You are here:
ADVANCEMENTS IN SOURCE-TO-DOSE ANALYSIS OF POPULATION EXPOSURES TO OZONE
Citation:
Vyas, V. M., S. W. Wang, Q. Sun, A. Chandrasekar, P. Shade, P. G. Georgopoulos, J M. Burke, AND A H. Ozkaynak. ADVANCEMENTS IN SOURCE-TO-DOSE ANALYSIS OF POPULATION EXPOSURES TO OZONE. Presented at International Society of Exposure Analysis 2002 Conference, Vancouver, Canada, August 11-15, 2002.
Impact/Purpose:
The primary objective of this research project is to develop a scientifically-robust, complete multimedia, multi-pathway human exposure source-to-dose modeling system with modules and computational tools that can estimate exposures and doses to the general population, as well as to identifiable susceptible subpopulations, and can predict and diagnose the complex relationships which exist between source and exposure and dose -- and to ensure that this scientifically sound model, with its associated tools and the various modules included within, meets the needs of Program Offices and the scientific community for conducting risk assessments.
Description:
The current study takes advantage of the observations from regional air quality monitoring networks, the data from the NE-OPS (North East Oxidant and Particulate Study) Project in the Philadelphia region, and regional photochemical air quality model predictions to obtain and compare estimates of population exposure distributions for a two-week period in the summer of 1999. Field data were available from 32 ozone monitoring stations (including the NE-OPS station) within a radius of 100km from the urban Philadelphia NE-OPS (Baxter) site. The availability of both observations and modeling results for O3, permitted the application of alternative spatiotemporal interpolation approaches in this study. Regional air quality predictions from runs of the USEPA Community Multiscale Air Quality (CMAQ) model, for the region and time period of interest, were used as alternative inputs to the population exposure model. The CMAQ runs were performed for the Eastern USA, for nested grids with higher resolution over the Philadelphia region. This study employed the consistent source-to-dose modeling framework of MENTOR-OPERAS/SHEDS (MENTOR-OPERAS: Modeling Environment for Total Risk studies/Ozone and Particles Exposure and Risk Analysis Systems; SHEDS: Stochastic Human Exposure and Dose Simulation). The region considered in the study encompassed the City of Philadelphia and a surrounding area. The study focused on a two-week episode, from July 11, 1999 to July 24, 1999. In order to link the modeling estimates of outdoor concentrations to population exposure models, the CMAQ results were used to calculate localized ambient ozone values for the 482 census tracts of urban Philadelphia. Two novel methods of spatiotemporal analysis were used and compared for obtaining the interpolated values at census tract level: the Spatiotemporal Random Field (STRF) method, and the Bayesian Maximum Entropy (BME) method. The STRF method uses a best linear unbiased estimate (BLUE) interpolation scheme that combines spatial and temporal information to provide more accurate estimates of interpolated concentrations than those resulting from purely spatial or purely temporal methods, thus reducing uncertainty associated with the estimates. The BME method uses sequential simulation techniques to combine hard and soft information on a random field to obtain the optimum posterior probability distribution (pdf) of the random process (the O3 concentrations) over the spatial domain and temporal period considered. The localized outdoor concentration estimates were used as input to the combined MENTOR-OPERAS/SHEDS model, to obtain estimates of population exposures to ozone. The model combines demographic characteristics of the population with outdoor concentration distributions, outdoor-indoor air exchange rates, time-activity diaries, and indoor concentration distributions. It accounts for measurement uncertainty as well as natural variability in input parameters, and provides as output distributions of exposures and does. The inhalation dosimetry module of this system uses equations that account for physiological and metabolic - as related to human activity - variability. Statistical tests and maps are used to compare and visualize the exposures and doses obtained using the alternative inputs for regional air quality concentrations (model predictions vs. monitor concentrations) and the localized estimates from different interpolation methods (STRF vs. BME). The results demonstrate the feasibility of developing population exposure assessments to ozone using a mechanistically consistent source-to-dose modeling framework.
This work had been funded in part by the US Environmental Protection Agency under Cooperative Agreement # EPAR-827033 to Environmental and Occupation Health Sciences Institute. It has been subjected to Agency review and approved for publication.