School of Environmental and Chemical Engineering, Xi'an Polytechnic University, Xi'an 710048, China; Shaanxi Key Laboratory of Land Consolidation, Xi'an 710075, China; State Key Laboratory of Urban and Regional Ecology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, China.
School of Environmental and Chemical Engineering, Xi'an Polytechnic University, Xi'an 710048, China.
Ecotoxicol Environ Saf. 2021 Dec 1;225:112772. doi: 10.1016/j.ecoenv.2021.112772. Epub 2021 Sep 13.
With rapid economic growth, urbanization and industrialization, fine particulate matter with aerodynamic diameters ≤ 2.5 µm (PM) has become a major pollutant and shows adverse effects on both human health and the atmospheric environment. Many studies on estimating PM concentrations have been performed using statistical regression models and satellite remote sensing. However, the accuracy of PM concentration estimates is limited by traditional regression models; machine learning methods have high predictive power, but fewer studies have been performed on the complementary advantages of different approaches. This study estimates PM concentrations from satellite remote sensing-derived aerosol optical depth (AOD) products, meteorological data, terrain data and other predictors in 2015 in Shaanxi, China, using a combined genetic algorithm-support vector machine (GA-SVM) method, after which the spatial clustering pattern was explored at the season and year levels. The results indicated that temperature (r = -0.684), precipitation (r = -0.602) and normalized difference vegetation index (NDVI) (r = -0.523) were significantly negatively correlated with the PM concentration, while AOD (r = 0.337) was significantly positively correlated with the PM concentration. Compared to conventional land use regression (LUR) and SVM models and previous related studies, the GA-SVM method demonstrated a significantly better prediction accuracy of PM concentration, with a higher 10-fold cross-validation coefficient of determination (R) of 0.84 and lower root mean square error (RMSE) and mean absolute error (MAE) of 12.1 μg/m and 10.07 μg/m, respectively. Y-scrambling test shows that the models have no chance correlation. The central and southern parts of Shaanxi have high PM concentrations, which are mainly due to the pollutant emissions and meteorological and topographical conditions in those areas. There was a positive spatial agglomeration characteristic of regional PM pollution, and the spatial spillover effect of PM pollution for seasonal and annual variations does exist. In general, the GA-SVM method is robust and accurately estimates PM concentrations via a novel modeling framework application and high-quality spatiotemporal information. It also has great significance for the exploration of PM pollution estimation and high-precision mapping methods, especially early warning in high-risk areas. Finally, the prevention and control of atmospheric pollution should take pollution control measures from major cities and surrounding cities, and focus on the joint pollution control measures for plain cities.
随着经济的快速增长、城市化和工业化，空气动力学直径≤2.5μm 的细颗粒物（PM）已成为主要污染物，对人类健康和大气环境都产生了不利影响。许多研究采用统计回归模型和卫星遥感来估算 PM 浓度。然而，传统回归模型的 PM 浓度估算精度有限；机器学习方法具有较高的预测能力，但不同方法的互补优势研究较少。本研究采用遗传算法-支持向量机（GA-SVM）方法，结合卫星遥感气溶胶光学厚度（AOD）产品、气象数据、地形数据和其他预测因子，估算了 2015 年中国陕西省的 PM 浓度，然后在季节和年水平上探索了空间聚类模式。结果表明，温度（r=-0.684）、降水（r=-0.602）和归一化差异植被指数（NDVI）（r=-0.523）与 PM 浓度呈显著负相关，而 AOD（r=0.337）与 PM 浓度呈显著正相关。与传统的土地利用回归（LUR）和 SVM 模型以及以前的相关研究相比，GA-SVM 方法对 PM 浓度的预测精度有了显著提高，10 倍交叉验证决定系数（R）更高，为 0.84，均方根误差（RMSE）和平均绝对误差（MAE）更低，分别为 12.1μg/m 和 10.07μg/m。Y 打乱检验表明模型没有机会相关性。陕西中南部地区 PM 浓度较高，主要是由于该地区的污染物排放和气象、地形条件。区域 PM 污染存在正向空间集聚特征，且 PM 污染的空间溢出效应在季节和年际变化中确实存在。总的来说，GA-SVM 方法通过新颖的建模框架应用和高质量的时空信息稳健且准确地估算 PM 浓度，对于 PM 污染估算和高精度制图方法的探索具有重要意义，尤其是在高风险地区的预警方面。最后，大气污染的防治应从大城市及其周边城市采取污染控制措施，并重点采取平原城市的联合污染控制措施。
