How pollen forecasts are made

A pollen forecast is a prediction, not a reading from an instrument.¹ AAAAI explains that counts come from an instrument that samples the air for 24 hours, while forecasts are predicted from the previous year’s counts and current weather conditions.¹ In AAFA’s words, a forecast is an estimate, “just like a weather forecast.”² What goes into one, and why it can miss, is set out below.

What goes into a forecast

The plants come first. Scientists developing an experimental pollen forecast at NOAA’s Global Systems Laboratory and a partner institute first give their model a dataset, from the University of Michigan’s Steiner Lab, of pollen emission potentials based on the types of plants growing across the United States and past seasons’ climatology.³

The weather comes next. As the experimental model produces its weather forecast, it modifies those emission potentials based on wind speed, humidity and precipitation.³ The system was experimental when NOAA’s Climate Program Office described it in 2024, and the office noted that no federal agency had an official, dedicated pollen forecasting system.³

Forecasts also tend to be selective: AAFA says that most of the time, pollen forecasts report only on the most common types of pollen.²

When a plant’s season starts

Accumulated warmth is one way to see a season coming. NC State Extension describes predicting the start of the pine pollen season by adding up degree-days above 55 °F from February 1, where a day’s degree-days are its high temperature minus 55, counted only when the high is above 55 °F.⁴ Pollen release typically begins at about 300 degree-days, and peak production comes at about 636.⁴

The weather can bend that calendar. NC State Extension notes that weather conditions have a major impact on the season and that a rainy spring or late spring frost can often reduce the amount of pollen in the air.⁴ A freeze can also slow a tree’s rate of producing pollen, according to University of Arizona Cooperative Extension.⁵ Cedar fever explained follows one season in detail.

How Google describes its model

Google, whose Pollen API supplies the levels on AllergyNest’s metro pages, said in its 2023 blog post announcing the API that the model behind it uses inputs such as land cover, climatological data and annual pollen production per plant to calculate the seasonality and daily amount of pollen grains.⁸ Its developer documentation says the Universal Pollen Index is based on a multilayer model that considers a concentration model for each plant, many local pollen indexes used worldwide and the allergenicity of different plant pollen types, to predict the number of pollen grains per cubic meter of air per day.⁶

Google says its API continuously calculates the value of various pollen types and plant indexes at a resolution of 1 x 1 kilometer (0.6 x 0.6 mile).⁷ It says the forecast runs up to five days, with daily information, and that the number of forecast days may vary between locations and plant species.⁷ Google also says that because a plant’s availability depends on actual data source measurements and reports, its plant data may vary across locations.⁶

Pollen index vs. pollen count explains the 0 to 5 scale, and how our pollen index works covers how AllergyNest shows it.

Why rain and wind change the day

The amount of pollen in the air tends to be highest early in the morning on warm, dry, breezy days and lowest during chilly, wet periods, according to NC State Extension.⁴ University of Arizona Cooperative Extension puts it simply: rain dampens pollen and reduces its ability to flow through the air, while windy and warm weather can increase pollen amounts.⁵

Rain’s effect can be short-lived. An AAAAI work group says light rainfall briefly washes pollen and fungal spores from the air but may release very small pollen particles.⁹ Wind, meanwhile, moves pollen far from its source: AAFA notes that ragweed pollen can travel hundreds of miles in the air.²

Thunderstorms can bring a different problem. A research team that included CDC scientists describes a proposed mechanism for thunderstorm asthma: early in a thunderstorm’s development, rising air can carry whole pollen grains from the ground to the cloud base; the grains then frequently rupture on contact with moisture in the cloud, releasing breathable pollen fragments in much higher concentration than the grains themselves; and precipitation and descending cool air could carry those particles back down to the ground.¹⁰ In the team’s study of asthma emergency department visits in Louisiana from 2010 through 2012, higher precipitation and lower temperature on thunderstorm days appeared to contribute to asthma attacks among people with asthma.¹⁰ Allergic asthma explains how allergies and asthma connect.

Where forecasts fall short

Weather is one of the inputs: in NOAA’s experimental model, the pollen estimates are adjusted for the wind speed, humidity and precipitation in the model’s own weather forecast.³ Weather forecasts have limits of their own. NOAA explains that the atmosphere is changing all the time, so weather forecasts are less reliable the further into the future they reach.¹¹

Checking a pollen forecast is hard too: NOAA names access to real-time pollen data for forecast verification, along with the heavy computing needs of the models, as fundamental challenges for its experimental forecast.³ The National Atmospheric Deposition Program’s aeroallergen committee says the high cost of identifying and counting pollen by hand substantially limits the spatial coverage and timeliness of forecasts, and that a central pollen data bank would aid in improving models for forecasting pollen where there are no monitoring stations.¹²

A forecast also describes the air, not a person: AAFA notes that people may feel allergy symptoms at times when the forecast predicted low or moderate amounts of pollen.²

Longer seasons and more pollen

The seasons themselves may be changing. CDC says climate change may increase pollen concentrations and extend pollen seasons.¹³ It also says climate change may cause shifts in precipitation, fewer frost days, warmer air temperatures and more carbon dioxide in the air, and that these changes can affect when the pollen season starts and ends, how long it lasts, how much pollen plants make and how much is in the air.¹³

An AAAAI work group notes that numerous reports have indicated pollen seasons are lengthening, driven by longer growing seasons, earlier last spring frosts and later first fall frosts.⁹ In the United States, it says, this has been most notable in the Northeast and West, while other studies have shown longer seasons at higher latitudes.⁹ Several North American studies of temperature and long-term pollen measurements have generally found earlier starts and longer seasons, according to the report.⁹ The same report says the balance of studies shows more total pollen collected in warmer years for trees and ragweed but not for grasses, and that, on the basis of available studies, more carbon dioxide will generally increase plants’ reproductive output, including pollen production.⁹ It adds that modeling studies indicate the effects of climate change on pollen seasons will vary substantially across regions of the United States.⁹

For questions about allergy symptoms in any season, talk to an allergist.