Air quality early-warning system for cities in China. (January 2017)

• Record Type: Journal Article
• Authors: Xu, Yunzhen; Yang, Wendong; Wang, Jianzhou
• Is Part Of: Atmospheric environment. Volume 148(2017)
• Page Start: 239
• Page End: 257
• Publication Date: 2017-01
• Subjects: Air quality early-warning systems -- Hybrid forecasting model -- Data preprocessing -- Bio-inspire optimization algorithm -- Fuzzy evaluation; Air -- Pollution -- Periodicals; Air -- Pollution -- Meteorological aspects -- Periodicals; 551.51
• DOI: 10.1016/j.atmosenv.2016.10.046
• Languages: English
• ISSNs: 1352-2310

Abstract

Abstract: Air pollution has become a serious issue in many developing countries, especially in China, and could generate adverse effects on human beings. Air quality early-warning systems play an increasingly significant role in regulatory plans that reduce and control emissions of air pollutants and inform the public in advance when harmful air pollution is foreseen. However, building a robust early-warning system that will improve the ability of early-warning is not only a challenge but also a critical issue for the entire society. Relevant research is still poor in China and cannot always satisfy the growing requirements of regulatory planning, despite the issue's significance. Therefore, in this paper, a hybrid air quality early-warning system was successfully developed, composed of forecasting and evaluation. First, a hybrid forecasting model was proposed as an important part of this system based on the theory of "decomposition and ensemble" and combined with the advanced data processing technique, support vector machine, the latest bio-inspired optimization algorithm and the leave-one-out strategy for deciding weights. Afterwards, to intensify the research, fuzzy evaluation was performed, which also plays an indispensable role in the early-warning system. The forecasting model and fuzzy evaluation approaches are complementary. Case studies using daily air pollution concentrations of six air pollutants from three cities in China ( i.e., Taiyuan, Harbin and Chongqing) are used as examples to evaluate the efficiency and effectiveness of the developed air quality early-warning system. Experimental results demonstrate that both the accuracy and the effectiveness of the developed system are greatly superior for air quality early warning. Furthermore, the application of forecasting and evaluation enables the informative and effective quantification of future air quality, offering a significant advantage, and can be employed to develop rapid air quality early-warning systems. Graphical abstract: Highlights: A novel air quality early-warning system is developed for cities in China. A hybrid model named ICEEMD-WOA-SVM is proposed for forecasting major pollutants. The fuzzy framework is used to transform forecast results into air-quality forecasts. A novel deciding weight method is proposed based on leave-one-out strategy and WOA. The results are validated well in Taiyuan, Harbin and Chongqing, China.