Evidence of free tropospheric and long-range transport of microplastic at Pic du Midi Observatory

The study addresses whether microplastics (MP) are present in the free troposphere (FT) and to what extent they can be transported over long distances. Prior research has documented atmospheric MPs primarily in the planetary boundary layer (PBL), focusing on deposition rates in urban and rural settings and identifying potential regional transport scales. The FT, characterized by higher wind speeds and reduced surface friction, can enable long-range and even intercontinental particulate transport. Demonstrating MP occurrence in the FT would imply that atmospheric MP pollution can reach remote regions (e.g., Arctic, Antarctic, high mountains) and that local emissions may have far-reaching impacts. The Pic du Midi (PDM) high-altitude observatory, largely free from local PBL influences and well-established for FT monitoring, is used to investigate MP occurrence, characteristics, and transport pathways in FT air masses.

Previous atmospheric MP studies have quantified deposition in megacities (Paris, London, Dongguan) reporting 175–1008 MP/m²/day and similar magnitudes in Hamburg for urban and rural sites. Mountain and wilderness studies (French Pyrenees; USA protected areas) showed MP deposition at altitude and over regional scales, with smaller particles transported longer distances and associations with large-scale atmospheric patterns such as the jet stream, suggesting FT influence. Modeling in North America indicated roads, agricultural soils, and oceans as MP sources for remote deposition. Marine aerosol sampling far offshore detected MPs (up to ~1.37 MP/m³), with HYSPLIT trajectories linking sources to East Asia. MPs have been detected in snow from the Alps to the Arctic, implying windborne transport akin to other long-range pollutants (e.g., mercury). Global assessments of tire and brake wear indicate substantial atmospheric transport and ocean deposition of these MPs. Collectively, these studies establish MPs in the PBL and suggest potential for long-range transport if entrained into the FT, motivating investigation of FT occurrence and transport.

Study site: Pic du Midi Observatory (PDM), French Pyrenees (42°56'11" N, 0°08'34" E), 2877 m a.s.l., a Global Atmospheric Watch station with minimal local influence. Sampling: High-volume TISCH PM10 sampler with quartz fiber filters (8×10 inch, 2.2 µm pore). Sampling from 06/23/2017 to 10/23/2017 across 15 sampling periods, each ~7–9 days (mean 8.2 ± 1.2), daily operation from 23:00–16:00 local (shutdown 16:00–23:00). Average sampled air volume 7880 ± 1206 m³ per sample. Two field blanks and two laboratory blanks included. Filters kiln-sterilized; handling under clean conditions (class 100 flow hood), cotton lab coats and nitrile gloves; samples stored at −20°C in sterilized aluminum foil. Analytical procedure: From each filter, three randomly selected 30 mm diameter subsamples were taken (total ~20% of the filter). Subsamples were rinsed with 250 mL MilliQ water; eluates filtered onto 25 mm Whatman Anodisc aluminum oxide filters (0.2 µm). Filters were vacuum-dried. µRaman spectroscopy (Horiba XploraPlus, 785 nm laser, 200–2000 cm⁻¹, 1 µm spatial resolution) was used to identify polymer types and count MPs, supported by open spectral libraries (SLOPP, SLOPP-E) and Spectragryph software. Particle morphology and size (including aerodynamic diameter) were measured using FIJI/ImageJ; fibres defined by length:width ≥3:1; all other shapes as fragments (including films/foams not separately classified). Confocal microscopy provided secondary validation of counts and dimensions. Aerodynamic diameter was computed using standard relationships accounting for polymer density and aspect ratio. Contamination control: Field and lab blanks processed identically; average 8 particles in field blanks and 3–4 in lab blanks; sample counts were blank-corrected. All materials sterilized; minimized airborne contamination. Meteorology and statistics: Local meteorological data (hourly) from P2OA included temperature, humidity, wind speed/direction, precipitation. Correlations (Pearson/Spearman as appropriate; log10 transformations as needed) assessed relationships between MP metrics and meteorology. Transport modeling: HYSPLIT v4 (backward trajectories) and FLEXPART v9.02 (backward particle dispersion) were used to reconstruct 7-day (168 h) air-mass/particle histories at hourly releases across sampling periods (1713 runs). ERA-Interim meteorology (0.5° × 0.5°, 60 vertical layers) drove models. Trajectory elevations above surface level and distances were extracted; PBL/FT mixing events identified and mapped. FLEXPART outputs provided potential emission sensitivities (source–receptor sensitivities) over land and ocean regions to infer likely entrainment areas.

