Filling the missing data gaps of daily MODIS AOD using spatiotemporal interpolation

Aerosol is an important component of the atmosphere that affects the environment, climate, and human health. Remote sensing is an efficient observation method for monitoring global aerosol distribution and changes over time. The daily Moderate Resolution Imaging Spectroradiometer (MODIS) level-2 aerosol optical depth (AOD) (Collection 6) product (10 km resolution) is often used to study climate change and air pollution. However, the product is prone to yielding large amounts of data gaps due to the unfeasibility of retrieving reliable estimates under cloudy conditions, and these data gaps inevitably affect the results and analysis of the product's application. In this study, a geostatistical data interpolation framework based on the spatiotemporal kriging method was implemented to interpolate satellite AOD products in Beijing, China. Compared to the ordinary kriging method for filling data gaps, the spatiotemporal interpolation not only utilizes spatial autocorrelation but also considers the temporal and spatiotemporal autocorrelations between different locations. In the study region, the completeness of the spatiotemporal-interpolated AOD product reaches 67.73%, which is significantly superior to the completeness of the original MODIS product (14.27%) and that of the spatial kriging-interpolated AOD product (33.3%). The cross-validation results show that the mean absolute error of the spatiotemporal kriging results (0.07) is lower than that of the ordinary kriging (0.09).