COS 75-2 - Modeling hourly pollen release for Juniperus ashei with observational and environmental data

Background/Question/Methods

Pollen allergies afflict millions of North Americans, triggering both allergic rhinitis (hay fever) and asthma attacks. Despite the considerable public health consequences of allergenic pollen exposure, much remains unknown about the underlying ecological processes and how they vary across space and time. In particular, more information is needed about the timing of pollen release at daily and hourly scales and how this is affected by environmental conditions for anemophilous plants.

In this field study, we quantified the timing of pollen release for Juniperus ashei, one of the most important sources of allergenic pollen in Texas. To do so, we observed trees and quantified the proportion of open cones, the proportion of open pollen sacs, and pollen release. To better understand the temporal distribution of pollen release, we designed and deployed novel sampling devices (‘pollen platters’) to monitor pollen deposition at 4-hour intervals across the entire season. The timing of pollen release was modeled as a function of environmental data from dataloggers that we deployed at several study sites (temperature, humidity, and wind speed) as well as from publicly available gridded environmental datasets (SMAP, Daymet, and gridMET).

Results/Conclusions

We observed 480 trees in the first year and 691 trees in the second year; some trees were observed repeatedly for a total of 653 and 1,084 total observations respectively. Individual trees released most of their pollen over approximately a week but due to inter-individual variation, pollen release within a site often extended for up to three weeks. Pollen release was generally highest in the mid-morning and lower in the afternoon and evening. Pollen release intensity was significantly correlated with temperature and inversely associated with precipitation and vapor pressure deficit.

The combination of observational and environmental data allowed us to model pollen release from this species on granular time scales. These models will be used in tandem with seasonal models of pollen cone phenology to predict pollen release on the hourly time scales required for effective atmospheric dispersion modeling. Ultimately, process-based models of pollen release will allow for more accurate predictions of airborne pollen concentrations and exposures, which are necessary to manage this public health issue.