EXPOLIS simulation model: PM2.5 application and comparison with measurements in Helsinki.

Area of Science:

• Environmental Health
• Exposure Science
• Computational Toxicology

Background:

• Particulate matter (PM2.5) exposure is a significant environmental health concern.
• Accurate estimation of population exposure distributions is crucial for epidemiological studies and risk assessment.

Purpose of the Study:

• To develop and validate a probabilistic simulation framework for predicting population-level PM2.5 exposure distributions.
• To compare simulation results with measured personal exposure data from the EXPOLIS study in Helsinki.

Main Methods:

• Utilized a probabilistic simulation framework to model PM2.5 exposure distributions for adult Helsinki citizens.
• Developed four microenvironment models, progressing from simpler to more complex versions incorporating factors like environmental tobacco smoke (ETS) and subpopulation-specific time-activity patterns.
• Accounted for correlations between input concentration and time fraction variables in advanced models.

Main Results:

• Simpler models (1 and 2) provided a reasonable outline of exposure distributions but underestimated the mean by up to 20% and the standard deviation by 23-35%.
• Improved models (3 and 4), which excluded ETS-exposed subjects and separated working/nonworking subpopulations, yielded results very close to observed distributions.
• Differences in means were less than 0.1 µg/m³, and standard deviation differences were less than 1% with the refined models.

Conclusions:

• Microenvironment modeling, when supported by reliable input data, can accurately predict population PM2.5 exposure distributions for practical applications.
• Advanced simulation models offer a significant improvement in accuracy for estimating personal PM2.5 exposure compared to simpler approaches.