Gentle Winds, Not Calm Air, Sweep the Most Dust From the Sky

When a dust storm dies down, the atmosphere is left loaded with suspended mineral particles, and what happens next determines how far that dust travels, where it lands, and how it alters climate and ecosystems. For decades, atmospheric models have treated the removal of this dust — dry deposition — as a process that either stays flat or speeds up as wind picks up from still air. A new wind-tunnel study published in Atmospheric Chemistry and Physics turns that assumption on its head. Ruochen Jia of the Institute of Applied Ecology of the Chinese Academy of Sciences, Zhimin Liu, Jianqiang Qian, and their colleagues found that fine and medium-sized dust particles do not settle most efficiently in calm air, nor in stronger winds. Instead, they accumulate fastest and most abundantly at a gentle breeze of about 2 metres per second — a surprising peak that challenges the monotonic formulas embedded in dust transport models worldwide.

The team carried out their experiments in a closed-circuit recirculating wind tunnel at the Experimental Center of Desert Forestry in Dengkou, Inner Mongolia. The facility’s test section stretches 13 metres, with a cross section 2 metres wide and 2 metres high, and four axial fans can drive wind speeds continuously from 0 to 20 metres per second. The researchers used dust sieved from the upper layers of local cropland soil, dried it in an oven at 105 degrees Celsius to standardize moisture, and restricted it to particles smaller than 63 micrometres. This choice matters: the finest of these particles are the ones most relevant to human health and long-range atmospheric transport, yet they are also the hardest to pin down in deposition measurements.

The experimental protocol unfolded in two stages. First, a strong-wind loading phase at 14 metres per second for roughly five minutes eroded the soil bed and filled the tunnel with a recirculating cloud of airborne dust. Then the soil source was sealed under steel plates to prevent further emission, and the wind was dropped to one of four target levels: 0, 2, 4, or 6 metres per second. Twenty-five cylindrical acrylic collection trays, each with a 10-centimetre radius, were laid out on the tunnel floor. Five trays were retrieved at each of five sampling times — 30, 60, 90, 120, and 180 minutes — and their captured dust was weighed on an analytical balance accurate to a tenth of a milligram. Crucially, no sticky coatings or adhesives were used, so the trays recorded net accumulation: the balance between particles arriving, rebounding, and being swept away again.

To make sense of the accumulating masses, the researchers fitted a first-order kinetic model in which the retained mass approaches an asymptote exponentially. The fit yields four interpretable quantities: the asymptotic retained mass, the effective net-accumulation rate coefficient, a characteristic timescale equal to the reciprocal of that coefficient, and the initial net accumulation flux, which is the product of the asymptote and the rate coefficient. The model performed remarkably well, with coefficients of determination between 0.889 and 0.992, and three-quarters of the fits exceeding 0.95. This means a single effective timescale captures the whole deposition process for every combination of wind speed and particle size tested — a useful simplification for anyone trying to parameterize dust removal.

The results revealed a striking pattern. In still air, fine particles accumulated slowly and almost linearly: the characteristic timescale for particles below 2.5 micrometres was about 4.2 hours, and for those below 10 micrometres about 1.4 hours. Coarse particles, by contrast, reached their asymptotes within roughly an hour, dominated by straightforward gravitational settling. When the wind was raised to 2 metres per second, everything changed. The characteristic time for the finest particles collapsed from 4.2 hours to just 0.3 hours — an order-of-magnitude acceleration — while the rate coefficient jumped from 0.24 to 3 per hour. The initial net accumulation flux was highest at 2 metres per second for all size classes from PM2.5 through PM30, and quadratic fits across the four wind speeds suggested a model-implied optimum somewhere between 2 and 3 metres per second.

