Premium
A comparision of low‐dimensional representations of sea‐surface temperature anomalies in the North Atlantic
Author(s) -
Thacker W. C.,
Lewandowicz R.
Publication year - 1997
Publication title -
international journal of climatology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.58
H-Index - 166
eISSN - 1097-0088
pISSN - 0899-8418
DOI - 10.1002/(sici)1097-0088(199707)17:9<953::aid-joc169>3.0.co;2-7
Subject(s) - principal component analysis , longitude , climatology , latitude , geographic coordinate system , covariance matrix , geology , sea surface temperature , geodesy , mathematics , statistics
Correlation‐matrix principal components of North Atlantic sea‐surface temperature anomalies for the interval 1950–1970 account for the anomalous variability observed during the interval 1972– 1992 better than do similar numbers of covariance‐matrix principal components, regional averages, or carefully selected local anomalies. When drift in the seasonal cycle is taken into account, local anomalies for 127 6°×4° longitude‐by‐latitude North Atlantic cells could be recovered with an average skill as high as 0·79. Surprisingly, skill increased monotonically with increasing numbers of principal components, and the maximum value was not obtained until 62 were used to characterize the field. Clearly, far more principal components carry useful information about local details than has been previously suspected. © 1997 the Royal Meteorological Society.