Confronting the controversy over neighborhood effect bias in green exposure: Using large-scale multi-temporal mobile signal data

Exposure to green spaces is known to enhance residents' physical and mental well-being, making accurate assessment of individual green exposure crucial. Traditional research often relies on fixed residential-based assessments, neglecting individual daily mobility, which can lead to estimation biases known as neighborhood effect biases, including the neighborhood effect averaging problem (NEAP) and neighborhood effect polarization problem (NEPP), due to varying sampling periods, seasonal changes, and sample selection biases. This study innovatively examines the spatiotemporal dynamics of residents' green exposure and neighborhood effect heterogeneity using large-scale (330,160 residents), multi-temporal (across four seasons in one year) mobile signal data (over 1.38 billion signal points). Overall, NEAP is dominant among the population. We found that "time restrictions" are key to neighborhood effect biases: on weekends or during spring and autumn (pleasant weather), NEAP is more likely to exhibit due to flexible travel, compensating for less greenery at home by visiting greener areas. Conversely, the probability of NEPP increases on weekdays due to strict commuting schedules or during summer and winter due to extreme weather conditions. Furthermore, socioeconomic factors such as income and gender differentially modulate access to green spaces, demonstrating complex spatiotemporal heterogeneity. These insights address the controversy over neighborhood effects of green exposure in previous studies and provide a new perspective for accurate environmental exposure assessments and their health outcomes.