An AI-based model to reduce disparities of NO2 exposure for communities: a case study in Metro New York

Ground-level NO2 has both adverse short and long-term effects on human health, while long-term exposure to NO2 is associated with morbidity and mortality (WHO, 2006, 2017). The sources of NOs are from human activities, including primarily combustions of fossil fuel, and other natural sources. Metro New York is suffering from NO2 pollution due to complicated emission sources and extremely high levels of human activity. This high level of ground-level NO2 is a significant threat to millions of people in Metro New York.

Machine learning and AI techniques have been applied to generate NO2 maps with high resolutions. However, communities and local governments need a tool that helps them manage NO2 pollution and reduce NO2-related environmental health and justice issues. The existing NO2 maps can show some helpful information but cannot project the consequences of communities’ actions on NO2 reduction and eventually assist communities with their air pollution management plan.

In this study, we focus on understanding factors that control the spatial and temporal distribution of ground-level NO2 in New York using machine learning and AI techniques. We have summarized the significant factors controlling ground-level distribution and concentrations of ground-level NO2 in Metro New York. We develop a model that can provide high-resolution NO2 maps, project consequences of community actions on NO2 management, and assist communities in improving their NO2 improvement plans.