Retrieving the Vertical Distribution of PM $_{2.5}$ Mass Concentration From Lidar Via a Random Forest Model

The vertical distribution of fine particles with a diameter <2.5 μm (PM $_{2.5}$ ) plays an important role in understanding the transport of air pollution and in making decisions regarding the prevention and control of regional air pollution. However, the studies of the vertical distribution of PM $_{2.5}$ were limited by the lack of monitoring data obtained with vertical sampling strategies. The lidar system can obtain the aerosol profile, which provides the possibility to measure PM $_{2.5}$ profile. Here, the vertical distributions of PM $_{2.5}$ concentrations were investigated on the basis of lidar data from January 2014 to October 2015. Linear regression, improved linear regression, and random forest (RF) models were used to retrieve the PM $_{2.5}$ concentration profile from lidar data. The models were built based on the relationship among extinction coefficient (EC), temperature (T), relative humidity (RH), and surface PM $_{2.5}$ mass concentration. Comparison of the estimated and observed PM $_{2.5}$ showed that the RF model exhibited the best inversion effect. The correlation coefficient reached 0.75, and the root mean absolute error (RMAE) and root mean square error (RMSE) were 3.94 and 21.1 μg/m³, respectively. Error analysis indicated that the estimated PM $_{2.5}$ retrieved using the linear and improved linear models (ILMs) was smaller than the observed PM $_{2.5}$ when EC was less than 0.7 km⁻¹, whereas PM $_{2.5}$ was evidently overestimated during winter pollution days. The reason might be that the effects of T and RH were inaccurately considered. Finally, the seasonal variation of the PM $_{2.5}$ profiles was investigated. Results indicated that the mass concentration of PM $_{2.5}$ was relatively large within 0.5-1.5 km, with a maximum of 60 μg/m³. The findings obtained here provide guidance for PM $_{2.5}$ vertical observation and regional pollutant transport.