Air pollution and planetary health
We aim to disentangle the complex chemical and physical processes during the air pollution development from either anthropogenic sources or wildfires. Numerical simulations as well as machine learning models are employed to characterize PM and trace gases and to assess their environmental impacts.
Machine Learning and Traffic Pollution
Our PNAS paper capitalizes on large variations of urban air quality during the COVID-19 pandemic and real-time observations of traffic, meteorology, and air pollution in Los Angeles to develop a machine-learning air pollution prediction model. We reveal heavy-duty truck emissions contribute primarily to the pollution variations, and demonstrate that the full benefit from fleet electrification cannot be attained if focused only on mitigation of local vehicle emissions.
AI/ML Boost Pollution and EJ Assessment
Air pollution in metropolitan areas varies with city-level, urban-rural, and neighborhood-level disparities. Our EST paper (featured on the journal cover) introduces a data-driven approach that leverages the real-world dynamic traffic profiles to continuously estimate community-level year-long air pollutants. The ML model shows promising ability to capture the traffic-induced exposure disparities and significantly improve residents’ exposure to PM2.5, especially for disadvantaged communities.
Pollution and Climate Change on Crops
Our paper published on Earth's Future develops a robust crop yield prediction model and reveals a critical role of particulate pollution in determining annual crop yields. The present study demonstrates the co-benefit of the recent air pollution control policy from agriculture and food perspectives. However, this benefit will eventually be diminished after the air pollution becomes alleviated in the full scale, while persisting or even exacerbated global warming will pose larger threat on the future food security.
Air pollution during the COVID-19 Lockdown
We investigate the surprising haze events during the lockdown period in China and untangle complex interplay between emissions, atmospheric chemistry, and meteorological variations. Particularly the importance of the multi-phase aerosol chemistry in the winter humid condition is identified. Our paper calls for a comprehensive regulatory strategy involving all emission sectors and accounting for meteorological variations for future emission control plans.
Predicting Smoke from Wildfires
Wildfires have become increasingly prevalent. Intense smoke consisting of particulate matter (PM) leads to an increased risk of morbidity and mortality. For example, the record-breaking Camp Fire ravaged Northern California for two weeks in 2018. We employ a comprehensive fire-chemical transport model along with ground-based and satellite observations to characterize the PM concentrations across California and to investigate the pollution sensitivity predictions to key parameters of the model.