Pollen allergies afflict over 35 million Americans and are a substantial public health concern. However, spatial variation in pollen exposure and its health consequences remain poorly quantified and ungrounded by plant biology, limiting our ability to predict or reduce exposures. Ecologically-based models could link specific plants to health outcomes and inform plant management solutions. These models have been considered unfeasible because they require data on plant: 1) location and size; 2) pollen production; 3) flowering phenology; 4) pollen dispersal characteristics; and 5) airborne pollen concentrations. In this talk, I report my findings from field studies, remote sensing analyses, allometric equations, and dispersion models for several taxa of allergenic importance in Detroit, Michigan including common ragweed (Ambrosia artemisiifolia), white mulberry (Morus alba), and oaks (Quercus). The resulting predictions of airborne pollen concentrations will reduce exposure measurement error in epidemiological investigations and provide insight on how to better manage plants in urban environments to reduce allergenic pollen exposures.