Scientists make game-changing breakthrough in battle against smog: 'We've gained a clearer picture'

Researchers from the Chinese Academy of Sciences have pioneered a complex detection model for elevated ground-level ozone powered by artificial intelligence.

In an ecological context, "ozone" — a gas that is both naturally occurring in the atmosphere and human made — is often associated with "the ozone layer," a protective portion of the Earth's stratosphere that shields humans from the most harmful effects of ultraviolet rays.

In the 1980s, researchers discovered that the use of refrigerants and aerosol sprays had caused the ozone layer to thin precipitously. In response, governments heeded the warnings of scientists and reversed course, with the two largest holes set to be "healed" completely by 2066.

Ground-level ozone (O3) is not protective, though, and it's a key component of a form of air pollution known as "smog." 

According to the Environmental Protection Agency, ground-level ozone forms when emissions and other airborne contaminants interact with sunlight.

Ground-level ozone is also responsible for an estimated 365,000 deaths worldwide each year, 70% of which occur in India and China. 

High levels of ground-level ozone are broadly detrimental to human health and longevity. Research indicates the contaminant can have adverse effects on fetal brain development. O3 is also devastating to crops and, ultimately, the global food supply. 

In China, where the issue is particularly pronounced, the team of researchers harnessed a sequential convolutional long short-term memory network, as the academy's report detailed. This technology allows for a deeper integrated analysis of visual data in a chronological and sequential framework.

"Conventional machine learning models often neglect these spatiotemporal dynamics, while numerical models suffer from high computational costs and limited ability to predict high-concentration ozone episodes," the researchers noted.

Their results were compelling. Their "model achieved high prediction accuracy" of 83% in the North China Plain and 56% for the Yangtze River Delta.

Do you worry about air pollution in and around your home? Yes — always 💯 Yes — often 😷 Yes — sometimes 😟 No — never 🙅 Click your choice to see results and speak your mind.

Overall, their model appeared to overcome existing limitations in forecasting models currently in use. It further "successfully quantified the impact of typhoon position shifts on regional ozone levels," adding another layer of functionality to their novel approach to quantifying ground-level ozone.

Professor Pinhua Xie led the study. Xie explained in the academy's report why the model could materially offset the adverse effects of O3 pollution in China due to the robust data it generates.

"We've gained a clearer picture of how weather patterns drive ozone pollution, which can really support better early warnings for high-risk ozone days," Xie remarked.

Join our free newsletter for weekly updates on the latest innovations improving our lives and shaping our future, and don't miss this cool list of easy ways to help yourself while helping the planet.