Abstract:Many epidemiological studies have evaluated the health risks of ambient fine particulate matter (PM2.5). However, few studies have investigated the potential exposure misclassification caused by using ambient PM2.5 concentrations as proxy for individual exposure to PM2.5 in regions with high-level of air pollution. This study aimed to compare the differences between personal and ambient PM2.5 constituent concentrations, and to predict the personal exposure of sixteen PM2.5 constituents. We collected 141 72-h personal exposure filter samples from a panel of 36 healthy non-smoking college students in Shanghai, China. We then used the liner mixed effects models to predict personal constituent-specific exposure using ambient observations and several possible influencing factors including time-activity patterns, temporal variables, and meteorological conditions. The final model of each component was further evaluated by determination coefficient (R2) and root mean square error (RMSE) from leave-one-out-cross-validation (LOOCV). We observed ambient concentrations were higher than personal concentrations for all PM2.5 components except for Mn, Fe, Ca, and V. Especially, ambient NH4+, As, and NO3– concentrations were 3.65, 5.65 and 7.33-fold higher than their corresponding personal concentrations, respectively. The ambient level was the strongest predictor of their corresponding personal PM2.5 components with the highest marginal R2 (RM2: 0.081 ~ 0.901), meteorological conditions (RM2: 0.000 ~ 0.357), time-activity pattern (RM2: 0.000 ~ 0.083) and temporal indicators (RM2: 0.031 ~ 0.562) were also important predictors. Our final models predicted at least 50% of the variance of all personal PM2.5 constituents and even over 90% for K, Pb, and SO42−. LOOCV analysis showed that R2 and RMSE ranged from 0.251 to 0.907 and 0.000 to 0.092 μg/m3, respectively. Our results showed that ambient concentration of most PM2.5 constituents along with time-activity patterns, temporal variables, and meteorological conditions, could adequately predict personal exposure concentration. Prediction models of individual PM2.5 constituent may help to improve the accuracy of exposure measurement in future epidemiological studies.

A Random Forest Model for Daily PM2.5 Personal Exposure Assessment for a Chinese Cohort

A Random Forest Model for Daily PM 2.5 Personal Exposure Assessment for a Chinese Cohort

Estimating Ground-Level PM 10 in a Chinese City by Combining Satellite Data, Meteorological Information and a Land Use Regression Model

National Scale Spatiotemporal Land-Use Regression Model for PM2.5, PM10 and NO2 Concentration in China

Estimation of Personal PM2.5 and BC Exposure by a Modeling Approach - Results of a Panel Study in Shanghai, China.

A Stochastic Exposure Model Integrating Random Forest and Agent-Based Approaches: Evaluation for PM2.5 in Jiangsu, China.

Assessment for personal PM2.5 exposure with a modeling method: A panel study in Wuhan, China

Necessity of Personal Sampling for Exposure Assessment on Specific Constituents of PM2.5: Results of a Panel Study in Shanghai, China

Predicting monthly high-resolution PM2.5 concentrations with random forest model in the North China Plain

A high-resolution computationally-efficient spatiotemporal model for estimating daily PM2.5 concentrations in Beijing, China

High-Resolution Spatiotemporal Modeling for Ambient PM2.5 Exposure Assessment in China from 2013 to 2019.

Estimation of hourly full-coverage PM2.5 concentrations at 1-km resolution in China using a two-stage random forest model

A similarity distance-based space-time random forest model for estimating PM2.5 concentrations over China

Estimating Annual Residential Exposure Averages by Extrapolating Snapshot PM2.5 Campaign Results and Translating Ambient Measurements

Estimating Elemental Constituents of Personal PM2.5: A Modeling Approach of Older Individuals of the China BAPE Study

Full-coverage 1 km daily ambient PM<sub>2.5</sub> and O<sub>3</sub> concentrations of China in 2005–2017 based on a multi-variable random forest model

High-Resolution Spatiotemporal Modeling for Ambient PM 2.5 Exposure Assessment in China from 2013 to 2019

Calibration of a Low-Cost PM2.5 Monitor Using a Random Forest Model

A spatio-temporally weighted hybrid model to improve estimates of personal PM2.5 exposure: Incorporating big data from multiple data sources

A Machine Learning Method to Estimate PM2.5 Concentrations Across China with Remote Sensing, Meteorological and Land Use Information.

Personal Exposures to PM2.5 in Rural Beijing, China: Spatial-temporal Variation Patterns and Modeling Approaches