How pollen counts are measured, and how accurate they are

Behind every pollen count is a pump, a sticky tape and a person at a microscope. Here is how the number is made, what it means, and where its limits are.

Blog · 10 min read · Published 28 September 2026 · Sources checked 28 September 2026

A pollen count is the number of pollen grains in one cubic metre of air, usually averaged over a day. Most monitoring stations measure it with a volumetric trap of the type designed by J. M. Hirst in 1952. A pump draws air in at 10 litres per minute, pollen sticks to a slowly moving adhesive tape, and a trained analyst later identifies and counts the grains under a microscope at about 400×.

The count covers only a sample of the tape, so it is scaled up and divided by the volume of air sampled to give grains per m³. Because the tape has to be collected, mounted and read, manual counts arrive a day or more after the air was sampled. Newer automatic monitors, such as the Swiss national network, now report hourly and in near real time.

Pollen counts are good at showing the timing and intensity of pollen seasons, but they are estimates for one location, with real statistical error at low values. This page explains measurement, not medical advice.

What a "pollen count" actually is

A pollen count is a concentration: grains of pollen per cubic metre of air (grains/m³). For most published counts it is a daily mean, the average over 24 hours, although some stations also report two-hourly or hourly values. It is usually broken down by type, such as birch, grass or ragweed, or grouped into trees, grasses and weeds.

That definition rules out a few common misunderstandings. A count is not the number of grains that land on a surface, not a measure of how many plants are flowering, and not a forecast. It is a measurement of what passed through one trap, often placed on a rooftop so that it samples air mixed over the surrounding area.

Scientists call the study of airborne pollen aerobiology; within palynology it is known as aeropalynology. The pollen that settles out of the air over a landscape is the same pollen rain that ends up preserved in lake sediments.

The volumetric spore trap

Airborne pollen was first monitored for medical purposes by Charles Blackley in the UK in 1870. The modern standard instrument came later: J. M. Hirst described "an automatic volumetric spore trap" in the Annals of Applied Biology in 1952. Cardiff holds the oldest continuous pollen record, dating back to 1943, and has used a Hirst trap since 1954.

Hirst-type traps are still the backbone of monitoring. A 2018 worldwide inventory found at least 879 active pollen monitoring stations, more than 500 of them in Europe, and more than 600 based on the Hirst principle.

How the trap works

1. Air intake. A calibrated pump draws air through a narrow slit at 10 litres per minute. A wind vane keeps the orifice facing the wind.
2. Impaction. Inside, a drum carries a transparent tape coated with an adhesive. Particles in the air stream hit the tape and stick.
3. Time track. A clock turns the drum at 2 mm per hour, so each position on the tape corresponds to a known time. One day equals 48 mm of tape.
4. Collection. The drum is typically changed every seven days. In the lab, the tape is cut into daily segments and mounted on microscope slides, for example in fuchsin-stained glycerine jelly, which makes the pollen stand out.

From tape to number: counting under the microscope

Analysts do not count every grain on the tape. They read a defined fraction of each daily segment, usually along parallel strips called transects, identify each pollen grain they meet, and scale the count up to the whole day's deposit. The result is then divided by the volume of air sampled that day.

How the fraction is read varies within the rules of the standard. One Italian station using EN 16868, for example, reported reading four transverse transects with the total analysed surface above 10% of the sample. The identification itself relies on the same features used in any palynology lab, which we explain in how analysts identify pollen types.

A worked example (illustrative numbers)

The example shows why low counts are fragile. At 25 grains/m³, the analyst actually saw 36 grains. At 3 grains/m³, they would have seen only four or five, and one grain more or less changes the result noticeably. The multiplication step hides how few grains a low value is based on.

Standards and networks

For decades, stations followed shared practice rather than a formal standard. In Europe that changed with EN 16868:2019, Ambient air – Sampling and analysis of airborne pollen grains and fungal spores for networks related to allergy – Volumetric Hirst method, released in September 2019. It sets out how networks sample, handle, identify and quantify pollen with Hirst-type traps. Minimum requirements for counting had earlier been discussed in the literature, for example by Galán and colleagues in 2014.

Coverage is uneven. The 2018 inventory by Buters and colleagues noted that, unlike air-quality networks for chemical pollutants, most pollen networks are not publicly funded and their data are not freely available. Africa had the weakest coverage of any continent. Our directory of pollen databases and networks lists further resources.

What counts as "high"?

There is no single international scale. The NAB publishes categories per station, and on 28 September 2026 most stations on its site used the following thresholds (grains/m³); a few stations set their own.

The very different thresholds show that a "high" grass count is a far smaller number than a "high" tree count. Other countries and networks use their own scales, so always read the scale used by your local source.

Real-time automatic monitors

Automatic monitors identify and count particles in the air as they pass through the instrument, without a tape or a human counter. Switzerland's MeteoSwiss runs the first national network of this kind. Air is drawn into the instrument, pollen grains are imaged in flight using holography, and an artificial-intelligence algorithm classifies them. According to MeteoSwiss, hourly data are now available in real time, where manual daily averages used to become available only after a week.

Japan was the pioneer in automatic monitoring, with 120 automatic stations run by its Environmental Agency in the 2018 inventory, which is possible partly because the pollen types of interest there are few. The same 2018 inventory counted only about eight automatic stations in Europe.

