Weather-Proofing the Data: Clean-Air Laws Really Are Cutting Toxic Metals in Europe’s Air

For decades, governments across Europe have poured money and legislation into cutting toxic heavy metals out of the air we breathe. But a stubborn question has haunted air quality scientists the whole time: how much of the improvement we measure is real progress, and how much is simply the weather being kind to us? A new twelve-year study from the Czech Republic offers one of the clearest answers yet, and it comes with a methodological twist that could change how pollution trends are read worldwide. By applying a technique called dispersion normalisation to measurements of arsenic, cadmium, lead and nickel collected at sixteen monitoring stations between 2010 and 2021, researchers at the Czech Hydrometeorological Institute have shown that the continent’s clean-air laws are working — and that without correcting for meteorology, we could easily be fooled into thinking otherwise.

The stakes are far higher than an academic curiosity. Heavy metals such as arsenic, cadmium, lead and nickel do not degrade in the environment. Once released into the atmosphere, usually attached to inhalable particles, they persist, bioaccumulate in the human body, and settle into soils and waters, spreading contamination far beyond their sources. This persistence is why they are governed by a dense web of international and European regulation, including the Convention on Long-range Transboundary Air Pollution’s Protocol on Heavy Metals, Directive 2004/107/EC and Directive 2008/50/EC, with Czech national limits set at 6 nanograms per cubic metre for arsenic, 5 for cadmium, 500 for lead and 20 for nickel. Verifying that these rules actually deliver cleaner air requires trend analyses that can be trusted — and that is precisely where weather has been muddying the picture.

The core problem is that the concentration of a pollutant measured at a monitoring station reflects two things at once: how much was emitted, and how well the atmosphere diluted it. A shallow mixing layer traps pollutants near the ground; a stiff wind sweeps them away. The Czech team tackled this with the ventilation coefficient, a simple but powerful parameter defined as the product of the atmospheric mixing layer height and the wind speed within it. Dispersion normalisation then adjusts each measured concentration by the ratio of the ventilation coefficient at the time of measurement to the average over the whole study period. In effect, the method asks a counterfactual question: what would the concentration have been if the atmosphere’s dilution capacity had been average all along? The inputs needed — mixing height and wind speed — are readily available from weather models such as ERA5, making the approach accessible far beyond the Czech Republic.

The study’s scope lends it unusual weight. The researchers drew on the national monitoring network, selecting sixteen manual stations that met strict data-completeness criteria under European directives — at least 90 percent valid measurements each year — and spanned six environment types: background rural mountain, background rural lowland, background suburban, background urban, industrial and traffic. Samples of PM10 particles were collected continuously for 24 hours every second day and analysed in accredited laboratories using inductively coupled plasma mass spectrometry. Trend calculations relied on the Theil–Sen method paired with the non-parametric Mann–Kendall approach, a robust statistical combination for noisy environmental data. The result is a dataset that captures everything from pristine mountain air to the industrial heartland of Ostrava and the street canyons of busy cities.

The headline finding is emphatic: arsenic, cadmium, lead and nickel all showed statistically significant declining trends at the p < 0.001 level across every station type, with one striking exception. At the industrial station in the Ostrava region, nickel showed no significant trend at all — a slight, non-significant increase of 0.74 percent per year. Annual reduction rates ranged from roughly 3.1 to 6.2 percent per year depending on metal and environment, with lead falling fastest at traffic stations (5.87 percent per year) and cadmium dropping most steeply at suburban sites (6.15 percent per year). These declines are substantially steeper than the roughly 1.8 percent per year reduction in total PM10 seen at European background stations, suggesting that the targeted heavy metal regulations have outperformed general particulate matter controls.

