GOCI Yonsei aerosol retrieval version 2 products: An improved algorithm and error analysis with uncertainty estimation from 5-year validation over East Asia

Myungje Choi, Jhoon Kim, Jaehwa Lee, Mijin Kim, Young Je Park, Brent Holben, Thomas F. Eck, Zhengqiang Li, Chul H. Song

Research output: Contribution to journalArticle

14 Citations (Scopus)

Abstract

The Geostationary Ocean Color Imager (GOCI) Yonsei aerosol retrieval (YAER) version 1 algorithm was developed to retrieve hourly aerosol optical depth at 550nm (AOD) and other subsidiary aerosol optical properties over East Asia. The GOCI YAER AOD had accuracy comparable to ground-based and other satellite-based observations but still had errors because of uncertainties in surface reflectance and simple cloud masking. In addition, near-real-time (NRT) processing was not possible because a monthly database for each year encompassing the day of retrieval was required for the determination of surface reflectance. This study describes the improved GOCI YAER algorithm version 2 (V2) for NRT processing with improved accuracy based on updates to the cloud-masking and surface-reflectance calculations using a multi-year Rayleigh-corrected reflectance and wind speed database, and inversion channels for surface conditions. The improved GOCI AOD τG is closer to that of the Moderate Resolution Imaging Spectroradiometer (MODIS) and Visible Infrared Imaging Radiometer Suite (VIIRS) AOD than was the case for AOD from the YAER V1 algorithm. The V2 τG has a lower median bias and higher ratio within the MODIS expected error range (0.60 for land and 0.71 for ocean) compared with V1 (0.49 for land and 0.62 for ocean) in a validation test against Aerosol Robotic Network (AERONET) AOD τA from 2011 to 2016. A validation using the Sun-Sky Radiometer Observation Network (SONET) over China shows similar results. The bias of error (τG - τA) is within -0.1 and 0.1, and it is a function of AERONET AOD and Ångström exponent (AE), scattering angle, normalized difference vegetation index (NDVI), cloud fraction and homogeneity of retrieved AOD, and observation time, month, and year. In addition, the diagnostic and prognostic expected error (PEE) of τG are estimated. The estimated PEE of GOCI V2 AOD is well correlated with the actual error over East Asia, and the GOCI V2 AOD over South Korea has a higher ratio within PEE than that over China and Japan.

Original languageEnglish
Pages (from-to)385-408
Number of pages24
JournalAtmospheric Measurement Techniques
Volume11
Issue number1
DOIs
Publication statusPublished - 2018 Jan 17

Fingerprint

error analysis
aerosol
surface reflectance
MODIS
ocean
NDVI
optical property
radiometer
optical depth
homogeneity
GOCI
product
Asia
reflectance
wind velocity
scattering

All Science Journal Classification (ASJC) codes

  • Atmospheric Science

Cite this

Choi, Myungje ; Kim, Jhoon ; Lee, Jaehwa ; Kim, Mijin ; Park, Young Je ; Holben, Brent ; Eck, Thomas F. ; Li, Zhengqiang ; Song, Chul H. / GOCI Yonsei aerosol retrieval version 2 products : An improved algorithm and error analysis with uncertainty estimation from 5-year validation over East Asia. In: Atmospheric Measurement Techniques. 2018 ; Vol. 11, No. 1. pp. 385-408.
@article{6409972371d1482bb2d81f80fee28152,
title = "GOCI Yonsei aerosol retrieval version 2 products: An improved algorithm and error analysis with uncertainty estimation from 5-year validation over East Asia",
abstract = "The Geostationary Ocean Color Imager (GOCI) Yonsei aerosol retrieval (YAER) version 1 algorithm was developed to retrieve hourly aerosol optical depth at 550nm (AOD) and other subsidiary aerosol optical properties over East Asia. The GOCI YAER AOD had accuracy comparable to ground-based and other satellite-based observations but still had errors because of uncertainties in surface reflectance and simple cloud masking. In addition, near-real-time (NRT) processing was not possible because a monthly database for each year encompassing the day of retrieval was required for the determination of surface reflectance. This study describes the improved GOCI YAER algorithm version 2 (V2) for NRT processing with improved accuracy based on updates to the cloud-masking and surface-reflectance calculations using a multi-year Rayleigh-corrected reflectance and wind speed database, and inversion channels for surface conditions. The improved GOCI AOD τG is closer to that of the Moderate Resolution Imaging Spectroradiometer (MODIS) and Visible Infrared Imaging Radiometer Suite (VIIRS) AOD than was the case for AOD from the YAER V1 algorithm. The V2 τG has a lower median bias and higher ratio within the MODIS expected error range (0.60 for land and 0.71 for ocean) compared with V1 (0.49 for land and 0.62 for ocean) in a validation test against Aerosol Robotic Network (AERONET) AOD τA from 2011 to 2016. A validation using the Sun-Sky Radiometer Observation Network (SONET) over China shows similar results. The bias of error (τG - τA) is within -0.1 and 0.1, and it is a function of AERONET AOD and {\AA}ngstr{\"o}m exponent (AE), scattering angle, normalized difference vegetation index (NDVI), cloud fraction and homogeneity of retrieved AOD, and observation time, month, and year. In addition, the diagnostic and prognostic expected error (PEE) of τG are estimated. The estimated PEE of GOCI V2 AOD is well correlated with the actual error over East Asia, and the GOCI V2 AOD over South Korea has a higher ratio within PEE than that over China and Japan.",
author = "Myungje Choi and Jhoon Kim and Jaehwa Lee and Mijin Kim and Park, {Young Je} and Brent Holben and Eck, {Thomas F.} and Zhengqiang Li and Song, {Chul H.}",
year = "2018",
month = "1",
day = "17",
doi = "10.5194/amt-11-385-2018",
language = "English",
volume = "11",
pages = "385--408",
journal = "Atmospheric Measurement Techniques",
issn = "1867-1381",
publisher = "Copernicus Gesellschaft mbH",
number = "1",

}

GOCI Yonsei aerosol retrieval version 2 products : An improved algorithm and error analysis with uncertainty estimation from 5-year validation over East Asia. / Choi, Myungje; Kim, Jhoon; Lee, Jaehwa; Kim, Mijin; Park, Young Je; Holben, Brent; Eck, Thomas F.; Li, Zhengqiang; Song, Chul H.