Automatic monitors have limits too. MeteoSwiss notes that only seven allergenic pollen types are measured at some of its automatic stations, and that automatic and manual data differ enough that current-year data are not fully comparable with the long-term average. Hirst traps remain the reference method that new instruments are compared against.

How accurate are pollen counts?

Pollen counts are reliable for tracking seasons and trends, but a single daily value is an estimate with measurable uncertainty. The main sources of error are:

• Counting error. Only a fraction of the tape is read, so the result is a statistical sample. The fewer grains seen, the larger the relative error. Studies cited by Fragola and colleagues (2022) indicate that the Hirst technique can carry significant uncertainty at counts below 20 grains/m³. Comtois and colleagues addressed pollen count statistics and precision in Aerobiologia in 1999.
• Identification limits. Many types can only be named to family by light microscopy. Grasses, for example, are usually counted as one group.
• Spatial representativeness. A rooftop trap samples the air at one site. Concentrations near a flowering tree, in a park or in another district can differ, which is why researchers study local variation with denser sensor networks.
• Time lag. Manual Hirst data reach users with a delay of one to several days, according to a 2024 study of automatic sensors around Brussels. By the time a daily count is published, the weather may have changed.
• Gaps. Instruments can fail and tapes can be damaged, leaving missing days. Researchers have developed statistical methods to fill such gaps, which adds estimation on top of measurement.

Counts vs forecasts vs "pollen apps"

These three are often confused:

• A count is a measurement of what was in the air at a station during a past period.
• A forecast is a model prediction of what will be in the air, based on weather, plant phenology and, increasingly, recent measurements. MeteoSwiss, for example, feeds real-time pollen data into its weather forecast model.
• A pollen app may show either, or a blend, sometimes for places far from any trap. Check whether the app names its source and whether the value is measured or modelled.

A useful rule: if a number is for today or tomorrow and you live far from a monitoring station, it is almost certainly modelled rather than counted.

What this means for allergy sufferers

Pollen counts are a tool for understanding exposure, not a diagnosis or a treatment plan. Sensitivity differs from person to person, so the same count can mean very different things to different people. For symptoms, treatment and prevention, follow your doctor's advice and official health sources such as the NHS guidance on hay fever.

What the measurement side can tell you is how to read the numbers: check which pollen types are being counted, whether the value is measured or forecast, how old it is, and which scale defines "high".

Frequently asked questions

How is the pollen count measured?

Most stations use a Hirst-type volumetric trap. A pump draws in air at 10 litres per minute, pollen sticks to an adhesive tape on a drum turning 2 mm per hour, and an analyst identifies and counts grains on part of the tape under a microscope. The count is scaled up and divided by the air volume to give grains per cubic metre.

Why is today's pollen count from yesterday?

Manual counts need the tape to be collected, mounted and read under a microscope, so results reach users a day or more after the air was sampled; in Switzerland, before automation, daily averages became available only after a week. Automatic monitors, such as those in the Swiss national network, now report hourly data in near real time.

Who counts pollen?

Trained aerobiologists and technicians at universities, hospitals, meteorological services and allergy organisations. In the US, stations report to the AAAAI National Allergy Bureau; in Europe, many stations share data through the European Aeroallergen Network. Many networks are not publicly funded.

What is a high pollen count?

It depends on the pollen type and the scale used. On the most common scale shown by the AAAAI National Allergy Bureau, "high" starts at 90 grains/m³ for trees, 20 for grasses and 50 for weeds. Other countries use different scales, so check the one your local source uses.

Are pollen counts accurate?

They are reliable for showing when seasons start and how strong they are at a station, but each daily value is an estimate. Only part of the tape is counted, low values carry large relative error, and a single rooftop trap cannot represent every street. Forecasts add model uncertainty on top.

Sources

• Hirst J.M. (1952). An automatic volumetric spore trap. Annals of Applied Biology 39: 257–265.
• Buters J.T.M. et al. (2018). Pollen and spore monitoring in the world. Clinical and Translational Allergy 8 (open access).
• CEN, EN 16868:2019. Ambient air – Sampling and analysis of airborne pollen grains and fungal spores for networks related to allergy – Volumetric Hirst method (DIN EN edition, released 2019-09).
• Galán C. et al. (2014). Pollen monitoring: minimum requirements and reproducibility of analysis. Aerobiologia 30: 385–395.
• Comtois P., Alcázar P., Néron D. (1999). Pollen counts statistics and its relevance to precision. Aerobiologia 15: 19–28.
• Fragola M. et al. (2022). Pollen monitoring by optical microscopy and DNA metabarcoding: comparative study and new insights. International Journal of Environmental Research and Public Health 19: 2624.
• Tagliaferro S. et al. (2024). A new method based on physical patterns to impute aerobiological datasets. PLOS ONE 19: e0314005.
• Renard J. et al. (2024). Spatial variation of airborne pollen concentrations locally around Brussels city, Belgium, during a field campaign in 2022–2023, using the automatic sensor Beenose. Sensors 24: 3731.
• Erb S. et al. (2024). Automatic real-time monitoring of fungal spores: the case of Alternaria spp. Aerobiologia 40: 123–127.
• MeteoSwiss, Automatic pollen monitoring network SwissPollen.
• AAAAI National Allergy Bureau, pollen and mold counts (station scales checked 28 September 2026).
• European Aeroallergen Network, EAN database.

This page explains how pollen is measured. It is not medical advice; for symptoms or treatment, consult a health professional.