After more than a decade of steady decreases, surface ozone (O₃) trends in North America reversed in 2015 despite policy mitigation efforts, report Weizhi Deng and colleagues. Their research links this reversal to O₃ emissions from wildfires and additionally documents a related rise in premature mortality. “Despite regulated reductions in anthropogenic emissions of O₃ precursors, observation stations indicate that policy-relevant surface O₃ levels have plateaued,” the authors write, tying this phenomenon to an increase in wildfire emissions. They describe the relationship between wildfires and surface O₃ trends more closely by using deep learning models to evaluate existing yet sparse EPA, satellite, and meteorological measurements, generating a dataset of daily surface O₃ measurements at a 1-kilometer resolution in North America from 2003 to 2024. Doing so revealed that O₃ trends flipped from a decrease of 0.65 parts per billion (ppb) per year before 2015 to an increase of 0.13 ppb per year after 2015. Further analyses determined this post-2015 rate would have stayed in decline (−0.25 ppb per year) if not for wildfire emissions. The authors then examined correlations between O₃ trends and premature deaths, attributing emissions to an additional 318 deaths per year since 2013. Essentially, after 2013, the mortality rate attributable to wildfire-sourced O₃ rose by 46%. Finally, Deng et al. examined O₃ emissions from 2022 to 2024, a period marked by extreme fires and smoke in Canada. Results showed wildfire emissions alone exposed 43 million people to unhealthy levels of air pollution, in excess of the United States’ O₃ air quality standard of 70 ppb. The authors suggest that these emissions prevented the United States from tightening its O₃ air quality standard by 4 ppb. They elaborate: “If the O₃ standard were lowered to 65 ppb, 60% of the population (202 million people) would fall into nonattainment, and under a 60-ppb standard, the fraction would increase to 87% (294 million people). These findings demonstrate the challenge in adopting a more stringent O₃ standard as growing wildfires contribute to high O₃ episodes.”

Concordia University researchers have developed an AI-powered operating room scheduling tool that plans surgeries, manages last-minute emergencies, and minimizes patient disruptions within a single framework. Tested on real hospital data, the model reduces variables compared to existing approaches, cutting wait times and day-of-surgery cancellations. The study appears in the International Journal of Production Research.

Boston University is joining the NSF AI Institute for Artificial Intelligence and Fundamental Interactions (IAIFI), which aims to use AI to inform physics breakthroughs—and vice versa.