In: Atmospheric Measurement Techniques, Vol. 11, No. 1, 17.01.2018, p. 385-408.

Research output: Contribution to journalArticle

TY - JOUR

T1 - GOCI Yonsei aerosol retrieval version 2 products

T2 - An improved algorithm and error analysis with uncertainty estimation from 5-year validation over East Asia

AU - Choi, Myungje

AU - Kim, Jhoon

AU - Lee, Jaehwa

AU - Kim, Mijin

AU - Park, Young Je

AU - Holben, Brent

AU - Eck, Thomas F.

AU - Li, Zhengqiang

AU - Song, Chul H.

PY - 2018/1/17

Y1 - 2018/1/17

N2 - The Geostationary Ocean Color Imager (GOCI) Yonsei aerosol retrieval (YAER) version 1 algorithm was developed to retrieve hourly aerosol optical depth at 550nm (AOD) and other subsidiary aerosol optical properties over East Asia. The GOCI YAER AOD had accuracy comparable to ground-based and other satellite-based observations but still had errors because of uncertainties in surface reflectance and simple cloud masking. In addition, near-real-time (NRT) processing was not possible because a monthly database for each year encompassing the day of retrieval was required for the determination of surface reflectance. This study describes the improved GOCI YAER algorithm version 2 (V2) for NRT processing with improved accuracy based on updates to the cloud-masking and surface-reflectance calculations using a multi-year Rayleigh-corrected reflectance and wind speed database, and inversion channels for surface conditions. The improved GOCI AOD τG is closer to that of the Moderate Resolution Imaging Spectroradiometer (MODIS) and Visible Infrared Imaging Radiometer Suite (VIIRS) AOD than was the case for AOD from the YAER V1 algorithm. The V2 τG has a lower median bias and higher ratio within the MODIS expected error range (0.60 for land and 0.71 for ocean) compared with V1 (0.49 for land and 0.62 for ocean) in a validation test against Aerosol Robotic Network (AERONET) AOD τA from 2011 to 2016. A validation using the Sun-Sky Radiometer Observation Network (SONET) over China shows similar results. The bias of error (τG - τA) is within -0.1 and 0.1, and it is a function of AERONET AOD and Ångström exponent (AE), scattering angle, normalized difference vegetation index (NDVI), cloud fraction and homogeneity of retrieved AOD, and observation time, month, and year. In addition, the diagnostic and prognostic expected error (PEE) of τG are estimated. The estimated PEE of GOCI V2 AOD is well correlated with the actual error over East Asia, and the GOCI V2 AOD over South Korea has a higher ratio within PEE than that over China and Japan.

AB - The Geostationary Ocean Color Imager (GOCI) Yonsei aerosol retrieval (YAER) version 1 algorithm was developed to retrieve hourly aerosol optical depth at 550nm (AOD) and other subsidiary aerosol optical properties over East Asia. The GOCI YAER AOD had accuracy comparable to ground-based and other satellite-based observations but still had errors because of uncertainties in surface reflectance and simple cloud masking. In addition, near-real-time (NRT) processing was not possible because a monthly database for each year encompassing the day of retrieval was required for the determination of surface reflectance. This study describes the improved GOCI YAER algorithm version 2 (V2) for NRT processing with improved accuracy based on updates to the cloud-masking and surface-reflectance calculations using a multi-year Rayleigh-corrected reflectance and wind speed database, and inversion channels for surface conditions. The improved GOCI AOD τG is closer to that of the Moderate Resolution Imaging Spectroradiometer (MODIS) and Visible Infrared Imaging Radiometer Suite (VIIRS) AOD than was the case for AOD from the YAER V1 algorithm. The V2 τG has a lower median bias and higher ratio within the MODIS expected error range (0.60 for land and 0.71 for ocean) compared with V1 (0.49 for land and 0.62 for ocean) in a validation test against Aerosol Robotic Network (AERONET) AOD τA from 2011 to 2016. A validation using the Sun-Sky Radiometer Observation Network (SONET) over China shows similar results. The bias of error (τG - τA) is within -0.1 and 0.1, and it is a function of AERONET AOD and Ångström exponent (AE), scattering angle, normalized difference vegetation index (NDVI), cloud fraction and homogeneity of retrieved AOD, and observation time, month, and year. In addition, the diagnostic and prognostic expected error (PEE) of τG are estimated. The estimated PEE of GOCI V2 AOD is well correlated with the actual error over East Asia, and the GOCI V2 AOD over South Korea has a higher ratio within PEE than that over China and Japan.

UR - http://www.scopus.com/inward/record.url?scp=85040793273&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85040793273&partnerID=8YFLogxK

U2 - 10.5194/amt-11-385-2018

DO - 10.5194/amt-11-385-2018

M3 - Article

AN - SCOPUS:85040793273

VL - 11

SP - 385

EP - 408

JO - Atmospheric Measurement Techniques

JF - Atmospheric Measurement Techniques

SN - 1867-1381

IS - 1

ER -