Air quality has influence on people's health and life. In addition to being directly affected by social activities and atmospheric emissions, the impacts of meteorological factors are also significant. Based on daily Air Quality Index (AQI) data and various meteorological parameters in the Yangtze River Delta (YRD), this paper summarized the spatiotemporal evolution characteristics of AQI over YRD, and quantitatively analyzed the contribution of different meteorological elements to air quality. We also evaluated different spatial interpolation methods to produce surface distribution of AQI, and noted that the Ordinary Kriging outperformed other methods. The spatial distribution of AQI in YRD showed seasonal and annual variations. However, the days with AQI over 100 (level ii) were mostly observed in winter. Generally, more severe air pollution was observed in the northern part of YRD than in the southern ones, for example the air quality of the Ningbo metropolitan area was the best, while in Hefei it was the worst. It was found that meteorological parameters have spatially varying effects on AQI. For instance, pressure has a significant positive effect on AQI, and others showed negative correlations. We also predicted AQI by exploiting different machine learning-based models. Through model comparison, it was found that the Autoregressive Integrated Moving Average Model (ARIMA (0,1,2)) has higher prediction accuracy for AQI than Multiple Linear Regression (MLR). The findings of this research can be used in future forecasting of air pollution, and also in air pollution controlling programs.

25.3.2020