Yao Liyin, Han Yan, Qi Xin, Huang Dasheng, Che Hanxiong, Long Xin, Du Yang, Meng Lingshuo, Yao Xiaojiang, Zhang Liuyi, Chen Yang
College of Environmental and Chemical Engineering, Chongqing Three Gorges University, Chongqing 404199, China; Research Center for Atmospheric Environment, Chongqing Institute of Green and Intelligent Technology, Chinese Academy of Sciences, Chongqing 400714, China.
Research Center for Atmospheric Environment, Chongqing Institute of Green and Intelligent Technology, Chinese Academy of Sciences, Chongqing 400714, China.
Sci Total Environ. 2024 Jul 15;934:173193. doi: 10.1016/j.scitotenv.2024.173193. Epub 2024 May 12.
O pollution in China has become prominent in recent years, and it has become one of the most challenging issues in air pollution control. We used data on atmospheric pollutants and meteorology from 2019 to 2021 to build an interpretable random forest (RF) model, applying this model to predict O concentration in 2022 in five cities in the Southwest North China Plain. The model was also used to identify and explain the influence of various factors on O formation. The correlation coefficient R between the predicted O concentration and observed O concentration was 0.82, the MAE was 15.15 μg/m, and the RMSE was 20.29 μg/m, indicating that the model can effectively predict O concentration in the studying area. The results of correlation analysis, feature importance, and the driving factor analysis from SHapley Additive exPlanations (SHAP) model indicated that temperature (T), NO, and relative humidity (RH) are the top three features affecting O prediction, while the weights of wind speed and wind direction were relatively low. Thus, O in the southwestern North China Plain may mainly come from the formation of local photochemical activities. The dominant factors behind O also varied in different seasons. In spring and autumn, O pollution is more likely to occur under high NO concentration and high-temperature conditions, while in summer, it is more likely to occur under high-temperature and precipitation-free weather. In winter, NO is the dominant factor in O formation. Finally, the interpretable RF model is used to predict future O concentration based on features provided by Community Multiscale Air Quality (CMAQ) and Weather Research & Forecast (WRF) model, and the simulation performance of CMAQ on O concentration is enhanced to a certain extent, improving the prediction of future O pollution situations and guiding pollution control.
近年来，中国的臭氧污染问题日益突出，已成为空气污染控制中最具挑战性的问题之一。我们利用2019年至2021年的大气污染物和气象数据构建了一个可解释的随机森林（RF）模型，并将该模型应用于预测华北平原西南部五个城市2022年的臭氧浓度。该模型还用于识别和解释各种因素对臭氧形成的影响。预测的臭氧浓度与观测的臭氧浓度之间的相关系数R为0.82，平均绝对误差（MAE）为15.15μg/m，均方根误差（RMSE）为20.29μg/m，表明该模型能够有效地预测研究区域内的臭氧浓度。相关性分析、特征重要性分析以及基于SHapley加性解释（SHAP）模型的驱动因素分析结果表明，温度（T）、一氧化氮（NO）和相对湿度（RH）是影响臭氧预测的前三大特征，而风速和风向的权重相对较低。因此，华北平原西南部的臭氧可能主要来自当地光化学活动的形成。臭氧背后的主导因素在不同季节也有所不同。在春季和秋季，臭氧污染更有可能在高NO浓度和高温条件下发生，而在夏季，更有可能在高温和无降水天气下发生。在冬季，NO是臭氧形成的主导因素。最后，利用可解释的RF模型，基于社区多尺度空气质量（CMAQ）和天气研究与预报（WRF）模型提供的特征预测未来的臭氧浓度，在一定程度上提高了CMAQ对臭氧浓度的模拟性能，改善了对未来臭氧污染状况的预测并指导污染控制。
