Show simple item record

Linearization of the Principal Component Analysis method for radiative transfer acceleration: Application to retrieval algorithms and sensitivity studies

dc.contributor.authorSpurr, R.
dc.contributor.authorNatraj, V.
dc.contributor.authorLerot, C.
dc.contributor.authorVan Roozendael, M.
dc.contributor.authorLoyola, D.
dc.date2013
dc.date.accessioned2016-03-25T11:11:12Z
dc.date.available2016-03-25T11:11:12Z
dc.identifier.urihttps://orfeo.belnet.be/handle/internal/2946
dc.descriptionPrincipal Component Analysis (PCA) is a promising tool for enhancing radiative transfer (RT) performance. When applied to binned optical property data sets, PCA exploits redundancy in the optical data, and restricts the number of full multiple-scatter calculations to those optical states corresponding to the most important principal components, yet still maintaining high accuracy in the radiance approximations. We show that the entire PCA RT enhancement process is analytically differentiable with respect to any atmospheric or surface parameter, thus allowing for accurate and fast approximations of Jacobian matrices, in addition to radiances. This linearization greatly extends the power and scope of the PCA method to many remote sensing retrieval applications and sensitivity studies. In the first example, we examine accuracy for PCA-derived UV-backscatter radiance and Jacobian fields over a 290-340. nm window. In a second application, we show that performance for UV-based total ozone column retrieval is considerably improved without compromising the accuracy.
dc.languageeng
dc.titleLinearization of the Principal Component Analysis method for radiative transfer acceleration: Application to retrieval algorithms and sensitivity studies
dc.typeArticle
dc.subject.frascatiEarth and related Environmental sciences
dc.audienceScientific
dc.subject.freeFast approximation
dc.subject.freePerformance enhancements
dc.subject.freePrincipal component analysis method
dc.subject.freePrincipal Components
dc.subject.freeRetrieval algorithms
dc.subject.freeRetrieval applications
dc.subject.freeSensitivity studies
dc.subject.freeTotal ozone column
dc.subject.freeJacobian matrices
dc.subject.freeLinearization
dc.subject.freeOptical properties
dc.subject.freePrincipal component analysis
dc.subject.freeRadiative transfer
dc.subject.freeData handling
dc.subject.freeaccuracy assessment
dc.subject.freealgorithm
dc.subject.freebackscatter
dc.subject.freeoptical property
dc.subject.freeozone
dc.subject.freeprincipal component analysis
dc.subject.freeradiative transfer
dc.subject.freeremote sensing
dc.subject.freesensitivity analysis
dc.source.titleJournal of Quantitative Spectroscopy and Radiative Transfer
dc.source.volume125
dc.source.page1-17
Orfeo.peerreviewedYes
dc.identifier.doi10.1016/j.jqsrt.2013.04.002
dc.identifier.scopus2-s2.0-84878592868


Files in this item

Thumbnail
Name:
Spurr(2013).pdf
Size:
1.817Mb
Format:
PDF
View/Open

This item appears in the following Collection(s)

Show simple item record