Atmosphere

Plastic falls out of the sky. Collectors at 11 western US national parks and wilderness areas averaged 132 microplastic particles per square meter per day (level C)¹. The authors extrapolated this to more than 1,000 metric tons a year on protected land in that region¹. How much the planet sends into the air is disputed. In January 2026 a Vienna team found that three emission inventories overestimated the medians of 2,782 measurements by a factor of 100 to 10,000 (level C)².

Two kinds of numbers: in the air and on the ground

Airborne plastic is reported in two ways that are easy to mix up. A concentration counts particles suspended in a cubic meter of air, sampled by pumping air through a filter. A deposition rate counts particles that settle onto a square meter of surface per day, caught in a funnel or a bucket that collects rain, snow and dry fallout. Wet deposition comes down with precipitation; dry deposition settles on its own.

The largest compilation so far, by Ioanna Evangelou, Silvia Bucci and Andreas Stohl at the University of Vienna, gathered 925 concentration and 1,857 deposition measurements made between 2014 and 2024 (level P)². The median concentration was 0.03 particles/m³ across all sites, 0.08 over land and 0.003 over the open sea, 27 times lower (level C)². The median bulk deposition over land was 0.0006 particles/m² per second, or about 50 particles/m² per day (our conversion from²).

Single sites vary more. Along the southeast coast of China, reported concentrations range from 0.004 to 190 particles/m³ (level C)². Reported deposition runs from 50 particles/m²/day in a Chinese suburb to 3,100 at a British urban site². The global mean concentration, 31 particles/m³, is a thousand times the overall median of 0.03, because a few studies report very high values².

Size limits drive much of this spread. Most compiled studies identified polymers by µ-Raman or µ-FTIR spectroscopy and reported particles from several micrometers to several millimeters². Particles smaller than a few micrometers are usually not counted, although particle numbers rise steeply as size falls^5,6.

Land leads by particle number; the mass split is disputed

Several independent studies find road traffic a major source of airborne plastic over land, at least near roads, in cities and in western US deposition (level B)^4,7,8. A 2026 global inversion by Ondřej Tichý and colleagues, still a preprint, ranks road dust, at 51 ± 31 Gg a year, below agriculture at 170 ± 125 Gg (preprint estimate, level C)³.

Janice Brahney and colleagues fitted a transport model to western US deposition data. It attributed 84% of deposited microplastic to roads, 11% to the ocean and 5% to agricultural dust (model estimate, level C)⁴. Tire and road wear particles were rare in their samples: in their reading, roads mainly resuspend plastic from many products after traffic grinds it down⁴.

Studies that target tire rubber chemically, by pyrolysis gas chromatography with mass spectrometry (py-GC/MS), give a different picture. At a roadside in Leipzig in September 2022, tire wear made up about 65% of the plastic mass in particles below 10 µm (level C)⁷. In Oslo, tire wear particles accounted for more than 90% of airborne microplastic mass (level C)⁸. These studies count tire wear as plastic; the spectroscopy-based western US samples rarely identified it⁴.

Textiles shed fibers into air. In Paris fallout, Rachid Dris and colleagues found about 29% of fibers synthetic or mixed (level C)⁹. They estimated that 3 to 10 metric tons of fibers settle on the Paris area each year⁹.

The ocean's role is disputed (level D). Two earlier models fitted to western US deposition data made sea spray the largest source by mass. Brahney's put the ocean at about 8.6 million metric tons (8.6 Tg) a year, 95% range 0 to 22⁴. An inversion by Nikolaos Evangeliou and colleagues put it at 8.9 ± 3.5 Tg of a 9.6 ± 3.6 Tg total (estimates, level C)¹⁰.

Newer models disagree in both directions. A 2025 global model made the ocean about 0.008% of emissions but about 15% of deposition, mainly a sink (estimate, level C)¹¹. The Tichý preprint counters that oceanic emissions are "likely non-negligible": 763 ± 611 of its 1,001 Gg a year³.

Hours to days aloft, as far as the poles

In the Brahney model, particles of about 4 µm and larger stayed up from one hour to 6.5 days (model estimate, level C)⁴. The authors suggest smaller particles could stay up for weeks⁴. In the Vienna model, removal from the air was dominated by settling under gravity rather than by rain².

That is enough to reach remote places: several independent studies find microplastics deposited at remote mountain, Arctic and polar sites far from local sources (level B)^{14,6,15,13,16}. At Bernadouze in the French Pyrenees, Steve Allen and colleagues measured deposition of about 365 particles/m²/day (5 µm and larger) from November 2017 to March 2018 (level C)¹⁴. Air-mass trajectories indicated transport over up to 95 km; the authors note that their data cannot prove long-range transport¹⁴.

