A study has found that a comparison of ozone (O₃) pollution is necessary not only in urban areas during the day but also in rural areas at night.
Ulsan National Institute of Science and Technology (UNIST) reported on the 6th that a research team led by Professor Lim Jeong-ho from the Department of Earth Environment and City Engineering analyzed using a self-developed artificial intelligence (AI) model, capturing the pattern of ozone persisting in rural areas overnight.
Ozone is a gas made up of three oxygen atoms and is a secondary pollutant generated by the reaction of sunlight with pollutants in the air. It appears most densely during the afternoon when temperatures are highest. Ozone is smaller than fine particulate matter and cannot be blocked by regular health masks, penetrating deeply into the alveoli and potentially causing inflammatory reactions.
The research team developed an AI-based all-sky model that can estimate surface ozone concentrations across East Asia at high resolution for 24 hours, regardless of cloud cover. While existing models struggle with accurate estimation due to observation gaps when clouds obscure the surface, the AI model developed by the research team can estimate ozone concentrations even in cloudy conditions. Additionally, it has a 2 km resolution that is 40 times denser than existing global atmospheric quality reanalysis data (CAMS), allowing it to capture localized high concentrations of ozone occurring in narrow areas.
The research team noted that they developed this model by combining various meteorological data, such as brightness temperature and temperature from the 'Himawari-8' satellite, wind speed, and solar radiation, applying interpretable AI techniques to analyze the information that the AI used for its predictions. Brightness temperature is the value converted from the infrared energy detected by the satellite from the surface or atmosphere into temperature, which is influenced by various environmental conditions, including actual temperature, sunlight intensity, and the thermal state of the atmosphere. AI can indirectly determine the potential for ozone generation through this brightness temperature.
Analysis of the East Asia region using this model found that ozone concentrations were high in urban areas during the day. In some rural areas near cities, ozone was observed to not decrease rapidly even after sunset, but rather tended to remain at high concentrations for extended periods.
Professor Lim explained that, since most ground observation stations are concentrated in urban areas, the model accurately reflects the regional and temporal characteristics of ozone that would have been missed, which can serve as precise data for establishing environmental policies such as seasonal ozone management measures in the future.
This study was conducted with support from the Ministry of Environment, Ministry of Oceans and Fisheries, and Ministry of Education. The research results were published in the Journal of Hazardous Materials on the 5th.
References
Journal of Hazardous Materials (2025) DOI: https://doi.org/10.1016/j.jhazmat.2025.137369