Gap-filling MODIS daily aerosol optical depth products by developing a spatiotemporal fitting algorithm

Aerosol loadings and their spatial distribution are among the most important atmospheric information needed for a range of applications such as air quality monitoring, climate research, public health. A key measure aerosol quantity is optical depth (AOD) it has been routinely observed from space by Earth observing satellites/instrument, especially Moderate Resolution Imaging Spectroradiometer (MODIS). Despite its global coverage daily temporal resolution, MODIS Multi-Angle Implementation Atmospheric Correction (MAIAC) AOD product fraught with missing values, severely limiting use. gap-filling method which suitable large-area application high efficiency to obtain gapless reasonable pattern complete still lacking. Here, we proposed novel spatiotemporal fitting algorithm gap-fill product. Our multi-stage aimed address non-stationary nature time series. First, trend in year each pixel was fitted via smoothing splines residual derived based on original data trend. Second, spatially interpolated, leveraging correlation between target neighboring pixels. Third, actual calculated sum interpolated residual. We tested against ground-based 2011 2018 China further evaluated cross-validation at scale 10 selected tiles. Compared ground-reference AOD, RMSE our datasets were 0.24 0.27 Terra Aqua, respectively; showed ranging 0.045 0.055 (Terra) 0.047 0.057 (Aqua) under different ratios. The outperforms Interpolation-based Correlation Weighting (ICW) Inverse Distance (IDW) algorithms accuracy. Meanwhile, using shows lower accuracy than MAIAC similar Long-term Gap-free High-resolution Air Pollutants (LGHAP) concentration dataset, higher Modern-Era Retrospective Analysis Research Applications, Version 2 (MERRA-2). Overall, this meets need typical relevant studies. transferable other regions, potential be used even operationally efficiently generating accurate input only data.