基于AERONET数据估算近地面PM2.5

Estimating the near-ground PM2.5 concentration based on AERONET data

近年来对PM2.5估算的研究大多是从卫星遥感的气溶胶光学厚度出发,而从气溶胶粒子谱本身出发的研究较少.基于此,本文提出一种新方法,利用2016-2017年AERONRT北京、香河、徐州和太湖4个站点的气溶胶粒径分布数据估算了近地面PM2.5质量浓度,经标高订正和湿度订正后用地面监测数据对估算结果进行评价检验.结果表明:①估算的2016-2017年PM2.5日均值和地面实测数据的拟合度R^2分别为北京0.42、香河0.31、徐州0.05和太湖0.49,经标高订正和湿度订正后分别提升至0.69(RMSE=39.33μg·cm^-3)、0.79(RMSE=35.36μg·cm^-3)、0.49(RMSE=32.93μg·cm^-3)和0.75(RMSE=15.24μg·cm^-3);②将估算的PM2.5季均值与地面实测季均值进行对比分析,结果也显示二者基本相当,同时基于该方法估算了2006-2017年北京和香河地区PM2.5年均值,分析了其变化趋势.由此可见,基于AERONET的粒子谱数据能够较好地估算近地面PM2.5质量浓度,并且可以利用该方法估算PM2.5历史数据,分析变化趋势.

Currently, a great many researches focus on the PM2.5 estimated from aerosol optical depth (AOD) retrieved through satellite remote sensing, but few ones may be presented which would be derived from aerosol size distribution itself by ground-based remote sensing. A new approach is presented in this paper which uses the aerosol size distribution of Beijing, Xianghe, Xuzhou and Taihu sites in Aerosol Robotic Network (AERONRT) from 2016 to 2017 to estimate the near-surface PM2.5 concentration. After corrections of aerosol layer height and relative humidity, the estimated results are evaluated by ground monitor data. They show that:①The R^2 of the estimated and observed daily averaged PM2.5 concentration at the near-ground surface from 2016 to 2017 is 0.42 for Beijing, 0.31 for Xianghe, 0.05 for Xuzhou and 0.49 for Taihu. After corrections of aerosol layer height and relative humidity, R2 values increase to 0.69(RMSE=39.33 μg·cm^-3 ),0.79(RMSE = 35.36 μg·cm^-3 ),0.49 (RMSE = 32.93 μg·cm^-3 ) and 0.75(RMSE = 15.24 μg·cm^-3 ), respectively;②Comparison with estimated and observed values of the seasonal average PM2.5 concentration shows that the two are basically close. Based on this method, the annual averaged PM2.5 concentrations in Beijing and Xianghe over 12 years from 2006 to 2017 have been estimated and the trend of changes has been analyzed. This study suggests that aerosol size distribution of AERONET can be used to estimate the near-surface PM2.5 mass concentration. This method can be also used to estimate PM2.5 historical data and analyze its trend.