A MISR-Based Method for the Estimation of Particle Size Distribution: Comparison with AERONET over China

• Published in: Journal of Remote Sensing
• Main Authors: Shao, Yanchuan, Liu, Riyang, Li, Weihan, Bi, Jun, Ma, Zongwei
• Format: Article in Journal/Newspaper
• Language: English
• Published: American Association for the Advancement of Science (AAAS) 2023
• Subjects: Aerosol Robotic Network
• Online Access: http://dx.doi.org/10.34133/remotesensing.0032; https://spj.science.org/doi/pdf/10.34133/remotesensing.0032

Aerosol particle size has a crucial impact on the environment and public health. Current satellite-based regression models focus on the total amount of particles and are limited by surface observations. This study proposes an algorithm to derive the long-term normalized volume size distribution (VSD) of aerosol particles, which is independent of ground measurements. The size distribution and aerosol optical depth of Multi-angle Imaging SpectroRadiometer (MISR) components are employed. We find the estimated MISR VSD is consistent with Aerosol Robotic Network (AERONET) observations, with R = 0.56, 0.54, 0.59, and 0.68 for daily, monthly, seasonal, and annual levels. The stratified validations of radius, stations, and years further confirm the stable performance of derived VSD ( R = 0.28 to 0.73). The application of the random forest model demonstrates the potential improvements of predicted VSD by 10-fold cross-validation R = 0.86 at the monthly level. We apply MISR VSD to quantify the normalized volume of fractional aerosol particles at a resolution of 0.2° × 0.2° during 2004 to 2016 in China. We also calculate the proportion of small and medium particles to indicate the contribution of anthropogenic aerosols. The highest ratios are concentrated in the northeastern regions especially during winter while relatively lower in the Taklamakan Desert of western China. The case study demonstrates that the application of MISR data can yield valuable and resolved size distributions of aerosol particles.