The measurement site, a small terminal depression called L’Bour near M’Hammid El Ghizlane, sits at the transition between the southern foothills of the Atlas Mountains and the open Saharan Desert. It is an ideal natural laboratory for dust emission studies. Mornings were typically calm, but as solar heating destabilized the near-surface air in the afternoons, winds strengthened and frequent local saltation events injected fresh dust into the lower atmosphere. On two days, 4 and 6 September 2019, the campaign captured intense cold-pool outflows known as haboobs, dramatic walls of dust that provided the researchers with abundant material for their instruments. This diurnal rhythm of quiet mornings and dusty afternoons allowed the team to build a rich dataset spanning a wide range of emission intensities.

The instrumentation combined a multi-angle integrating nephelometer, which measured light scattering at three wavelengths and eight angular positions, with a seven-wavelength dual-spot aethalometer that tracked absorption from the near-ultraviolet to the near-infrared. A Fine Dust Aerosol Spectrometer recorded particle size distributions in 63 logarithmically spaced bins covering optical diameters from roughly 0.2 to 19 micrometers. During the first part of the campaign, the optical instruments sampled through a PM2.5 inlet; later they switched to a PM10 inlet, a change that profoundly influenced what the team observed, because most of the mass of mineral dust resides in the coarse fraction. All measurements were averaged into 15-minute intervals to smooth out instrumental noise while preserving the turbulent structures that carry dust aloft.

The headline optical results are strikingly consistent with previous Saharan dust studies, which is itself reassuring for the field. During dust events, the single scattering albedo, the ratio of scattering to total extinction, was 0.95 for PM2.5 particles at 370 nanometers, rising to 0.97 at 660 nanometers, with slightly lower values of 0.94 and 0.96 for the PM10 fraction. These numbers tell us that Saharan dust is predominantly a scatterer of sunlight, but that its absorption is far from negligible, especially at short wavelengths. The absorption Ångström exponent, a measure of how strongly absorption varies with wavelength, reached values of up to 2.5, far above the values near 1 that characterize soot-like anthropogenic particles. This steep spectral signature is the fingerprint of iron oxide minerals such as hematite and goethite, which absorb much more strongly in the ultraviolet and blue than in the red and near-infrared.

Particle size emerged as a dominant control on nearly every optical property the team measured. The asymmetry parameter, which describes whether light is preferentially scattered forward or backward, was 0.7 at 520 nanometers for PM2.5 dust, and the backscatter fraction was correspondingly low at 0.11, both hallmarks of large particles that funnel scattered light strongly into the forward direction. The scattering Ångström exponent dropped to negative values when coarse dust dominated, a well-known indicator of coarse-mode particles. Mass absorption efficiency fell from about 0.30 to 0.15 square meters per gram at 370 nanometers as particle size increased, showing that larger grains absorb less light per unit mass. Meanwhile, the mass scattering efficiency shifted its peak toward longer wavelengths for larger particles, exactly as scattering theory predicts when the size parameter, the ratio of particle circumference to wavelength, grows beyond the regime of wavelength-independent scattering.

The most consequential part of the study, however, goes beyond describing the dust. The researchers used their combined measurements to retrieve the imaginary part of the complex refractive index, the quantity usually written as k that directly controls how much light a particle absorbs. Retrieving k from field data is an exercise in conditional reasoning: the answer depends on assumptions about particle density, the real part of the refractive index, and, critically, the shape of the particles. The team therefore ran their inversion repeatedly under a menu of prescribed shape assumptions, including perfect spheres, biaxial spheroids using aspect ratios measured on individual particles by electron microscopy during the campaign, the spheroid distribution commonly used in AERONET satellite retrievals, triaxial ellipsoids, and a set of six irregular particle geometries from the MOPSMAP modeling database.

