Spatiotemporal Analysis of PM2.5 in Chicago Using Data from EPA and Low-Cost Sensor Network

Abstract

PM2.5 has been linked to numerous pollution-mediated adverse health effects and their monitoring is key for taking preventative and mitigative measures. Accurate measurements of PM2.5 concentrations are available from EPA sites, but such data lacks spatial resolution due to a limited number of monitoring locations. In recent years the deployment of low-cost sensor networks has opened up the possibility of acquiring air quality data at a high spatiotemporal resolution. In this study, we aimed to build a model to estimate the PM2.5 concentrations at high resolution in the Chicago area, using PM2.5 measurements from EPA and low-cost PurpleAir (PA) sensors network.Time series clustering analysis was first performed to identify spatial patterns in daily and weekly PM2.5 data collected from 4 EPA and 10 PA sampling sites between January to December 2020. The predictor variables in our model included land-used variables such as road length, annual average traffic, land type, and chemicals releasing industries around the sampling locations, and also included meteorological variables such as temperature, humidity, wind speed, and wind direction. Multivariate linear regression models for daily and weekly PM2.5 concentrations were developed after checking multicollinearity among the predictor variables. By selective incorporation of variables, a reasonable prediction of PM2.5 was obtained with coefficient of determination (R2) of 0.6-0.8. Our findings show promise of using low-cost sensor network data to complement EPA data to build models with higher spatiotemporal accuracy .