Data-driven prediction strategies for low-frequency patterns of North Pacific climate variability

Darin Comeau, Zhizhen Zhao, Dimitrios Giannakis, Andrew J. Majda

Research output: Contribution to journalArticle

Abstract

The North Pacific exhibits patterns of low-frequency variability on the intra-annual to decadal time scales, which manifest themselves in both model data and the observational record, and prediction of such low-frequency modes of variability is of great interest to the community. While parametric models, such as stationary and non-stationary autoregressive models, possibly including external factors, may perform well in a data-fitting setting, they may perform poorly in a prediction setting. Ensemble analog forecasting, which relies on the historical record to provide estimates of the future based on past trajectories of those states similar to the initial state of interest, provides a promising, nonparametric approach to forecasting that makes no assumptions on the underlying dynamics or its statistics. We apply such forecasting to low-frequency modes of variability for the North Pacific sea surface temperature and sea ice concentration fields extracted through Nonlinear Laplacian Spectral Analysis, an algorithm which produces clean time-scale separation from data without pre-filtering. We find such methods may outperform parametric methods and simple persistence with increased predictive skill, and are more skillful when initialized in an active phase, rather than a quiescent phase. We also apply these methods to the predict integrated sea ice extent anomalies in the North Pacific from both models and observations, and find an increase of predictive skill over the persistence forecast by about 2 months.

Original languageEnglish (US)
Pages (from-to)1-18
Number of pages18
JournalClimate Dynamics
DOIs
StateAccepted/In press - May 27 2016

Fingerprint

climate
prediction
sea ice
persistence
timescale
historical record
spectral analysis
sea surface temperature
trajectory
anomaly
method
statistics
forecast

Keywords

  • Dimension reduction
  • North Pacific climate variability
  • PDO
  • Predictability

ASJC Scopus subject areas

  • Atmospheric Science

Cite this

Data-driven prediction strategies for low-frequency patterns of North Pacific climate variability. / Comeau, Darin; Zhao, Zhizhen; Giannakis, Dimitrios; Majda, Andrew J.

In: Climate Dynamics, 27.05.2016, p. 1-18.

Research output: Contribution to journalArticle

@article{f7a00316f4ed4de7b203a624f8b28e41,
title = "Data-driven prediction strategies for low-frequency patterns of North Pacific climate variability",
abstract = "The North Pacific exhibits patterns of low-frequency variability on the intra-annual to decadal time scales, which manifest themselves in both model data and the observational record, and prediction of such low-frequency modes of variability is of great interest to the community. While parametric models, such as stationary and non-stationary autoregressive models, possibly including external factors, may perform well in a data-fitting setting, they may perform poorly in a prediction setting. Ensemble analog forecasting, which relies on the historical record to provide estimates of the future based on past trajectories of those states similar to the initial state of interest, provides a promising, nonparametric approach to forecasting that makes no assumptions on the underlying dynamics or its statistics. We apply such forecasting to low-frequency modes of variability for the North Pacific sea surface temperature and sea ice concentration fields extracted through Nonlinear Laplacian Spectral Analysis, an algorithm which produces clean time-scale separation from data without pre-filtering. We find such methods may outperform parametric methods and simple persistence with increased predictive skill, and are more skillful when initialized in an active phase, rather than a quiescent phase. We also apply these methods to the predict integrated sea ice extent anomalies in the North Pacific from both models and observations, and find an increase of predictive skill over the persistence forecast by about 2 months.",
keywords = "Dimension reduction, North Pacific climate variability, PDO, Predictability",
author = "Darin Comeau and Zhizhen Zhao and Dimitrios Giannakis and Majda, {Andrew J.}",
year = "2016",
month = "5",
day = "27",
doi = "10.1007/s00382-016-3177-5",
language = "English (US)",
pages = "1--18",
journal = "Climate Dynamics",
issn = "0930-7575",
publisher = "Springer Verlag",

}

TY - JOUR

T1 - Data-driven prediction strategies for low-frequency patterns of North Pacific climate variability

AU - Comeau, Darin

AU - Zhao, Zhizhen

AU - Giannakis, Dimitrios

AU - Majda, Andrew J.

PY - 2016/5/27

Y1 - 2016/5/27

N2 - The North Pacific exhibits patterns of low-frequency variability on the intra-annual to decadal time scales, which manifest themselves in both model data and the observational record, and prediction of such low-frequency modes of variability is of great interest to the community. While parametric models, such as stationary and non-stationary autoregressive models, possibly including external factors, may perform well in a data-fitting setting, they may perform poorly in a prediction setting. Ensemble analog forecasting, which relies on the historical record to provide estimates of the future based on past trajectories of those states similar to the initial state of interest, provides a promising, nonparametric approach to forecasting that makes no assumptions on the underlying dynamics or its statistics. We apply such forecasting to low-frequency modes of variability for the North Pacific sea surface temperature and sea ice concentration fields extracted through Nonlinear Laplacian Spectral Analysis, an algorithm which produces clean time-scale separation from data without pre-filtering. We find such methods may outperform parametric methods and simple persistence with increased predictive skill, and are more skillful when initialized in an active phase, rather than a quiescent phase. We also apply these methods to the predict integrated sea ice extent anomalies in the North Pacific from both models and observations, and find an increase of predictive skill over the persistence forecast by about 2 months.

AB - The North Pacific exhibits patterns of low-frequency variability on the intra-annual to decadal time scales, which manifest themselves in both model data and the observational record, and prediction of such low-frequency modes of variability is of great interest to the community. While parametric models, such as stationary and non-stationary autoregressive models, possibly including external factors, may perform well in a data-fitting setting, they may perform poorly in a prediction setting. Ensemble analog forecasting, which relies on the historical record to provide estimates of the future based on past trajectories of those states similar to the initial state of interest, provides a promising, nonparametric approach to forecasting that makes no assumptions on the underlying dynamics or its statistics. We apply such forecasting to low-frequency modes of variability for the North Pacific sea surface temperature and sea ice concentration fields extracted through Nonlinear Laplacian Spectral Analysis, an algorithm which produces clean time-scale separation from data without pre-filtering. We find such methods may outperform parametric methods and simple persistence with increased predictive skill, and are more skillful when initialized in an active phase, rather than a quiescent phase. We also apply these methods to the predict integrated sea ice extent anomalies in the North Pacific from both models and observations, and find an increase of predictive skill over the persistence forecast by about 2 months.

KW - Dimension reduction

KW - North Pacific climate variability

KW - PDO

KW - Predictability

UR - http://www.scopus.com/inward/record.url?scp=84971001488&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84971001488&partnerID=8YFLogxK

U2 - 10.1007/s00382-016-3177-5

DO - 10.1007/s00382-016-3177-5

M3 - Article

AN - SCOPUS:84971001488

SP - 1

EP - 18

JO - Climate Dynamics

JF - Climate Dynamics

SN - 0930-7575

ER -