Research

Current projects include creating models of airborne pollen at municipal scales in Detroit for wind-pollinated trees and ragweed, at regional scales in Texas for Ashe juniper pollen (the cause of 'cedar fever') , and at continental scales in the United States and Canada for wind-pollinated trees. We trace the contributions of individual plants to airborne pollen concentrations and exposures, allowing for:

more informed plant management decisions in cities;
improved pollen forecasts so people can reduce their exposures and better manage their allergy medications; and
better epidemiological analyses through improved estimates of allergen exposures.

urban-scale pollen models

In this project, we leveraged methods from plant ecology and remote sensing to understand and predict airborne pollen concentrations on scales of hundreds of meters to kilometers. Specific projects include:

Ragweed habitat modeling within Detroit
Identification of trees within Detroit using remote sensing data
Development of allometric equations for pollen production by several tree species
Quantification of flowering phenology using field measurements and remote sensing
Parameterization of pollen atmospheric dispersion models
Creation of portable and programmable pollen samplers
Empirical assessment of spatio-temporal variation in airborne pollen concentrations
Synthesis of these datasets into a process-based model of airborne pollen concentrations and model assessment

This project was supported by an F32 postdoctoral fellowship from the National Institute of Environmental Health Sciences and by the Michigan Institute of Clinical and Health Research and was advised by Drs. Stuart Batterman at the (University of Michigan School of Public Health) and Alan Baptist (University of Michigan Department of Internal Medicine).

rEGIONAL-scale pollen models

Ashe juniper (Juniperus ashei) is one of the most abundant trees in central Texas and it is releases the infamous allergenic pollen that causes 'cedar fever.' The goal of this research is to create process-based models of airborne pollen concentrations for J. ashei. This includes:

manual measurements of pollen cone phenology
design of an automated pollen release monitoring device (the 'Pollen Platter')
creation of the citizen science project Pollen Trackers in collaboration with the National Phenology Network
combining pollen cone measurements with gridded environmental data to model pollen cone phenology
modeling atmospheric dispersion of pollen
assessment of model results with National Allergy Bureau monitoring stations
quantifying associations between airborne pollen concentrations and asthma-related emergency department visits.

This project is advised by Drs. Elizabeth Matsui at the Dell Medical School and Shalene Jha at the University of Texas at Austin.

Continental-scale pollen models

We are starting to build continental scale models of airborne pollen based on maps of plant locations and abundance. This project will leverage National Phenology Network observations, plant distribution and abundance models, local-scale temperature metrics and phenology modeling, laboratory experiments to better understand the conditions of pollen release, and atmospheric dispersion modeling. We're actively searching for collaborators on this research.

Other projects

Allergenic pollen is just one of many ways that people interact with plants. We readily and enthusiastically apply our knowledge and skillsets to other topics, including ecosystem services. For example, plant location, structure, and identity can be used to create better estimates of shading from trees, an important ecosystem service in Austin! This physically accurate animation of shade on July 22, 2020 in Hyde Park is based on LiDAR, aerial imagery, and is coded in R using lidR and rayshader.

What our work looks like

The plants

We work with temperate wind-pollinated plants, from oak trees to junipers and ragweed.

the Process

Our lab uses methods from plant ecology, remote sensing, Bayesian statistics, citizen science, atmospheric dispersion modeling, engineering, and epidemiology. We work at scales from microns to thousands of kilometers.

Credit: Gregory Fox Photography

the models

We use remote sensing imagery collected from airplanes, satellites, and drones to map plants, their characteristics, and their habitats. We also use environmental datasets (self-collected and publicly available) to understand and predict plant phenology. This knowledge is combined to estimate pollen release and inform atmospheric dispersion modeling efforts, which we then assess empirically.

HOME PEOPLE RESEARCH PUBLICATIONS CONTACT MEDIA