MAIAC algorithm calibration uncertainties for sites with low AOD values. Case study: Metropolitan Area of Buenos Aires

URQUIZA, JOSEFINA; ARIEL SCAGLIOTTI; DIEZ, SEBASTIAN

On - line

Congreso; 16th IGAC Scientific Conference Atmospheric Chemistry from a Distance: Real Progress through Virtual Interaction; 2021

International Global Atmospheric Chemistry (IGAC) Project

Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm has been developed from the MODIS sensor measurements (on-board Terra and Aqua satellite), and combines a time series approach and image processing to derive atmospheric aerosols products. The dataset obtained includes Aerosol Optical Depth (AOD) at 470 nm and 550 nm with high spatial resolution (1 km), which for example could be useful for studying the aerosol variability at the urban-scale. This product is generally calibrated with help of ground-based AOD measurements from solar photometers, being the AERONET network a common method to validate satellite AOD retrievals. Even though the AERONET network provides AOD measurements at various wavelengths, they are not at the same MAIAC wavelength. In order to validate MAIAC, it is used the closest AERONETwavelength or an interpolation is introduced in order to match it up (i.e. to 550 nm and 470 nm). Furthermore, the MAIAC calibration is a complex task since the ground-truth data corresponds to a specific location and 15-min time resolution, while the satellite product provides data few times a day for a wider area. As the whole procedure has associated errors that will impact on the calibrated data they must be considered and propagated when analyzing the data. This work aims to study the MAIAC calibration uncertainties using as a case study the Metropolitan Area of Buenos Aires, Argentina, where relatively low AOD values are observed. The results showed that for low AOD values (0.05 - 0.15), the MAIAC relative expanded uncertainty (REU) is always greater than 20% while for higher AOD values, the REU decreases, observing similar patterns for the consideredwavelengths.