Environmental exposure assessment has largely relied on static, residence-based proxies that assume people remain continuously exposed to home surroundings. The neighborhood effect averaging problem (NEAP) showed that mobility-based estimates tend to converge toward population means, exposing bias in residence-based measures. Yet mobility-based studies often rely on observation windows of days or weeks and implicitly treat them as representative of long-term conditions. Environmental exposures vary across weekdays, weekends, seasons, weather and life-course transitions; short windows capture only partial temporal heterogeneity. Given these challenges, in-depth research is needed on temporal representativeness in environmental exposure assessment.
The study, published (DOI: 10.1016/j.ese.2026.100782) in 2026 in Environmental Science and Ecotechnology, comes from researchers at Hong Kong University of Science and Technology (Guangzhou), The Chinese University of Hong Kong, Tongji University and Peking University Shenzhen Graduate School. The team formalized the temporal exposure averaging problem (TEAP) and tested it using year-long anonymized mobile signaling records from 110,000 Shenzhen residents. They linked daily mobility to greenspace and air pollution exposure and varied observation-window length to quantify temporal sampling bias. The work provides the first empirical demonstration of TEAP.
The team analyzed anonymized mobile signaling records from March 2023 to February 2024 for 110,000 Shenzhen residents, mapping locations to 250 m or 500 m grid cells. Environmental layers included the monthly normalized difference vegetation index (NDVI) for greenspace and fine particulate matter (PM₂.₅), sulfur dioxide (SO₂), carbon monoxide (CO) and nitrogen dioxide (NO₂). They compared monthly mobility-based exposure (MBE) and year-long MBE, and further randomly sampled one to 12 months, repeating each window 100 times, to check the effect of the length of observation windows on the TEAP. Monthly coefficient of variations (CV) consistently exceeded annual CVs, and individuals with low monthly exposure tended to have higher annual exposure, and vice versa—a bidirectional compression toward the mean. The TEAP held across age and gender groups, though it was weakest among middle-aged adults and stronger among younger and older groups and among females. Both intraindividual deviation, measured by root mean squared deviation (RMSD), and interindividual variability, measured by standard deviation (STD) and CV, declined as window length increased, converging toward full-year values. Seasonally stratified four-month sampling produced results nearly identical to unrestricted four-month sampling, suggesting that extending observation duration matters more than selecting nominally representative months.
The authors said the TEAP is not random measurement noise but a structural bias that reshapes how exposure inequality appears. They said short windows can make environmental disparities look sharper than they are, while longer windows reveal convergence toward population averages. They said the findings argue for matching temporal coverage to the question: long-term averages for chronic burden and persistent inequality, and high-frequency data for acute peaks and sensitive periods. They said planners should not assume that a few weeks of tracking captures a person’s annual exposure.
The study has direct implications for environmental justice assessment and urban planning. Exposure disparities estimated from short or seasonally constrained data may misidentify disadvantaged groups or places; long-term, mobility-informed estimates offer a more reliable basis for identifying persistent inequality. At the same time, planners should not rely only on long-term averages: rush-hour pollution peaks, heat waves and other acute episodes require high-frequency monitoring and targeted interventions. The authors propose a tolerance-based approach for determining minimum observation windows, and they call for model-based temporal extrapolation, adaptive sampling and bias-correction methods when full-year data are unavailable. Integrating longitudinal mobility data into environmental monitoring can help distinguish persistent disparities from short-term fluctuations.
###
References
DOI
10.1016/j.ese.2026.100782
Original Source URL
https://doi.org/10.1016/j.ese.2026.100782
Funding information
National Natural Science Foundation of China (Grant No. 42430106);
Hong Kong Research Grants Council (RGC): General Research Fund (Grant Nos. 14605920, 14606922, 14603724 and 14604326); Collaborative Research Fund (Grant No. C4023-20 GF);
Research Matching Grants (RMG 8601219, 8601242 and 3110151);
Research Committee on Research Sustainability of Major Research Grants Council Funding Scheme (Grant No. 3133235) of The Chinese University of Hong Kong (CUHK);
1+1+1 CUHK-CUHK(SZ)-GDSTC Joint Collaboration Fund (Grant Nos. 4760974 and 2025A0505000062);
Innovation and Technology Fund (ITF) (Grant No. ITS/018/24SC);
Vice-Chancellor's One-off Discretionary Fund (Smart and Sustainable Cities: City of Commons) (Grant No. 4930787) of CUHK;
RGC Postdoctoral Fellowship Scheme (Award No. PDFS2324-4H04) for Jianying Wang;
About Environmental Science and Ecotechnology
Environmental Science and Ecotechnology (ISSN 2666-4984) is an international, peer-reviewed, and open-access journal published by Elsevier. The journal publishes significant views and research across the full spectrum of ecology and environmental sciences, such as climate change, sustainability, biodiversity conservation, environment & health, green catalysis/processing for pollution control, and AI-driven environmental engineering. The latest impact factor of ESE is 14.3, according to the Journal Citation ReportsTM 2024.