The outcome was unambiguous. For central assumptions about density and the real refractive index, the retrieved k at 520 nanometers ranged from approximately 0.00104 for spherical particles to approximately 0.00145 for triaxial ellipsoids, a difference of about 40 percent. Non-spherical shapes systematically demanded larger values of k to reproduce the same observed absorption-to-scattering ratio, and the more elongated or irregular the assumed particle, the higher the retrieved absorption. The team found a nearly linear relationship between k and a shape descriptor, the ratio of volume-equivalent to cross-section-equivalent particle radius, with the effect weakening at longer wavelengths. Particle density assumptions also mattered, because density controls how the measured size distribution is converted between optical and aerodynamic diameters, whereas the assumed real refractive index had only a weak influence across its plausible range.

To put their in-situ results in context, the researchers compared them with two decades of dust-filtered retrievals from the AERONET network, using stations across western and eastern North Africa, the Sahel, the Arabian Peninsula, central Asia, and India, including Ouarzazate, the station closest to their field site. The comparison showed qualitative agreement with the Ouarzazate climatology, especially when the more realistic non-spherical shape assumptions were used, with median values converging closely at 550 and 675 nanometers. But the analysis also exposed pronounced regional variability in dust absorption, with retrieved k increasing from western toward eastern North Africa and further still across Asia, likely reflecting regional differences in iron oxide content and mineralogy. Differences in particle size distributions between surface measurements and column-integrated retrievals, along with dust ageing and residual aerosol contamination, complicate any strict comparison, so the authors frame the AERONET analysis as context rather than validation.

Why does all this matter? Mineral dust is one of the most abundant aerosol types in the atmosphere by mass, and its direct radiative effect is estimated at a central value of minus 0.40 watts per square meter, but with an uncertainty of plus or minus 0.30 that leaves open the possibility of a substantially different sign or magnitude. Dust absorption is the least constrained ingredient in these calculations, and most existing measurements come either from laboratory chambers using resuspended soils or from remote sensing of dust that has already been transported, aged, and mixed with other aerosols. Direct measurements of dust at the point of emission remain scarce. By showing that the shape assumptions embedded in retrieval algorithms and climate models can shift inferred absorption by tens of percent, the study provides a concrete explanation for part of the persistent spread among dust absorption estimates, and it suggests a practical path forward: past datasets built on spherical assumptions could potentially be reinterpreted using the relationship between particle asphericity and retrieved k.

The authors are careful about the limitations of their work. Their retrieval is explicitly conditional, restricted to the PM2.5 dust-event period, and their observing system simulation experiments revealed that size-dependent distortions in the measured size distributions, arising from different sampling inlets and possible particle losses, cannot be fully eliminated even by basing the cost function on the ratio of truncated scattering to extinction. Future campaigns, they argue, should prioritize co-located and size-consistent measurements, well-characterized inlets, and independent constraints on particle density and morphology. Still, the message of this desert field campaign is clear and broadly resonant: the light absorbed by a billowing Saharan dust plume depends not only on what the dust is made of, but on what shape we assume its grains to be, and getting that shape right may be essential for pinning down one of climate science’s most stubborn uncertainties.

Subject of Research: Optical properties and shape-dependent refractive index retrieval of freshly emitted Saharan mineral dust

Article Title: Optical properties and shape-dependent retrievals of the imaginary part of the refractive index of freshly emitted Saharan dust

Article References: Yus-Díez, J., Escribano, J., Pandolfi, M., Alastuey, A., González-Flórez, C., Gonzalez-Romero, A., Gonçalves Ageitos, M., Ivančič, M., Klose, M., Kandler, K., Obiso, V., Panta, A., Reche, C., Rigler, M., Querol, X., & Perez García-Pando, C. (2026). Optical properties and shape-dependent retrievals of the imaginary part of the refractive index of freshly emitted Saharan dust. Atmospheric Chemistry and Physics, 26(19), 14531-14557. https://doi.org/10.5194/acp-26-14531-2026

Image Credits: AI Generated

DOI: 10.5194/acp-26-14531-2026

Keywords: Saharan dust, mineral dust aerosol, refractive index, light absorption, single scattering albedo, particle shape, iron oxides, AERONET, Morocco, dust emission, radiative forcing, aerosol optics