TY - JOUR
T1 - Impact of aerosol layering, complex aerosol mixing, and cloud coverage on high-resolution MAIAC aerosol optical depth measurements
T2 - Fusion of lidar, AERONET, satellite, and ground-based measurements
AU - Rogozovsky, Irina
AU - Ansmann, Albert
AU - Althausen, Dietrich
AU - Heese, Birgit
AU - Engelmann, Ronny
AU - Hofer, Julian
AU - Baars, Holger
AU - Schechner, Yoav
AU - Lyapustin, Alexei
AU - Chudnovsky, Alexandra
N1 - Publisher Copyright:
© 2020 Elsevier Ltd
PY - 2021/2/15
Y1 - 2021/2/15
N2 - Knowledge of the vertical distribution and layering of aerosols and identification of the corresponding aerosol sources are needed to improve our understanding of the spatial and temporal variability of aerosol pollution. To achieve this goal, we combined both passive and active remote-sensing techniques to provide a 3D view of local aerosol levels and regional to long-range pollution transport. We studied aerosol optical depth (AOD) data from the Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm at 1-km spatial resolution along with active multiwavelength polarization lidar observations of vertical aerosol profiles in Haifa, Israel, Aerosol Robotic Network (AERONET) sun photometer observations at the lidar site, and local-network observations of aerosol concentrations (PM2.5). This comprehensive dataset enabled analyzing the performance of the MAIAC AOD retrieval in cases of complex aerosol layering and mixing states which are typical of the Eastern Mediterranean. While satellite-derived and ground-based AOD measurements generally showed good agreement, 35 out of 100 measurements showed low correspondence. Analysis of those cases revealed that overestimation of AOD was mostly related to cloud-contaminated pixels and aerosol water-uptake effects in moist, cloud-free air at cloud level. Furthermore, AOD over- and underestimations were related to the presence of complex aerosol mixture and layering conditions, especially when dust was mixed with aged anthropogenic aerosol pollution and marine aerosols with lofted anthropogenic pollution. In these cases 50–70% of measurements were outside of the expected error limit. Perhaps these conditions are not considered in the MAIAC retrieval. Finally, we investigated the link between AOD spatial variability and the MAIAC AOD bias, and performed a cluster analysis corroborating the strong impact of cloud contamination on MAIAC AOD quality. Our observation-based results raise the importance of carefully analyzing the uncertainties in satellite AOD measurements that are used as an important input variable in numerous health-related exposure studies and climate models.
AB - Knowledge of the vertical distribution and layering of aerosols and identification of the corresponding aerosol sources are needed to improve our understanding of the spatial and temporal variability of aerosol pollution. To achieve this goal, we combined both passive and active remote-sensing techniques to provide a 3D view of local aerosol levels and regional to long-range pollution transport. We studied aerosol optical depth (AOD) data from the Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm at 1-km spatial resolution along with active multiwavelength polarization lidar observations of vertical aerosol profiles in Haifa, Israel, Aerosol Robotic Network (AERONET) sun photometer observations at the lidar site, and local-network observations of aerosol concentrations (PM2.5). This comprehensive dataset enabled analyzing the performance of the MAIAC AOD retrieval in cases of complex aerosol layering and mixing states which are typical of the Eastern Mediterranean. While satellite-derived and ground-based AOD measurements generally showed good agreement, 35 out of 100 measurements showed low correspondence. Analysis of those cases revealed that overestimation of AOD was mostly related to cloud-contaminated pixels and aerosol water-uptake effects in moist, cloud-free air at cloud level. Furthermore, AOD over- and underestimations were related to the presence of complex aerosol mixture and layering conditions, especially when dust was mixed with aged anthropogenic aerosol pollution and marine aerosols with lofted anthropogenic pollution. In these cases 50–70% of measurements were outside of the expected error limit. Perhaps these conditions are not considered in the MAIAC retrieval. Finally, we investigated the link between AOD spatial variability and the MAIAC AOD bias, and performed a cluster analysis corroborating the strong impact of cloud contamination on MAIAC AOD quality. Our observation-based results raise the importance of carefully analyzing the uncertainties in satellite AOD measurements that are used as an important input variable in numerous health-related exposure studies and climate models.
KW - AOD spatial Variance
KW - Aerosol optical depth (AOD)
KW - Cluster analyses
KW - Multi-angle implementation of atmospheric correction (MAIAC)
KW - Polly-lidar
UR - http://www.scopus.com/inward/record.url?scp=85099848230&partnerID=8YFLogxK
U2 - 10.1016/j.atmosenv.2020.118163
DO - 10.1016/j.atmosenv.2020.118163
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???
AN - SCOPUS:85099848230
SN - 1352-2310
VL - 247
JO - Atmospheric Environment
JF - Atmospheric Environment
M1 - 118163
ER -