Modeling airborne pollen concentrations at an urban scale with pollen release from individual trees

Airborne pollen can trigger allergic reactions, but exposure is poorly understood because neither regional pollen models nor monitoring networks adequately capture the extensive spatial variation in pollen concentrations observed at urban scales. Here, we test whether pollen emissions from individual source plants can predict spatial variation in airborne pollen at scales of hundreds of meters to kilometers. To do so, we quantified pollen release within a city for oaks (Quercus) by mapping individual trees using remote sensing, calculating each tree’s pollen production with allometric equations, and estimating the timing of flowering with satellite-derived temperature data. We also measured airborne pollen concentrations multiple times a week at 9 sites in the first year and at 15 sites in the second year. Predicted pollen release explained 86% of the spatial variation in measured airborne pollen across the pollen season and 55% of local airborne pollen concentrations on any given day, whereas a traditional monitoring station measurements explained only 34% of spatiotemporal variation. Airborne pollen was best predicted by pollen release within approximately 1–2 km. Our results demonstrate that airborne pollen can be effectively modeled within cities by quantifying pollen release from individual trees. This type of approach could potentially be applied elsewhere, improving predictions of airborne pollen within cities and providing opportunities to avoid allergen exposure, fine-tune medication use, and better inform tree management decisions.

Graphical abstract