At the Pic du Midi Observatory (2,877 m), also in the Pyrenees, pumped air from the free troposphere held 0.09–0.66 particles/m³ from June to October 2017 (concentration, µ-Raman, level C)¹⁶. Back-trajectories crossed the Atlantic and the Mediterranean, which the authors read as intercontinental and transoceanic transport¹⁶.

Snow from the Alps to the Fram Strait held microplastics, with Arctic samples reaching 14,400 particles per liter at 11 µm and above (FTIR imaging, level C)⁶. Cloud water on Japanese summits from 1,300 to 3,776 m held 6.7 to 13.9 particles per liter, sized 7.1–94.6 µm (level C)¹⁷.

Over the Norwegian Sea up to Bear Island, py-GC/MS found up to 37.5 ng/m³ of microplastic in air, with tire wear reaching 35 ng/m³ (level C)¹². On Ross Island, Antarctica, snow at all 19 sites sampled in late 2019 held microplastics, about 29 particles per liter of meltwater at 50 µm and above (µFTIR, level C)¹⁸. Back-trajectories allow transport of up to 6,000 km, but nearby research stations are a possible local source¹⁸.

How much reaches the ocean from the air is a model question. A 2020 simulation estimated that about 100,000 metric tons of tire wear particles a year are deposited in the world ocean (estimate, level C)¹⁹. From the few marine deposition measurements, all near coasts, Evangelou and colleagues infer that most suspended microplastic settles close to its land sources².

Models versus measurements: the 2026 correction

Evangelou, Bucci and Stohl ran three published emission inventories through the FLEXPART transport model, matching each simulated value to the size range of its measurement². All three overestimated the global median concentrations and deposition by two to four orders of magnitude (level C)². Over land, simulated median concentrations were 105 to 1,506 particles/m³, against a measured 0.08 (level C)². Before scaling, even the best correlation with measured concentrations was weak (r = 0.33)².

With data too sparse for an inversion, two global factors cut land emissions to 0.13% of the bottom-up inventory and raised ocean emissions 3.6 times². The scaled total is 0.0045 Tg (range 0.0006–0.0089), about 4,500 metric tons a year for 5–100 µm (estimate, level C)². Land emits over 20 times more particles than the ocean, but the ocean emits more mass, about 4,000 against 500 t a year (estimate, level C)². That split rests on assumed emission size distributions².

The emission magnitude is disputed (level D). Soroush Abolfathi (University of Warwick) called the study careful but warned that size cut-offs and mass-to-number conversions can shift results by orders of magnitude²⁰. Richard Lampitt (National Oceanography Centre) argued that the compiled techniques differ too much to compare and that model inputs often come from chains of other models²⁰. Isidro Pérez (University of Valladolid) cited measurements spread over years of changing emissions, sparse sites and an outlier-sensitive correlation coefficient²¹.

The Tichý preprint is a testable alternative, not a midpoint: fitted to remote deposition data, it estimates about 1 Mt a year for 5–250 µm³. On 3 November 2026 it was under review for Atmospheric Chemistry and Physics, with public discussion open until 11 November (level P)³.

By our comparison, Brahney's extrapolation puts more than 1,000 t a year onto western US protected lands¹, twice the scaled model's 500 t a year for all land emissions worldwide². It is not like for like. Land deposition also receives plastic from the sea and particles lifted again after settling, and the size windows differ (counts from about 4 µm converted to mass, against 5–100 µm). This is a diagnostic to explain, not a refutation of either study.

Indoor air and what we breathe

People in industrialized countries spend about 90% of their time indoors⁵. Small studies report more plastic indoors than outdoors. In the Paris area, Dris and colleagues counted 1.0–60.0 fibers/m³ in two apartments and an office, against 0.3–1.5 fibers/m³ outdoors; about a third of the indoor fibers were petrochemical (level C)²².

Smaller particles change the scale. Nadiia Yakovenko and colleagues sampled three apartments and two cars in southern France, identifying particles of 1–10 µm by Raman spectroscopy (level C)⁵. Median concentrations were 528 particles/m³ in homes and 2,238 in car cabins; 94% of particles were smaller than 10 µm⁵. Combining these with published data, they estimate that an adult inhales about 68,000 particles of 1–10 µm a day indoors, 100 times earlier estimates extrapolated from larger sizes (estimate, level C)⁵.

By mass, the Leipzig study estimated an inhaled dose of about 2.1 µg of airborne plastic a day for someone breathing that roadside air (estimate, level C)⁷. The WHO concluded in 2022 that inhalation exposure data were too limited for a human risk assessment²³. What inhaled particles do in the body is covered on Human Body; filtration indoors is covered on Solutions by System.

