Integrating machine learning with chemical transport modeling to assess ozone-related mortality and crop yield losses in China

Surface ozone (O 3 ) pollution has become an increasingly important environmental issue in China, posing substantial risks to human health and crop yield. In this study, we developed a machine learning bias-correction framework that integrates Weather Research and Forecasting (WRF) and Community Multiscale Air Quality (CMAQ) simulations with multisource meteorological and emission predictors to improve surface O 3 estimates across China during 2015–2018. The corrected O 3 fields were used to assess non-accidental premature mortality and relative yield losses of four major crops. The machine learning model substantially improved O 3 simulation performance, increasing R 2 from 0.60 to 0.79 and reducing the normalized mean bias (NMB) and normalized mean error (NME) by 64% and 22%, respectively. Based on the corrected O 3 fields, O 3 exposure was associated with 76.9 thousand non-accidental premature deaths per year during 2015–2018 (95% Posterior Interval (PI): 41.8–111.7 thousand), lower than CMAQ estimates of 97.7 thousand yr⁻¹ (95% PI: 53.1–141.8 thousand) and closer to observation-interpolated estimates of 82.9 thousand yr⁻¹ (95% PI: 45.1–120.4 thousand). The highest health burdens were concentrated in the North China Plain (NCP), the Yangtze River Delta (YRD), the Pearl River Delta (PRD), and the Sichuan-Chongqing region (SC). O 3 also caused substantial crop yield losses, with machine learning-corrected estimates showing the highest relative yield losses for wheat (9.4%–19.0%), followed by late rice (6.6%–8.9%), while maize, single rice, and early rice showed comparatively lower losses. These results demonstrate that machine learning bias correction can effectively reduce systematic errors in chemical transport model outputs and improve downstream health and crop impact assessments. This study provides a useful framework for refining O 3 exposure estimates and informing coordinated air quality management, public health protection, and food security in China.