At higher winds, the picture reversed. At 4 metres per second, deposition remained fast but the ultimate retained mass fell by roughly 25 to 60 percent below still-air values, especially for fine and medium particles. At 6 metres per second, the suppression became dramatic: the asymptotic retained mass for PM2.5 and PM10 dropped by about an order of magnitude relative to calm conditions, indicating near-complete inhibition of effective accumulation even though the rate coefficients stayed high. An auxiliary screening test identified an apparent critical wind-speed range of 5.87 to 6.01 metres per second for mobilizing dust from the tray surface, placing the 6-metre-per-second treatment right at the edge of re-entrainment. In other words, stronger winds kept particles aloft longer and stripped away what had landed, while a gentle breeze delivered particles to the surface without dislodging them.

The particle-size distributions of the retained dust added a fascinating clue about mechanism. At 2 metres per second, the median diameter of accumulated material grew from 7.1 to 9.8 micrometres between the half-hour and three-hour samples, and the modal diameter shifted from 4.3 to 18.3 micrometres. The researchers suggest this effective coarsening may reflect shear-induced aggregation: moderate turbulence increases collisions between dust particles, allowing small ones to stick together into transient clusters that settle faster. This interpretation aligns with a rich literature on cohesive sediments in two-phase flows, where aggregate size consistently peaks at intermediate shear — strong enough to bring particles together, too weak to break the clusters apart. The geometry is suggestive too: the volume associated with the observed diameter shift corresponds to roughly two to three similarly sized primary particles merging, though the authors caution that size-selective deposition alone could also produce the pattern.

Quantified against a non-decreasing benchmark built from the same data, the departure from conventional expectations is substantial. For the PM2.5 through PM40 classes, the observed initial flux at 2 metres per second exceeded the monotonic benchmark by 15 to 72 percent. At 6 metres per second, the observed flux fell 45 to 91 percent below the benchmark — meaning deposition-only schemes could overestimate net accumulation by factors of roughly 2 to 11 for the affected sizes under such conditions. The physics is consistent with what is known about near-surface turbulence: for 2-to-10-micrometre particles, settling velocities are only fractions of a millimetre per second, so penetration of the thin viscous sublayer above the surface depends on turbulent diffusion rather than gravity. The calculated Rouse numbers, well below unity, confirm that turbulent mixing dominates over settling for these sizes — and at 2 metres per second the turbulence is strong enough to deliver particles downward but too weak to resuspend them.

The implications ripple outward through the global dust cycle. If net dry deposition in source regions under calm post-storm winds is higher than models currently predict, atmospheric dust loads would decline faster than simulated, shortening transport distances and reducing the dust reaching remote oceans and high latitudes. Conversely, where background winds are stronger, resuspension could extend particle lifetimes and enhance long-range transport. This helps explain a persistent bias in climate models, which tend to underestimate near-source deposition of coarse and accumulation-mode particles while overestimating dust burdens far downwind. The authors are careful about limitations: the experiments used a single agricultural soil dust, tray geometry and smooth acrylic surfaces may influence rebound, soil moisture and particle shape were not varied, and the inferred optimum rests on only four wind-speed levels. Still, the size-resolved kinetic parameters they report — asymptotic mass, rate coefficient, timescale, and initial flux — offer concrete benchmarks for building coupled deposition-retention-resuspension schemes. The gentle breeze, it turns out, is the atmosphere’s own dust vacuum, and models are only now learning to see it.

Subject of Research: Size-resolved kinetics of mineral dust dry deposition under calm and low-wind conditions

Article Title: Non-monotonic response of net dust deposition kinetics to low wind speeds

Article References: Jia, R., Ba, C., Zhang, C., Kuang, W., Liu, Y., Wang, Z., Liu, F., Xin, Z., Liu, Y., Ma, Q., Zong, L., Liang, W., Qian, J., & Liu, Z. (2026). Non-monotonic response of net dust deposition kinetics to low wind speeds. Atmospheric Chemistry and Physics, 26(19), 14245-14258. https://doi.org/10.5194/acp-26-14245-2026

Image Credits: AI Generated

DOI: 10.5194/acp-26-14245-2026

Keywords: mineral dust, dry deposition, wind tunnel, atmospheric aerosols, dust cycle, turbulence, particle aggregation, re-entrainment, deposition parameterization, atmospheric chemistry, climate modeling, PM2.5