High temporal frequency global positioning system (GPS) data are increasingly employed in mobility-based exposure assessments, yet their advantages over lower-frequency data remain uncertain for long-term applications. Understanding how GPS sampling rates affect exposure estimates is crucial for identifying the temporal scale at which human daily mobility affects exposure assessments and for informing practical strategies for GPS data collection. This study evaluated the influence of sampling rates on long-term personal exposure to air pollution, noise, green space, and fast food environments. We collected 2-week GPS data from 345 Dutch adults at 20 s (i.e., benchmark data) and systematically down-sampled the data to 1-25 min (i.e., resampled data). Exposure estimates from benchmark and resampled data were compared using Bland-Altman plots, showing minimal impact of sampling rate on personal exposures (mean biases ≈ 0 and concordance correlation coefficients = 1). We found that over 80% of air pollution, noise, and green space exposures originated from the two activity locations where participants spent most of their time, whereas in-transit periods were the dominant source of fast food exposure. These findings indicate that high-frequency GPS data are not essential for estimating long-term, time-weighted personal exposures. Activity location-based approaches can effectively approximate long-term personal exposures to air pollution, noise, and green space and present a promising pathway to facilitate scalable, person-centered exposure studies across large populations.
Wei et al. (Fri,) studied this question.