The spatial patterns tell their own story. Mountain background stations recorded the cleanest air, with median arsenic concentrations of just 0.35 nanograms per cubic metre and lead of 2.1, while the industrial station registered medians an order of magnitude higher — 1.72 for arsenic and a striking 14.90 for lead. Correlation analyses revealed that cadmium and lead behave as regionally uniform pollutants, with many station pairs showing strong Spearman correlations above 0.7, pointing to shared, widespread sources across Central Europe. Nickel, by contrast, showed weak correlations everywhere, consistent with its patchy, source-specific behaviour — chiefly oil combustion, metallurgical processes and tyre and brake wear rather than the seasonal heating and industrial combustion that drive the other metals.

Perhaps the most eye-opening result concerns how badly raw measurements can mislead. The divergence between original and dispersion-normalised concentrations reached up to 52 percent across station types and metals, meaning that weather alone could inflate or deflate apparent pollution levels by half. The year 2018 is a case in point: elevated heavy metal concentrations that year coincided with an exceptionally dry and warm spring and summer across Europe, which suppressed wet deposition and stabilised the atmosphere, allowing pollutants to accumulate. Without normalisation, such a meteorologically driven spike could be misread as an emissions failure. The reverse trap appeared during 2020 and 2021, when rising arsenic and cadmium at a traffic station might have been attributed to pandemic-era changes — but the normalised data suggest adverse dispersion conditions, not altered emissions, were the likelier culprit.

The method also sharpened the view of policy successes buried in the raw record. A short-term rise in emission-related concentrations during 2011–2013, visible as a distinct deviation from the long-term trend in the normalised data at traffic and industrial stations, was largely masked by meteorological variability in the original series. Episodic lead peaks at the industrial station in 2012, 2015 and 2018 likewise point to periodic industrial releases persisting beneath the overall decline. Only one station in the entire network exceeded a legal limit: an urban background site where arsenic surpassed the 6 nanogram per cubic metre threshold in 2010, 2012 and 2013, consistent with the well-documented influence of residential coal combustion during heating seasons.

The Czech findings also resonate across borders. Central Europe remains one of the continent’s regions with the highest heavy metal concentrations and deposition rates, and transboundary transport can account for anywhere from 7 to 99 percent of cadmium deposition in EMEP member countries. Neighbouring Germany and Poland remain among the largest heavy metal emitters in the EU, and Poland recorded among the smallest reductions in cadmium and lead emissions between 2005 and 2023. This regional dimension means that even the clean mountain stations, where long-range transport dominates, are partly shaped by decisions made beyond Czech borders — reinforcing the case for the international frameworks whose effectiveness the study confirms.

What elevates this work beyond a national audit is its practicality. The authors emphasise that dispersion normalisation is simple, transparent and reproducible: the ventilation coefficient can be computed from freely available reanalysis data such as ERA5 or from the Hysplit modelling system, so any air quality agency in the world could adopt the approach without specialised tools. As climate change alters dispersion conditions and extreme weather events like the 2018 drought become more common, the risk of meteorologically induced distortions in pollution statistics will only grow. For regulators seeking to verify whether billions spent on emission controls are paying off, and for scientists tracking the slow detoxification of Europe’s atmosphere, this Czech study delivers both reassurance and a warning: the air is genuinely getting cleaner, but only careful, weather-aware statistics can prove it.

Subject of Research: Long-term trends of heavy metal concentrations in ambient air in the Czech Republic evaluated using a dispersion normalisation method

Article Title: Dispersion normalisation method for improved long-term trend evaluation: heavy metals in ambient air in the Czech Republic, Central Europe (2010–2021)

Article References: Holubová Šmejkalová, A., Lhotka, R., Škáchová, H., & Pacner, J. (2026). Dispersion normalisation method for improved long-term trend evaluation: heavy metals in ambient air in the Czech Republic, Central Europe (2010–2021). Aerosol Research, 4(2), 441-456. https://doi.org/10.5194/ar-4-441-2026

Image Credits: AI Generated

DOI: 10.5194/ar-4-441-2026

Keywords: heavy metals, air quality, dispersion normalisation, ventilation coefficient, Czech Republic, arsenic, cadmium, lead, nickel, long-term trends, meteorological variability, emission regulations