Bayesian Aerosol Retrieval-Based PM[sub.2.5] Estimation through Hierarchical Gaussian Process Models

Satellite-based aerosol optical depth (AOD) data are widely used to estimate land surface PM[sub.2.5] concentrations in areas not covered by ground PM[sub.2.5] monitoring stations. However, AOD data obtained from satellites are typically at coarse spatial resolutions, limiting their applications on small or medium scales. In this paper, we propose a new two-step approach to estimate 1-km-resolution PM[sub.2.5] concentrations in Shanghai using high spatial resolution AOD retrievals from MODIS. In the first step, AOD data are refined to a 1×1 km[sup.2] resolution via a Bayesian AOD retrieval method. In the second step, a hierarchical Gaussian process model is used to estimate PM[sub.2.5] concentrations. We evaluate our approach by model fitting and out-of-sample cross-validation. Our results show that the proposed approach enjoys accurate predictive performance in estimating PM[sub.2.5] concentrations.