Ensemble-trained PM2.5 source apportionment approach for health studies

Area of Science:

• Environmental Science
• Atmospheric Chemistry
• Public Health

Background:

• Particulate matter (PM) source apportionment is crucial for understanding air quality and its health impacts.
• Traditional receptor models can exhibit significant day-to-day variability and biases in source impact estimations.
• Integrating chemical transport models (CTMs) with receptor models offers potential for improved source apportionment accuracy.

Purpose of the Study:

• To develop and evaluate an ensemble-trained Chemical Mass Balance (CMB) approach for PM source apportionment.
• To assess the performance of the ensemble CMB method in reducing variability and biases compared to traditional methods.
• To generate new, seasonally-specific source profiles for improved CMB applications in health studies.

Main Methods:

• An ensemble approach combining a short-term emission-based CTM with multiple receptor-based models was developed.
• New source profiles were derived using ensemble results and observational data for summer and winter periods.
• The ensemble-trained CMB model was applied to a 12-month dataset of daily PM2.5 measurements in Atlanta, GA.

Main Results:

• Ensemble CMB results demonstrated reduced day-to-day variation in source impacts compared to original receptor models.
• The ensemble approach showed fewer biases between observed and estimated PM2.5 mass.
• Increases in road dust, biomass burning, and coal impacts were observed, while secondary organic carbon (SOC) impacts decreased.

Conclusions:

• Ensemble-trained CMB approaches effectively reduce day-to-day variability in source impact estimates.
• The method improves upon traditional receptor modeling by reducing instances of zero impact from known sources.
• This enhanced CMB approach provides more reliable source apportionment data for air quality and health studies.