4PLANET MAGAZINE
The air filter that became a biodiversity time machine
For 34 years, a Swedish monitoring station stored air filters designed to catch radioactive particles. Scientists have now found that those same filters were preserving traces of thousands of organisms — turning old infrastructure into an accidental archive of ecological change.
For decades, a monitoring station in northern Sweden pulled enormous volumes of air through glass-fibre filters and stored them away. The system was built to detect radioactive particles. It was also, without anyone designing it for the job, collecting fragments of forests, fungi, insects, birds and mammals.
A Nature Communications study published in December 2025 turned that accidental archive into a 34-year ecological record. Researchers sequenced environmental DNA trapped on weekly aerosol filters from the station at 67.84°N in the boreal forest. Their analysis reconstructed relative-abundance signals for more than 2,700 genera across the tree of life.
The surprise is not simply that DNA can travel through air. That has been demonstrated before. The deeper shift is temporal: a piece of infrastructure built for one environmental risk had quietly preserved material that could be re-read decades later to answer a different question.
Air as an ecological sampling medium
Environmental DNA has already transformed aquatic monitoring because water gathers biological traces from many organisms into a substrate that can be sampled. Terrestrial ecosystems are harder. Soil, scat, vegetation and insects each capture different fragments of the living system. Air offers another route because the near-surface atmosphere carries particles from microbes, plants, fungi and animals.
The Swedish station made that idea unusually powerful. Air was continuously pumped through 0.2-micrometre filters at roughly 1,000 cubic metres per hour, and the filters were changed weekly. That meant the archive did not contain a few isolated snapshots. It contained a repeated sampling process stretching across decades.
The most radical sensor in this story may be an old sensor whose data we had not yet learned how to read.
What the researchers could see
The team detected organisms across broad branches of life and produced quantitative indices that, for some vertebrates, were congruent with traditional monitoring. Bird eDNA estimates, for example, showed a statistically significant relationship with point-transect counts during overlapping years. The researchers also identified a long-term decline in diversity consistent with changes in regional forest management.
That last sentence needs care. Airborne eDNA is not a magical census. The study is explicit about uncertainty in transport distance, reference databases, biological shedding, weather and sampling. Most sequence reads — 76% — could not be classified, largely because genetic reference libraries still represent only a fraction of life. Relative abundance is not the same thing as counting every animal in the landscape.
The infrastructure hiding in plain sight
The paper points to a bigger opportunity. Radionuclide and air-quality stations already operate under standardised protocols in many countries. Europe alone has hundreds of radionuclide monitoring stations, while global networks built for nuclear-test surveillance are deliberately distributed across the planet. If archived filters from some of those systems preserve usable biological material, the world may already possess fragments of ecological history that have never been analysed as biodiversity records.
That possibility changes how we should think about environmental infrastructure. A sensor network does not have to belong to one problem forever. The physical act of sampling can outlive the question that justified it. As sequencing methods improve and reference databases grow, old material can become newly informative.
The useful question is not ‘can AI or DNA replace field ecology?’
It cannot, and that framing misses the point. The interesting future is layered monitoring: field surveys, acoustic recorders, camera traps, satellites, environmental DNA and existing infrastructure each contributing different kinds of evidence. The result is not one perfect eye on nature. It is a denser, more resilient observation system — one in which an air filter stored for the wrong reason can still become a window into a forest that has already changed.
Sources
- Airborne eDNA captures three decades of ecosystem biodiversity
- Advancing airborne eDNA sampling methods for monitoring diverse terrestrial vertebrate communities