Mobility-driven estimate reveals elevated air pollution exposure and socioeconomic disparities beyond residence-based
Nail F Bashan1, Yang Zhang1, Michelle L Bell2
1Department of Civil and Environmental Engineering, Northeastern University, Boston, 02115, MA, USA.
Background:
Residence-based air pollution exposure assessments ignore daily human mobility and may misrepresent exposure levels and disparities across population groups.
Objective:
We hypothesize that incorporating high-resolution mobility trajectories into exposure modeling will reveal higher average PM2.5 exposures and uncover sociodemographic disparities that traditional residence-based methods underestimate or conceal.
Methods:
We analyzed 155,000 trip records from 990 Boston-area participants (June-December 2023) collected via smartphone GPS, linked to PM2.5 measurements from 294 calibrated PurpleAir air quality sensors collected at 2-min intervals. For each stay location, we computed a daily adjusted exposure as the average PM2.5 within a 4 km buffer minus the region's daily average. We compared these mobility-informed exposures to home-based estimates, assessed temporal (weekday vs. weekend, peak vs. off-peak) and spatial variability (Moran's I), and used weighted least squares regressions and t-tests to evaluate differences across race, income, education, age, and occupation.
Results:
Mobility-informed exposures averaged 0.10 µg/m3 higher than residence-based estimates on weekdays (up to 0.45 µg/m3 on high-pollution days). Employed and higher-income individuals, as well as White participants, experienced significantly elevated exposures during peak travel hours (up to +0.30 µg/m3; p < 0.01). Spatial clustering of mobility exposures was stronger on weekdays (Moran's I = 0.4) than weekends (I = 0.2), and regression coefficients confirmed systematic underestimation by traditional methods.
Significance:
These findings demonstrate that neglecting mobility systematically underestimates exposure levels and obscures environmental injustices.
Impact Statement:
Integrating dynamic mobility data with hyperlocal air quality monitoring provides a refined framework for accurate exposure assessment, informing equitable public health policies and targeted interventions.
Related Concept Videos
Bias in Epidemiological Studies
The Availability Heuristic
Sampling Plans
Random sampling is a method where each member of the population has an equal chance of being selected for the sample. It involves selecting individuals randomly, often using random number generators or lottery-type methods. For example, when analyzing the properties of a...
Strategies for Assessing and Addressing Confounding
Confounding can be addressed at both the design phase of a study and through analytical methods after data...
External and Internal Respiration
Confounding in Epidemiological Studies


