Long-term statistical assessment of Aqua-MODIS aerosol optical depth over coastal regions: bias characteristics and uncertainty sources

• Main Authors: Jacob C Anderson, Jun Wang, Jing Zeng, Gregory Leptoukh, Maksym Petrenko, Charles Ichoku, Chuanmin Hu
• Other Authors: The Pennsylvania State University CiteSeerX Archives
• Format: Text
• Language: English
• Subjects: Merra; Aerosol Robotic Network
• Online Access: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.1050.8587; http://eas2.unl.edu/%7Ejwang/docs/publication/paper_pdf/2013/Anderson_2013.pdf

A B S T R A C T Coastal regions around the globe represent a major source for anthropogenic aerosols in the atmosphere, but the surface characteristics may not be optimal for the Moderate Resolution Imaging Spectroradiometer (MODIS) algorithms designed for aerosol retrievals over dark land or ocean surfaces. Using data collected from 62 coastal stations worldwide by the Aerosol Robotic Network (AERONET) in 2002Á2011, statistical assessments of uncertainties are conducted for coastal aerosol optical depth (AOD) retrieved from MODIS measurements aboard the Aqua satellite (i.e., the Collection 5.1 MYD04 data product generated by the MODIS atmosphere group). It is found that coastal AODs (at 550 nm) characterised respectively by the Dark Land algorithm and the Dark Ocean algorithm all exhibit a log-normal distribution, which contrasts to the near-normal distribution of their corresponding biases. After data filtering using quality flags, the MODIS AODs from both the Dark Land and Dark Ocean algorithms over coastal regions are highly correlated with AERONET AODs (R 2 :0.8), but both have larger uncertainties than their counterparts (of MODIS AODs) over land and open ocean. Overall, the Dark Ocean algorithm overestimates the AERONET coastal AOD by 0.021 for AOD B 0.25 and underestimates it by 0.029 for AOD ! 0.25. This dichotomy is shown to be related to the ocean-surface wind speed and cloud-contamination effects on the MODIS aerosol retrievals. Consequently, an empirical correction scheme is formulated that uses cloud fraction and sea-surface wind speed from Modern Era Retrospective-Analysis for Research and Applications (MERRA) to correct the AOD bias from the Dark Ocean algorithm, and it is shown to be effective over the majority of the coastal AERONET stations to (a) simultaneously reduce both the mean and the spread of the bias and (b) improve the trend analysis of AOD. Further correlation analysis performed after such an empirical bias correction shows that the MODIS AOD is also likely impacted by the ...