TY - JOUR
T1 - Detection of daytime arctic clouds using MISR and MODIS data
AU - Shi, Tao
AU - Clothiaux, Eugene E.
AU - Yu, Bin
AU - Braverman, Amy J.
AU - Groff, David N.
N1 - Funding Information:
Tao Shi and Bin Yu were partially supported by NSF grants CCR-0106656. Bin Yu also benefited from support of an NSF grant DMS-03036508 and ARO grant W911NF-05-1-0104 and a Miller Research Professorship in spring 2004 from the Miller Institute for Basic Research at University of California at Berkeley. For this research Eugene Clothiaux and David Groff were supported by NASA grant NNG04GL93G and Jet Propulsion Laboratory, California Institute of Technology, contract 1259588. Amy Braverman's work is performed at the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. All MISR data were obtained from the NASA Langley Research Center Atmospheric Science Data Center. All MODIS data were obtained from the NASA Goddard Space Flight Center Earth Sciences Data and Information Services Center. The authors would like to thank Dominic Mazzoni, Larry Di Girolamo, David Diner, Roger Davies, and Ralph Kahn for helpful discussions and suggestions. The reviewers of this paper were extraordinary, providing many, many detailed comments and useful insights. If problems remain in the paper, they are there in spite of the considerable help from the reviewers and are completely our own.
PY - 2007/3/15
Y1 - 2007/3/15
N2 - Expert labels were used to evaluate arctic cloud detection accuracies of several methods based on MISR angular radiances and MODIS spectral radiances. The accuracy of cloud detections was evaluated relative to 5.086 million expert labels applied to 7.114 million 1.1-km resolution pixels with valid radiances from 57 scenes. The accuracy of the MODIS operational cloud mask was 90.72% for the 32 partly cloudy scenes and 93.37% for the 25 completely clear and overcast scenes. An automated, simple threshold algorithm based on three features extracted from MISR radiances and the MODIS operational cloud mask agreed with each other for 74.91% of the pixels in the 32 partly cloudy scenes and 78.44% of the pixels in the 25 completely clear and overcast scenes. These subsets of pixels had, relative to the expert labels, classification accuracies of 96.53% for the 32 partly cloudy scenes and 99.05% for the 25 completely clear and overcast scenes. Fisher's quadratic discriminate analysis (QDA) trained on expert labels from the 32 partly cloud scenes was applied to MISR radiances, three features based on MISR radiances, MODIS radiances, and the six features of the MODIS operational cloud mask with accuracies ranging from 87.51% to 96.43%. Accuracies increased to about 97% when QDA with expert labels was applied to combined radiances or features from both MISR and MODIS. Operational QDA-based classifiers were developed using as training labels those pixels for which the MISR automated, simple threshold and MODIS operational cloud mask algorithms agreed. Training the QDA classifier on these automatic labels using MISR radiances, three features based on MISR radiances, MODIS radiances, and the six features of the MODIS operational cloud mask led to accuracies ranging from 85.23% to 93.62% for the 32 partly cloudy scenes. Classification accuracies increased to 93.74% (93.40%) when combined MISR and MODIS radiances (features) were used. The highest accuracy attained with any operational algorithm tested on all 57 scenes was 94.51%. These results suggest that both MISR and MODIS radiances are sufficient for cloud detection in daytime polar regions.
AB - Expert labels were used to evaluate arctic cloud detection accuracies of several methods based on MISR angular radiances and MODIS spectral radiances. The accuracy of cloud detections was evaluated relative to 5.086 million expert labels applied to 7.114 million 1.1-km resolution pixels with valid radiances from 57 scenes. The accuracy of the MODIS operational cloud mask was 90.72% for the 32 partly cloudy scenes and 93.37% for the 25 completely clear and overcast scenes. An automated, simple threshold algorithm based on three features extracted from MISR radiances and the MODIS operational cloud mask agreed with each other for 74.91% of the pixels in the 32 partly cloudy scenes and 78.44% of the pixels in the 25 completely clear and overcast scenes. These subsets of pixels had, relative to the expert labels, classification accuracies of 96.53% for the 32 partly cloudy scenes and 99.05% for the 25 completely clear and overcast scenes. Fisher's quadratic discriminate analysis (QDA) trained on expert labels from the 32 partly cloud scenes was applied to MISR radiances, three features based on MISR radiances, MODIS radiances, and the six features of the MODIS operational cloud mask with accuracies ranging from 87.51% to 96.43%. Accuracies increased to about 97% when QDA with expert labels was applied to combined radiances or features from both MISR and MODIS. Operational QDA-based classifiers were developed using as training labels those pixels for which the MISR automated, simple threshold and MODIS operational cloud mask algorithms agreed. Training the QDA classifier on these automatic labels using MISR radiances, three features based on MISR radiances, MODIS radiances, and the six features of the MODIS operational cloud mask led to accuracies ranging from 85.23% to 93.62% for the 32 partly cloudy scenes. Classification accuracies increased to 93.74% (93.40%) when combined MISR and MODIS radiances (features) were used. The highest accuracy attained with any operational algorithm tested on all 57 scenes was 94.51%. These results suggest that both MISR and MODIS radiances are sufficient for cloud detection in daytime polar regions.
UR - http://www.scopus.com/inward/record.url?scp=33846988606&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=33846988606&partnerID=8YFLogxK
U2 - 10.1016/j.rse.2006.10.015
DO - 10.1016/j.rse.2006.10.015
M3 - Article
AN - SCOPUS:33846988606
SN - 0034-4257
VL - 107
SP - 172
EP - 184
JO - Remote Sensing of Environment
JF - Remote Sensing of Environment
IS - 1-2
ER -