Predicting regional and pan-Arctic sea ice anomalies with kernel analog forecasting

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

Research output: Contribution to journalArticle

Abstract

Predicting Arctic sea ice extent is a notoriously difficult forecasting problem, even for lead times as short as one month. Motivated by Arctic intraannual variability phenomena such as reemergence of sea surface temperature and sea ice anomalies, we use a prediction approach for sea ice anomalies based on analog forecasting. Traditional analog forecasting relies on identifying a single analog in a historical record, usually by minimizing Euclidean distance, and forming a forecast from the analog’s historical trajectory. Here an ensemble of analogs is used to make forecasts, where the ensemble weights are determined by a dynamics-adapted similarity kernel, which takes into account the nonlinear geometry on the underlying data manifold. We apply this method for forecasting pan-Arctic and regional sea ice area and volume anomalies from multi-century climate model data, and in many cases find improvement over the benchmark damped persistence forecast. Examples of success include the 3–6 month lead time prediction of Arctic sea ice area, the winter sea ice area prediction of some marginal ice zone seas, and the 3–12 month lead time prediction of sea ice volume anomalies in many central Arctic basins. We discuss possible connections between KAF success and sea ice reemergence, and find KAF to be successful in regions and seasons exhibiting high interannual variability.

Original languageEnglish (US)
Pages (from-to)5507-5525
Number of pages19
JournalClimate Dynamics
Volume52
Issue number9-10
DOIs
StatePublished - May 1 2019

Fingerprint

sea ice
anomaly
prediction
marginal ice zone
historical record
climate modeling
sea surface temperature
persistence
trajectory
geometry
winter
basin
forecast

ASJC Scopus subject areas

  • Atmospheric Science

Cite this

Predicting regional and pan-Arctic sea ice anomalies with kernel analog forecasting. / Comeau, Darin; Giannakis, Dimitrios; Zhao, Zhizhen; Majda, Andrew J.

In: Climate Dynamics, Vol. 52, No. 9-10, 01.05.2019, p. 5507-5525.

Research output: Contribution to journalArticle

Comeau, Darin ; Giannakis, Dimitrios ; Zhao, Zhizhen ; Majda, Andrew J. / Predicting regional and pan-Arctic sea ice anomalies with kernel analog forecasting. In: Climate Dynamics. 2019 ; Vol. 52, No. 9-10. pp. 5507-5525.
@article{ec37e692f25447f8bf2a0f925da1b869,
title = "Predicting regional and pan-Arctic sea ice anomalies with kernel analog forecasting",
abstract = "Predicting Arctic sea ice extent is a notoriously difficult forecasting problem, even for lead times as short as one month. Motivated by Arctic intraannual variability phenomena such as reemergence of sea surface temperature and sea ice anomalies, we use a prediction approach for sea ice anomalies based on analog forecasting. Traditional analog forecasting relies on identifying a single analog in a historical record, usually by minimizing Euclidean distance, and forming a forecast from the analog’s historical trajectory. Here an ensemble of analogs is used to make forecasts, where the ensemble weights are determined by a dynamics-adapted similarity kernel, which takes into account the nonlinear geometry on the underlying data manifold. We apply this method for forecasting pan-Arctic and regional sea ice area and volume anomalies from multi-century climate model data, and in many cases find improvement over the benchmark damped persistence forecast. Examples of success include the 3–6 month lead time prediction of Arctic sea ice area, the winter sea ice area prediction of some marginal ice zone seas, and the 3–12 month lead time prediction of sea ice volume anomalies in many central Arctic basins. We discuss possible connections between KAF success and sea ice reemergence, and find KAF to be successful in regions and seasons exhibiting high interannual variability.",
author = "Darin Comeau and Dimitrios Giannakis and Zhizhen Zhao and Majda, {Andrew J.}",
year = "2019",
month = "5",
day = "1",
doi = "10.1007/s00382-018-4459-x",
language = "English (US)",
volume = "52",
pages = "5507--5525",
journal = "Climate Dynamics",
issn = "0930-7575",
publisher = "Springer Verlag",
number = "9-10",

}

TY - JOUR

T1 - Predicting regional and pan-Arctic sea ice anomalies with kernel analog forecasting

AU - Comeau, Darin

AU - Giannakis, Dimitrios

AU - Zhao, Zhizhen

AU - Majda, Andrew J.

PY - 2019/5/1

Y1 - 2019/5/1

N2 - Predicting Arctic sea ice extent is a notoriously difficult forecasting problem, even for lead times as short as one month. Motivated by Arctic intraannual variability phenomena such as reemergence of sea surface temperature and sea ice anomalies, we use a prediction approach for sea ice anomalies based on analog forecasting. Traditional analog forecasting relies on identifying a single analog in a historical record, usually by minimizing Euclidean distance, and forming a forecast from the analog’s historical trajectory. Here an ensemble of analogs is used to make forecasts, where the ensemble weights are determined by a dynamics-adapted similarity kernel, which takes into account the nonlinear geometry on the underlying data manifold. We apply this method for forecasting pan-Arctic and regional sea ice area and volume anomalies from multi-century climate model data, and in many cases find improvement over the benchmark damped persistence forecast. Examples of success include the 3–6 month lead time prediction of Arctic sea ice area, the winter sea ice area prediction of some marginal ice zone seas, and the 3–12 month lead time prediction of sea ice volume anomalies in many central Arctic basins. We discuss possible connections between KAF success and sea ice reemergence, and find KAF to be successful in regions and seasons exhibiting high interannual variability.

AB - Predicting Arctic sea ice extent is a notoriously difficult forecasting problem, even for lead times as short as one month. Motivated by Arctic intraannual variability phenomena such as reemergence of sea surface temperature and sea ice anomalies, we use a prediction approach for sea ice anomalies based on analog forecasting. Traditional analog forecasting relies on identifying a single analog in a historical record, usually by minimizing Euclidean distance, and forming a forecast from the analog’s historical trajectory. Here an ensemble of analogs is used to make forecasts, where the ensemble weights are determined by a dynamics-adapted similarity kernel, which takes into account the nonlinear geometry on the underlying data manifold. We apply this method for forecasting pan-Arctic and regional sea ice area and volume anomalies from multi-century climate model data, and in many cases find improvement over the benchmark damped persistence forecast. Examples of success include the 3–6 month lead time prediction of Arctic sea ice area, the winter sea ice area prediction of some marginal ice zone seas, and the 3–12 month lead time prediction of sea ice volume anomalies in many central Arctic basins. We discuss possible connections between KAF success and sea ice reemergence, and find KAF to be successful in regions and seasons exhibiting high interannual variability.

UR - http://www.scopus.com/inward/record.url?scp=85053878478&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85053878478&partnerID=8YFLogxK

U2 - 10.1007/s00382-018-4459-x

DO - 10.1007/s00382-018-4459-x

M3 - Article

AN - SCOPUS:85053878478

VL - 52

SP - 5507

EP - 5525

JO - Climate Dynamics

JF - Climate Dynamics

SN - 0930-7575

IS - 9-10

ER -