| Summary: | 碩士 === 崑山科技大學 === 環境工程研究所 === 104 === In this study we analyze open 6-year PM2.5 (particles with aerodynamic diameters less than 2.5 μm) data for tipping points of high haze pollution events in Beijing, China. First, we analyze basic statistic parameters, such as annual mean and standard deviation. Second, we select three statistic indicators (standard deviation, skewness, and autocorrelation) and try various sliding windows (300, 370, 500, 1000, and 1200-days). Finding that 500-day of leading indicators can extract clear signals of critical transitions (CTs). Based on this, we identify five CTs, four are flickering and one is critical slowing down (CSD) in the period of 2009–2014. Third, we also find alternative stable states are continuing in process that implies atmospheric PM2.5 variation caused by various external driving forces, e.g., wind speeds and aerosol optical depth. That implies need effective pre-warning technologies to predict heavy haze pollution events.
|
|---|
