A Hierarchical Bayesian Approach for Aerosol Retrieval Using MISR Data

• Activity Number: 420
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Yueqing Wang*+ and Xin Jiang and Ming Jiang and Bin Yu
• Companies: University of California at Berkeley and Peking University and Peking University and University of California at Berkeley
• Keywords: Aerosol retrieval ; Multi-angle Imaging SpectroRadiometer ; Hierarchical Bayesian mode ; Markov Chain Monte Carlo ; Maximum a Posteriori ; Remote sensing

Abstract

Atmospheric aerosols can cause serious damage to human health. Using observations by NASA's Multi-angle Imaging SpectroRadiometer (MISR), the current operational algorithm retrieves Aerosol Optical Depth (AOD) at a resolution of 17.6 km. A systematic study of aerosols and their impact on public health, especially in highly-populated urban areas, requires a finer-resolution estimate of AOD. We embed MISR's operational weighted least squares criterion and forward simulations for AOD retrieval in a likelihood framework and expand it into a Bayesian hierarchical model to adapt to a finer spatial scale of 4.4 km. Our method borrows strength from data at neighboring pixels by postulating a Gaussian Markov Random Field prior for AOD. Our model considers both AOD and aerosol mixing vectors as continuous variables. Their inference is carried out using Metropolis-within-Gibbs sampling methods. We also implement a parallel MCMC algorithm to reduce computational cost. Based on case studies in the greater Beijing area, we show that a 4.4 km resolution can improve the accuracy and coverage of remotely-sensed AOD retrievals and our understanding of the spatial and seasonal behaviors of aerosols.