Errors in coarse particulate matter mass concentrations and spatiotemporal characteristics when using subtraction...

Area of Science:

• Environmental Science
• Atmospheric Chemistry
• Air Quality Monitoring

Background:

• Particulate matter (PM) mass concentrations are crucial for air quality assessment.
• Estimating coarse PM (PM10-2.5) often involves subtracting PM2.5 from PM10 measurements.
• The Filter Dynamic Measurement System (FDMS) was developed to correct for semi-volatile losses in heated Tapered Element Oscillating Microbalance (TEOM) filters.

Purpose of the Study:

• To evaluate the accuracy of PM10-2.5 mass concentration estimations using various combinations of TEOM instruments with and without FDMS.
• To assess the biases introduced by different measurement configurations on absolute mass, temporal variability, spatial correlation, and homogeneity.
• To develop a method for correcting biased measurements from FDMS-equipped instruments.

Main Methods:

• Utilized data from three monitoring sites in the Colorado Coarse Rural-Urban Sources and Health (CCRUSH) study.
• Simulated four PM10-2.5 estimation methods using collocated PM10 and PM2.5 TEOMs, with and without FDMS.
• Assessed biases in absolute mass, temporal variability, spatial correlation, and homogeneity.
• Developed a regression-based model to estimate and remove semi-volatile bias from PM2.5 measurements.

Main Results:

• TEOM units without FDMS provided accurate PM10-2.5 mass and spatial characteristic estimations in Colorado due to low semi-volatile PM10-2.5.
• Estimations using PM2.5 or PM10 monitors without FDMS introduced absolute biases ranging from -24% to 25%.
• FDMS-equipped instruments, when not corrected for semi-volatile loss, introduced significant biases in measurements.
• The developed regression model effectively corrected semi-volatile bias in PM2.5 measurements.

Conclusions:

• In regions with low semi-volatile coarse PM, TEOM instruments without FDMS can accurately estimate PM10-2.5.
• The use of FDMS without proper correction can lead to significant biases in PM10-2.5 estimations.
• A regression-based approach can successfully correct for semi-volatile mass loss, improving the accuracy of PM measurements for regulatory and epidemiological studies.