Spatial source, simulating improvement, and short-term health effect of high PM_2.5 exposure during mutation event in...

Air quality in China has significantly improved owing to the effective implementation of pollution control measures. However, mutation events caused by short-term spikes in PM_2.5 in urban agglomeration regions continue to occur frequently. Identifying the spatial sources and influencing factors, as well as improving the prediction accuracy of high PM_2.5 during mutation events, are crucial for public health. In this study, we firstly introduced discrete wavelet transform (DWT) to identify the mutation events with high PM_2.5 concentration in the four key urban agglomerations, and evaluated the spatial sources for the polluted scenario using Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model. Additionally, DWT was combined with a widely used artificial neural network (ANN) to improve the prediction accuracy of PM_2.5 concentration seven days in advance (seven-day forecast). Results indicated that mutation events commonly occurred in the northern regions during winter time, which were under the control of both short-range transportation of dirty airmass as well as negative meteorology conditions. Compared with the ANN model alone, the average band errors decreased by 9% when using DWT-ANN model. The average correlation coefficient (R) and root mean square error (RMSE) obtained using the DWT-ANN improved by 10% and 12% compared to those obtained using the ANN, indicating the efficiency and accuracy of simulating PM_2.5, by combining the DWT and ANN. The short-term mortality during mutation events was then calculated, with the total averted all-cause, cardiovascular, and respiratory deaths in the four regions, being 4751, 2554, and 582 persons, respectively. A declining trend in prevented deaths from 2018 to 2020 demonstrated that the pollution intensity during mutation events gradually decreased owing to the implementation of the Three-Year Action Plan to Win the Blue Sky Defense War. The method proposed in this study can be used by policymakers to take preventive measures in response to a sudden increase in PM_2.5, thereby ensuring public health.