The temporal exposure averaging problem (TEAP)

Accurately capturing environmental exposure is essential for environmental health research and equitable urban planning. Many studies estimate long-term greenspace and air pollution exposure from short GPS or mobile signaling tracks lasting only days or weeks, treating these brief periods as representative of chronic conditions. Here we show that this assumption systematically exaggerates differences between people. We conceptualize this temporal sampling bias as the temporal exposure averaging problem (TEAP). Using year-long mobile signaling records from 110,000 individuals in Shenzhen, we reconstructed daily mobility and calculated exposure to greenspace and multiple air pollutants. When we randomly sampled increasing numbers of months, short windows produced large intraindividual deviations from each person's annual exposure and inflated interindividual variability across the population. Longer windows caused both measures to converge steadily toward the full-year values and the population means. This temporal averaging pattern held across age and gender groups and environmental indicators. Seasonally stratified and unrestricted random four-month sampling yielded similar estimates. The results show that observation length shapes whether environmental exposure appears highly unequal or relatively averaged, even when daily mobility is accounted for. Accurately assessing chronic exposure burdens and persistent inequality therefore requires observation periods long enough to capture temporal heterogeneity, rather than brief samples that amplify apparent differences.