Ground-Level PM_2.5 Concentration Estimation from Satellite Data in the Beijing Area Using a Specific Particle Swarm Extinction Mass Conversion Algorithm

Abstract

Particulate matter (PM) has a substantial influence on the environment, climate change and public health. Due to the limited spatial coverage of a ground-level PM2.5 monitoring system, the ground-based PM2.5 concentration measurement is insufficient in many circumstances. In this paper, a Specific Particle Swarm Extinction Mass Conversion Algorithm (SPSEMCA) using remotely sensed data is introduced. Ground-level observed PM2.5, planetary boundary layer height (PBLH) and relative humidity (RH) reanalyzed by the European Centre for Medium-Range Weather Forecasts (ECMWF) and aerosol optical depth (AOD), fine-mode fraction (FMF), particle size distribution, and refractive indices from AERONET (Aerosol Robotic Network) of the Beijing area in 2015 were used to establish this algorithm, and the same datasets for 2016 were used to test the performance of the SPSEMCA. The SPSEMCA involves four steps to obtain PM2.5 values from AOD datasets, and every step has certain advantages: (I) In the particle correction, we use η2.5 (the extinction fraction caused by particles with a diameter less than 2.5 μm) to make an accurate assimilation of AOD2.5, which is contributed to by the specific particle swarm PM2.5. (II) In the vertical correction, we compare the performance of PBLHc retrieved by satellite Lidar CALIPSO data and PBLHe reanalysis by ECMWF. Then, PBLHc is used to make a systematic correction for PBLHe. (III) For extinction to volume conversion, the relative humidity and the FMF are used together to assimilate the AVEC (averaged volume extinction coefficient, μm2/μm3). (IV) PM2.5 measured by ground-based air quality stations are used as the dry mass concentration when calculating the AMV (averaged mass volume, cm3/g) in humidity correction, that will avoid the uncertainties derived from the estimation of the particulate matter density ρ. (V) Multi-Angle Implementation of Atmospheric Correction (MAIAC) 1 km × 1 km AOD was used to retrieve high resolution PM2.5, and a LookUP Table-based Spectral Deconvolution Algorithm (LUT-SDA) FMF was used to avoid the large uncertainties caused by the MODIS FMF product. The validation of PM2.5 from the SPSEMCA algorithm to the AERONET observation data and MODIS monitoring data achieved acceptable results, R = 0.70, RMSE (root mean square error) = 58.75 μg/m3 for AERONET data, R = 0.75, RMSE = 43.38 μg/m3 for MODIS data, respectively. Furthermore, the trend of the temporal and spatial distribution of Beijing was revealed.

Keywords

• PM_2.5
• AOD
• fine mode fraction
• MODIS