Why your allergy app says “low” on a terrible day
You wake up with a blocked nose, you burn through half a pack of tissues before lunch, and the app on your phone says pollen is low today. It is not necessarily lying. It is answering a different question from the one you are asking.
A pollen index is an average of things you do not all react to
Almost every general pollen number you see, in a weather app, on a news site, in a widget, is a blend. It takes the species that are airborne in your region right now, weights them, and reduces them to one figure or one word: low, moderate, high.
That blend is useful if you react to everything. Very few people do. Sensitisation is species-specific: you might react to birch and not to alder, to timothy and ryegrass but barely to mugwort. The blend flattens exactly the distinction that matters to you.
So on a day when birch has finished, grasses have not started, and the only thing in the air is the one weed you happen to be sensitised to, the blended index is genuinely low. It is low on average. The average is not what is landing in your eyes.
The bands are wider than your symptoms
The second problem is resolution. Pollen counts are usually bucketed into three or four bands, and those buckets are wide. For grass pollen, a commonly used scheme in central Europe puts roughly 1 to 10 grains per cubic metre in the lowest band, 11 to 30 in the middle, and anything above that in the top band.
Look at what that means in practice. Someone who is strongly sensitised can start reacting in the single digits. The whole bottom band, the one your app renders as a reassuring green, covers the entire range from "nothing at all" to "a bad morning for a sensitive person". Once you are above 30, the band stops resolving anything: 35 and 350 read the same, and they are not the same day.
Rule of thumb: the more strongly you react, the less information the low band carries for you. If you have had a bad day that the index called low, that is not a bug in your body.
Forecasts are models, not measurements
It is worth being clear about where these numbers come from. A live pollen count means someone ran a volumetric trap, and somebody looked at a slide under a microscope and counted grains. That is accurate, local, and typically a day or two behind.
What you get in an app is nearly always a forecast instead: a model that combines the historical flowering calendar, land cover data on what actually grows around you, and current weather. Models are the only way to get today and tomorrow rather than last Tuesday, and they are good at the shape of a season. They are less good at your street.
They are especially weak on two things:
- Very local sources. A single birch outside your bedroom window produces a personal pollen climate that no regional grid cell knows about.
- Sharp weather turns. Rain scrubs the air quickly; a dry, warm, breezy afternoon after rain can reload it just as quickly. Thunderstorm outflows can also break grains into smaller fragments that travel deeper into the airways, which is the mechanism behind so-called thunderstorm asthma.
Your symptoms lag the air
There is a third gap, and it is the one people most often mistake for the app being wrong. The inflammatory response does not reset overnight. After a heavy exposure day, the airways stay primed, and a moderate day that follows a severe one can feel worse than the severe one did. This is why a chart of your symptoms rarely lines up neatly with a chart of yesterday's counts, and why one bad day is weak evidence about a trigger while a season of logged days is strong evidence.
What to do with all this
None of it means pollen data is useless. It means you should read it with the right expectations:
- Find out what you actually react to. A skin prick or specific IgE test from an allergist turns "pollen" into a short list of names. Everything downstream gets sharper once you have that list.
- Look at the species breakdown, not the headline number. Most data sources publish per-species values. The headline is the part that throws your information away.
- Prefer hourly to daily. A day labelled moderate can hide a three-hour spike. See grass season, hour by hour.
- Keep a diary for one season. Your own record beats any model at predicting you specifically, because it captures the local birch, your commute, and your own thresholds all at once.
How Polly handles it
This is the problem Polly was built around. Instead of showing you a regional average, it takes the per-species forecast for your location and weights it by the allergens in your own profile, so the score you see moves when your pollens move and stays quiet when they do not. The species driving the number are always listed underneath it, because a score you cannot interrogate is just another opaque band.
The diary works the other way round: two taps to log how a day went, and over a season Polly compares your logged days against what was actually in the air to surface triggers you had not connected. That is also how it catches the species you never got tested for.
Polly scores the day against your own allergens
Not a city average. Free on the App Store, no account needed.
Get Polly