What is uncertain about nanoplastics in air

For nanoplastics below 1 µm, the airborne record is close to empty. We found no peer-reviewed, size-resolved number concentrations of nanoplastics below 1 µm in outdoor air (level E). Search scope: this site's registry, Crossref and web search, papers from 2019–2026 on "airborne nanoplastics", "atmospheric nanoplastics" and "TD-PTR-MS air", November 2026. The models in the 2026 comparison stop at 5 µm, and the compiled measurements start at several micrometers².

The nanoplastic evidence comes from snow and ice. Dušan Materić and colleagues used thermal desorption proton transfer reaction mass spectrometry (TD-PTR-MS) to find PET nanoplastics in Alpine snow (level C)²⁴.

At the Sonnblick Observatory, 3,106 m up in the Austrian Alps, surface snow in 2017 held on average 46.5 ng/mL, implying deposition of about 42 kg per km² a year (level C)¹⁵. Greenland firn averaged 13.2 ng/mL and Antarctic sea ice 52.3 ng/mL, with tire wear nanoparticles identified in Greenland (level C)¹³. The sea-ice value does not show whether the particles arrived by air or from seawater.

These are masses, not counts or sizes. The Leipzig study reports a fine fraction by mass, not resolved below 1 µm⁷. Sub-micrometer estimates in the indoor study are extrapolations from a size distribution, not measurements⁵.

Illustration

Measured vs modeled microplastic concentrations in air, global medians

Data table

label	value	unit	kind	scope
Over land, measured median	0.08	particles/m³	measurement	2014–2024, coastal sites excluded
Over land, model (TD-B inventory)	105	particles/m³	model estimate	lowest of three inventories over land; TD-B is the Brahney et al. 2021 top-down estimate
Over land, model (TD-E inventory)	1506	particles/m³	model estimate	highest of three inventories over land; TD-E is the Evangeliou et al. 2022 top-down estimate
Over ocean, measured median	0.003	particles/m³	measurement	2014–2024, open ocean
Over ocean, model (lowest inventory)	0.2	particles/m³	model estimate	range across three inventories
Over ocean, model (highest inventory)	13.9	particles/m³	model estimate	range across three inventories

Global microplastic emissions into the air: published estimates

Data table

label	value	unit	size_range	basis	source
Evangeliou et al. 2022, top-down inversion	9.6	Tg/yr	microplastics up to 250 µm	total of all sources; western US deposition data, global extrapolation; ocean 8.9 ± 3.5 of the 9.6 ± 3.6; microfibers (6.5 ± 2.9 Tg/yr) estimated separately and not included	S-199
Brahney et al. 2021, ocean source only	8.6	Tg/yr	model bins 0.3–70 µm; observations 4–250 µm	ocean term only, 95% range 0–22, not a total; western US deposition data, global extrapolation; roads (0.096) and agricultural dust (0.069) add about 2%	S-050
Tichý et al. 2026, preprint	1	Tg/yr	5–250 µm	remote deposition data, Bayesian inversion; not peer reviewed; under review for Atmospheric Chemistry and Physics on 3 November 2026	S-198
Fu et al. 2023, GEOS-Chem	0.32	Tg/yr	0.3–70 µm	second-hand: as reported in Evangelou et al. 2026; the primary paper was not checked by this site	S-052
Evangelou et al. 2026, scaled	0.0045	Tg/yr	5–100 µm	2,782 measurements, two global scaling factors; 90% confidence interval 0.0006–0.0089	S-052

Airborne plastic indoors and outdoors: what each study counted

Data table

label	value	unit	statistic	size_range	place	source	low
Outdoor air over land, global median	0.08	particles/m³	median	several µm and larger	283 locations worldwide, 2014–2024	S-052
Outdoor air, Paris area (all fibers)	1.5	fibers/m³	range; value is the high end	fibers, natural and synthetic	one outdoor site, published 2017	S-203	0.3
Indoor air, Paris area (all fibers)	60	fibers/m³	range; value is the high end	fibers; about one-third petrochemical	two apartments and one office	S-203	1
Homes, median	528	particles/m³	median	1–10 µm, Raman	three apartments, Toulouse, seven samples	S-063
Car cabins, median	2238	particles/m³	median	1–10 µm, Raman	two cars, southern France, five samples	S-063

Questions

Where to go next

What happens to plastic once it lands at sea is on Oceans. The instruments and their size limits are explained on How We Measure It, and open questions on Knowledge Gaps.