That error propagates directly into estimates of aerosol radiative forcing. The propagated uncertainty on the true black carbon mass absorption efficiency was twenty-six to thirty-one percent across wavelengths, reflecting uncertainties in the dust refractive index, the fraction of organic carbon that actually absorbs, and gaps in size-resolved chemical sampling below about 0.7 micrometers. Yang and colleagues are candid about these limits. The brown carbon mass absorption efficiency is a lower bound. The dust optical properties were inferred rather than directly measured. These are not fatal flaws — the correlations and multi-instrument cross-checks hold up — but they define what future campaigns would need to close. Direct spectral absorption measurements for brown carbon, better-constrained dust refractive indices, and finer size-resolved chemical sampling are the gaps this work maps out. What the EAST-AIRE results demonstrate is that the single absorption number from a standard filter instrument is a mixture, and decomposing that mixture is not just academically interesting. In coal-dominated airsheds like the one around Xianghe in 2005, a substantial fraction of absorbing mass lies outside black carbon. Climate models that ignore that will misrepresent how much ultraviolet radiation reaches the surface, misattribute warming, and potentially misguide the emission controls designed to address it. Yang and colleagues gave the field a method — wavelength-resolved absorption plus particle size plus careful air mass classification — that makes those contributions legible. The aethalometer's single number becomes three separate answers, each pointing in a different direction. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.