Evaluation of NASA Deep Blue/SOAR Aerosol Retrieval Algorithms Applied to AVHRR Measurements

• Main Authors: Smirnov, A., Sayer, A. M., Lee, J., Carletta, N., Chen, S -H, Hsu, N. C.
• Format: Other/Unknown Material
• Language: unknown
• Published: 2017
• Subjects: Earth Resources and Remote Sensing; Aerosol Robotic Network
• Online Access: http://hdl.handle.net/2060/20180003221

The Deep Blue (DB) and Satellite Ocean Aerosol Retrieval (SOAR) algorithms have previously been applied to observations from sensors like the Moderate Resolution Imaging Spectroradiometers (MODIS) and Sea-viewing Wide Field-of-view Sensor (SeaWiFS) to provide records of mid visible aerosol optical depth (AOD) and related quantities over land and ocean surfaces respectively. Recently, DB and SOAR have also been applied to Advanced Very High Resolution Radiometer (AVHRR) observations from several platforms (NOAA11, NOAA14, and NOAA18), to demonstrate the potential for extending the DB and SOAR AOD records. This study provides an evaluation of the initial version (V001) of the resulting AVHRR-based AOD data set, including validation against Aerosol Robotic Network (AERONET)and ship-borne observations, and comparison against both other AVHRR AOD records and MODIS/SeaWiFS products at select long-term AERONET sites. Although it is difficult to distil error characteristics into a simple expression,the results suggest that one standard deviation confidence intervals on retrieved AOD of plus or minus (0.03+15 %) over water and plus or minus (0.05+25 %) over land represent the typical level of uncertainty, with a tendency towards negative biases in high-AOD conditions, caused by a combination of algorithmic assumptions and sensor calibration issues. Most of the available validation data are for NOAA18 AVHRR, although performance appear to be similar for the NOAA11 and NOAA14 sensors as well.