- Detection and concentration: MPs were found in all 15 samples. Air concentrations ranged 0.09–0.66 MP/m³ (mean 0.23 ± 0.15 MP/m³). Thirteen of fifteen samples exceeded 0.1 MP/m³; four exceeded 0.33 MP/m³ (upper quartile). - Size and shape: On average, 51% of MPs were ≤10 µm (range 21–74%; SD ±17%); 96% were ≤20 µm aerodynamic diameter. Particles were predominantly fragments (70%) with fibres at 30%. Larger MPs (<30 µm) were fibres; maximum aerodynamic diameter observed was 53 µm. Smallest spectrally characterized particle was 3.5 µm (study LOQ 3.5 µm). - Polymer composition: LD/HDPE 44%, PS 18%, PVC 15%, PET 14%, PP 10%. Polymer type showed no significant correlation with particle size distribution or shape. - Local meteorology: Mean wind speed 7.9 ± 3.6 m/s (range 1.4–22.6 m/s), predominantly W–SW. Larger MP (>10 µm) counts correlated with maximum northerly wind velocity (r = 0.79, P < 0.05) and generally with stronger winds (r = 0.61, P < 0.05). No significant correlations between total MP counts and local temperature, humidity, precipitation, or wind direction. - FT occurrence and transport: Back trajectories and dispersion modeling indicated that air masses/particles remained on average >2000 m above surface level and traveled at least 275 km over the 168 h period, with average pathway distances ~4550 km (range ~2047–6631 km by sample; maximum straight-line distance up to 10,212 km for A13). Lower MP samples (<0.33 MP/m³) had longer average pathway distances (~4992 ± 1097 km) than higher MP samples by ~1660 km. - PBL/FT mixing influence: Samples with higher MP (>0.33 MP/m³) had lower average trajectory elevations (2747 ± 373 m ASL) and higher frequency of trajectories entering the PBL (9% ± 6%) than lower MP samples (3276 ± 425 m; 2% ± 1%). Correlations: MP count vs fraction of trajectories with PBL/FT mixing r = 0.69 (P < 0.05); MP >10 µm vs PBL mixing frequency r = 0.78 (P < 0.05). Minimum FT-only transport (no PBL influence) prior to arrival was greater for low-MP samples (e.g., 887 km for MP < 0.13 MP/m³) than for high-MP samples (343 km for MP > 0.33 MP/m³). - Source-region indications: All periods showed Atlantic influence. Elevated-MP samples had more trajectories over the Mediterranean and Northern Africa (52%) than low-MP samples (21); conversely, lower-MP samples had a higher proportion over the Atlantic/North America (53% vs 34%). Considering low elevation (<500 m ASL) trajectory points, elevated-MP samples had more over the Mediterranean/North Africa (73% vs 55%). Larger MP (>10 µm) positively correlated with the number of Northern African trajectories (log10(MP > 10 µm) r = 0.8, P < 0.05). FLEXPART potential emission sensitivities highlighted Europe, the Atlantic, Northern Africa, and extending to North America as influential regions. - Comparative magnitudes: PDM FT concentrations are lower than inner-city and coastal sea-spray studies by one to several orders of magnitude; when matched on comparable size thresholds (≥20 µm and ≥58 µm), PDM counts (avg ~0.01 MP/m³ at ≥20 µm; 0 at ≥58 µm) are within an order of magnitude and slightly lower than offshore marine air studies (<1.37 and <0.077 MP/m³, respectively).

Findings demonstrate the presence of microplastics in free-tropospheric air masses at a high-altitude, remote observatory, supporting the hypothesis that MPs can be transported over intercontinental and trans-oceanic distances. The strong association between higher MP concentrations and increased PBL/FT mixing frequency indicates that entrainment processes at lower altitudes (over both land and ocean) significantly influence FT MP burdens observed at PDM, especially for larger particles (>10 µm). While higher MP concentrations tended to accompany more recent or proximal PBL mixing (i.e., lower average trajectory elevations and shorter FT-only distances), lower MP concentrations were associated with longer continuous FT transport without PBL influence. Comparisons with prior studies underscore that PDM FT concentrations are substantially lower than urban and coastal environments, consistent with its remote setting, but still comparable in magnitude to offshore marine aerosols when harmonized by size thresholds. Source-region indications from trajectory and dispersion analyses implicate both continental (Europe, Northern Africa) and marine (Atlantic, Mediterranean) regions, aligning with known areas of high plastic presence and potential aerosolization. Overall, the results highlight the FT as an effective vector for widespread MP dispersal and suggest that both terrestrial and marine sources contribute to FT MP loading via PBL entrainment and subsequent uplift/mixing.

This study provides direct evidence of microplastics in free-tropospheric air masses at the Pic du Midi Observatory, with concentrations of 0.09–0.66 MP/m³ dominated by small particles (≤20 µm) and a mix of common polymers (PE, PS, PVC, PET, PP). Air-mass back trajectories and dispersion modeling indicate long-range transport pathways spanning thousands of kilometers, including trans-oceanic and intercontinental scales. Elevated MP concentrations coincide with greater frequencies of PBL/FT mixing and lower trajectory elevations, suggesting that the location and frequency of PBL entrainment critically influence FT MP loads at high-altitude sites. Implications include the potential for MPs to reach and impact remote ecosystems and human populations far from sources, possibly transporting associated chemicals and microbes. Future work should expand temporal coverage, refine size-resolved and polymer-specific measurements, and advance process-level understanding of MP emission, entrainment, scavenging, deposition, and in-cloud removal, supported by targeted laboratory experiments and multi-model intercomparisons.

- Instrument/sampler bias: The TISCH PM10 inlet relies on particle inertia tuned for mineral dust (~2.65 g/cm³), whereas plastics (~1 g/cm³) may navigate the 180° turn more readily, potentially altering size/type capture efficiencies. No pre-filter was used; potential under/over-capture of certain sizes/shapes may occur. Authors recommend dedicated QA/QC to quantify MP collection efficiency relative to total particulate and PM10 performance for MPs. - Limited field validation for MP atmospheric dynamics: Modeling used passive tracers due to scarce data on MP densities, deposition velocities, wet scavenging/in-cloud processes; assumptions may affect trajectory and source sensitivity interpretations. - Temporal and sampling constraints: Long sampling durations (multi-day) may smooth short-term variability; nightly shutdowns (to accommodate observatory operations) reduce continuous sampling. A limited number of samples (15) and seasonal window (summer–autumn) restrict generalizability. - Blank contamination: Although controlled and corrected, field/lab blanks contributed up to <30% of counts in blanks, introducing uncertainty at low concentrations. - Trajectory/model uncertainties: Back-trajectory analyses, while robust with reanalysis data, carry uncertainties in PBL height, mixing depth, and representativeness; correlations with PBL mixing frequency are suggestive rather than strictly causal.