Data-Driven Predictive Modeling of Ground-Level Ozone for Climate Action and Public Health

The potential effects of emission reduction efforts in the electric power and transportation sectors may affect future ozone levels, as indicated by data-driven predictive models. The proposed study focuses on developing models for forecasting ground-level ozone (O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf>) concentrations, which are critical for climate action and public health. Using air quality index (AQI) data from 2018 to 2022 collected in Colorado and Washington DC, along with meteorological factors such as temperature and humidity, three forecasting models—ARIMA, SARIMAX, and Holt-Winters—were applied to predict temporal and seasonal trends in O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf> pollution. The ARIMA model captured broader trends but lacked accuracy during extreme events. SARIMAX improved predictions by incorporating seasonal variations and exogenous variables, such as temperature and wildfire events. The Holt-Winters model effectively modeled short-term trends and seasonal fluctuations. Model accuracy was validated using Mean Squared Error (MSE), Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE). Results revealed strong correlations between O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf> AQI and other pollutants, with significant regional variations between the two locations and the importance of integrating environmental and meteorological data to improve predictive accuracy for O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf> pollution, which can inform climate action and public health strategies.