Using dispersion and mesoscale meteorological models to forecast pollen concentrations

This work describes the results of research into a source-oriented pollen concentration forecasting technique. Tests were conducted using the National Center for Atmospheric Research/ Penn State Fifth Generation Mesoscale Model (MM5), the National Oceanographic and Atmospheric Administration (NOAA) Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT_4) Model combined with the locations of oak trees and their aerial coverage from biogenic emissions land cover database version 3.1 (BELD3). Daily forecasts of pollen concentrations via MM5 and HYSPLIT_4 were made with 30-min increments and tested against 30-min oak pollen data collected by the St. Louis County Department of Health in Clayton, Missouri, for the month of April 2000.Results from these tests show that the combination of MM5 and HYSPLIT_4 with accurate source locations can provide short-term forecasts as indicated by the levels of forecast pollen and actual oak pollen levels, which follow similar profiles for the day. From the 30 individual pollen concentration forecasts, two example forecasts are presented. Additional studies need to be conducted to further validate these results, using an array of pollen collectors. A better understanding of the biology of pollen release is critical to improving these pollen concentration forecasts.