Real-time mobile sensing calibrated by random forest for spatiotemporal assessment of air pollution exposure...

Air pollution has been broadly acknowledged as a significant contributor to various health issues. However, air pollution has conventionally been measured by fixed monitoring stations with limited spatiotemporal resolution. This research leverages real-time mobile sensing to conduct the spatiotemporal assessment of air pollution distribution and inequality in Hong Kong across multiple temporal levels including daytime/nighttime, weekday/weekend, and four seasons. Using the population-weighted exposure method and the Gini coefficient, the study reveals that districts like Sha Tin, Central & Western, Yau Tsim Mong, and Sham Shui Po consistently exhibit elevated pollution levels compared to the Hong Kong average across various timeframes. Furthermore, regarding intra-district pollution inequality, Large Subunit Groups (LSUGs) near the border with Mainland China within the districts of North, Tai Po, Tuen Mun, and Yuen Long in the New Territories, as well as LSUGs within Wan Chai near the district boundary with Central & Western on Hong Kong Island, and LSUGs within Kowloon City near the district boundary with Yau Tsim Mong and Sham Shui Po in Kowloon, exhibit higher levels of air pollution exposure compared to their intra-district LSUG counterparts located further inland. These findings help policymakers formulate targeted interventions to improve air quality across Hong Kong.