Skill comparisons between neural networks and canonical correlation analysis in predicting the equatorial Pacific sea surface temperatures

Benyang Tang, William W. Hsieh, Adam H. Monahan, Fredolin Tangang @ Tajudin Mahmud

Research output: Contribution to journalArticle

53 Citations (Scopus)

Abstract

Among the statistical methods used for seasonal climate prediction, canonical correlation analysis (CCA), a more sophisticated version of the linear regression (LR) method, is well established. Recently, neural networks (NN) have been applied to seasonal climate prediction. Unlike CCA and LR, NN is a nonlinear method, which leads to the question whether the nonlinearity of NN brings any extra prediction skill. In this study, an objective comparison between the three methods (CCA, LR, and NN) in predicting the equatorial Pacific sea surface temperatures (in regions Nino 1+2, Nino3, Nino3.4, and Nino4) was made. The skill of NN was found to be comparable to that of LR and CCA. A cross-validated t test showed that the difference between NN and LR and the difference between NN and CCA were not significant at the 5% level. The lack of significant skill difference between the nonlinear NN method and the linear methods suggests that at the seasonal timescale the equatorial Pacific dynamics is basically linear.

Original languageEnglish
Pages (from-to)287-293
Number of pages7
JournalJournal of Climate
Volume13
Issue number1
Publication statusPublished - 1 Jan 2000
Externally publishedYes

Fingerprint

sea surface temperature
climate prediction
nonlinearity
analysis
comparison
method
timescale
prediction

ASJC Scopus subject areas

  • Atmospheric Science

Cite this

Skill comparisons between neural networks and canonical correlation analysis in predicting the equatorial Pacific sea surface temperatures. / Tang, Benyang; Hsieh, William W.; Monahan, Adam H.; Tangang @ Tajudin Mahmud, Fredolin.

In: Journal of Climate, Vol. 13, No. 1, 01.01.2000, p. 287-293.

Research output: Contribution to journalArticle

@article{3bc479c392274791883be500b4328320,
title = "Skill comparisons between neural networks and canonical correlation analysis in predicting the equatorial Pacific sea surface temperatures",
abstract = "Among the statistical methods used for seasonal climate prediction, canonical correlation analysis (CCA), a more sophisticated version of the linear regression (LR) method, is well established. Recently, neural networks (NN) have been applied to seasonal climate prediction. Unlike CCA and LR, NN is a nonlinear method, which leads to the question whether the nonlinearity of NN brings any extra prediction skill. In this study, an objective comparison between the three methods (CCA, LR, and NN) in predicting the equatorial Pacific sea surface temperatures (in regions Nino 1+2, Nino3, Nino3.4, and Nino4) was made. The skill of NN was found to be comparable to that of LR and CCA. A cross-validated t test showed that the difference between NN and LR and the difference between NN and CCA were not significant at the 5{\%} level. The lack of significant skill difference between the nonlinear NN method and the linear methods suggests that at the seasonal timescale the equatorial Pacific dynamics is basically linear.",
author = "Benyang Tang and Hsieh, {William W.} and Monahan, {Adam H.} and {Tangang @ Tajudin Mahmud}, Fredolin",
year = "2000",
month = "1",
day = "1",
language = "English",
volume = "13",
pages = "287--293",
journal = "Journal of Climate",
issn = "0894-8755",
publisher = "American Meteorological Society",
number = "1",

}

TY - JOUR

T1 - Skill comparisons between neural networks and canonical correlation analysis in predicting the equatorial Pacific sea surface temperatures

AU - Tang, Benyang

AU - Hsieh, William W.

AU - Monahan, Adam H.

AU - Tangang @ Tajudin Mahmud, Fredolin

PY - 2000/1/1

Y1 - 2000/1/1

N2 - Among the statistical methods used for seasonal climate prediction, canonical correlation analysis (CCA), a more sophisticated version of the linear regression (LR) method, is well established. Recently, neural networks (NN) have been applied to seasonal climate prediction. Unlike CCA and LR, NN is a nonlinear method, which leads to the question whether the nonlinearity of NN brings any extra prediction skill. In this study, an objective comparison between the three methods (CCA, LR, and NN) in predicting the equatorial Pacific sea surface temperatures (in regions Nino 1+2, Nino3, Nino3.4, and Nino4) was made. The skill of NN was found to be comparable to that of LR and CCA. A cross-validated t test showed that the difference between NN and LR and the difference between NN and CCA were not significant at the 5% level. The lack of significant skill difference between the nonlinear NN method and the linear methods suggests that at the seasonal timescale the equatorial Pacific dynamics is basically linear.

AB - Among the statistical methods used for seasonal climate prediction, canonical correlation analysis (CCA), a more sophisticated version of the linear regression (LR) method, is well established. Recently, neural networks (NN) have been applied to seasonal climate prediction. Unlike CCA and LR, NN is a nonlinear method, which leads to the question whether the nonlinearity of NN brings any extra prediction skill. In this study, an objective comparison between the three methods (CCA, LR, and NN) in predicting the equatorial Pacific sea surface temperatures (in regions Nino 1+2, Nino3, Nino3.4, and Nino4) was made. The skill of NN was found to be comparable to that of LR and CCA. A cross-validated t test showed that the difference between NN and LR and the difference between NN and CCA were not significant at the 5% level. The lack of significant skill difference between the nonlinear NN method and the linear methods suggests that at the seasonal timescale the equatorial Pacific dynamics is basically linear.

UR - http://www.scopus.com/inward/record.url?scp=0034035880&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0034035880&partnerID=8YFLogxK

M3 - Article

VL - 13

SP - 287

EP - 293

JO - Journal of Climate

JF - Journal of Climate

SN - 0894-8755

IS - 1

ER -