Analyzing El Niño–Southern Oscillation Predictability Using Long-Short-Term-Memory Models

Andrew Huang, Ben Vega-Westhoff, Ryan Sriver

Research output: Contribution to journalArticle

Abstract

El Niño–Southern Oscillation (ENSO) can have global impacts, affecting daily temperature and precipitation, and extreme weather, such as hurricanes and tornadoes. Because of its importance, scientists strive to understand the processes that govern ENSO and develop models to predict its evolution and changes in variability. Here long-short-term-memory models (LSTMs) were compared to linear regression models (LR) to explore the benefits of simple, deep neural networks in predicting ENSO, in addition to quantifying the relative importance of the sources of ENSO's predictability. The models use central Pacific sea surface temperatures (SST), equatorial Pacific warm water volumes, and western Pacific zonal winds as predictors, individually and in combinations, on monthly and daily resolutions, from 1- to 11-month leads. By using these predictors, many characteristic time scales are encompassed—from days-to-weeks in the atmosphere, to months-to-seasons in the coupled system, and interseasonal-to-interannual in the subsurface ocean. Results show, with monthly input, predictions from LSTM were like predictions from LR. However, with daily SST at longer leads, LSTM exhibited some advantage over LR in terms of the correlation coefficient. This suggests that daily SST may contain some nonlinear element that improves LSTM predictability compared to LR. In addition, this suggests that more information, such as gridded data and additional variables, would likely improve predictability using LSTM, but results would be more difficult to interpret. Overall, LSTM may be appealing because once the computationally expensive training of LSTM is complete, the predictions employing the trained model can be relatively cheap to perform thereafter.

Original languageEnglish (US)
Pages (from-to)212-221
Number of pages10
JournalEarth and Space Science
Volume6
Issue number2
DOIs
StatePublished - Feb 1 2019

Fingerprint

oscillation
sea surface temperature
prediction
tornado
zonal wind
warm water
hurricane
El Nino-Southern Oscillation
timescale
weather
atmosphere
ocean

Keywords

  • fuzzy logic
  • model verification and validation
  • modeling
  • neural networks

ASJC Scopus subject areas

  • Environmental Science (miscellaneous)
  • Earth and Planetary Sciences(all)

Cite this

Analyzing El Niño–Southern Oscillation Predictability Using Long-Short-Term-Memory Models. / Huang, Andrew; Vega-Westhoff, Ben; Sriver, Ryan.

In: Earth and Space Science, Vol. 6, No. 2, 01.02.2019, p. 212-221.

Research output: Contribution to journalArticle

@article{bfdac69a43bb4887bbceb4eda77c790f,
title = "Analyzing El Ni{\~n}o–Southern Oscillation Predictability Using Long-Short-Term-Memory Models",
abstract = "El Ni{\~n}o–Southern Oscillation (ENSO) can have global impacts, affecting daily temperature and precipitation, and extreme weather, such as hurricanes and tornadoes. Because of its importance, scientists strive to understand the processes that govern ENSO and develop models to predict its evolution and changes in variability. Here long-short-term-memory models (LSTMs) were compared to linear regression models (LR) to explore the benefits of simple, deep neural networks in predicting ENSO, in addition to quantifying the relative importance of the sources of ENSO's predictability. The models use central Pacific sea surface temperatures (SST), equatorial Pacific warm water volumes, and western Pacific zonal winds as predictors, individually and in combinations, on monthly and daily resolutions, from 1- to 11-month leads. By using these predictors, many characteristic time scales are encompassed—from days-to-weeks in the atmosphere, to months-to-seasons in the coupled system, and interseasonal-to-interannual in the subsurface ocean. Results show, with monthly input, predictions from LSTM were like predictions from LR. However, with daily SST at longer leads, LSTM exhibited some advantage over LR in terms of the correlation coefficient. This suggests that daily SST may contain some nonlinear element that improves LSTM predictability compared to LR. In addition, this suggests that more information, such as gridded data and additional variables, would likely improve predictability using LSTM, but results would be more difficult to interpret. Overall, LSTM may be appealing because once the computationally expensive training of LSTM is complete, the predictions employing the trained model can be relatively cheap to perform thereafter.",
keywords = "fuzzy logic, model verification and validation, modeling, neural networks",
author = "Andrew Huang and Ben Vega-Westhoff and Ryan Sriver",
year = "2019",
month = "2",
day = "1",
doi = "10.1029/2018EA000423",
language = "English (US)",
volume = "6",
pages = "212--221",
journal = "Earth and Space Science",
issn = "2333-5084",
publisher = "Wiley-Blackwell Publishing Ltd",
number = "2",

}

TY - JOUR

T1 - Analyzing El Niño–Southern Oscillation Predictability Using Long-Short-Term-Memory Models

AU - Huang, Andrew

AU - Vega-Westhoff, Ben

AU - Sriver, Ryan

PY - 2019/2/1

Y1 - 2019/2/1

N2 - El Niño–Southern Oscillation (ENSO) can have global impacts, affecting daily temperature and precipitation, and extreme weather, such as hurricanes and tornadoes. Because of its importance, scientists strive to understand the processes that govern ENSO and develop models to predict its evolution and changes in variability. Here long-short-term-memory models (LSTMs) were compared to linear regression models (LR) to explore the benefits of simple, deep neural networks in predicting ENSO, in addition to quantifying the relative importance of the sources of ENSO's predictability. The models use central Pacific sea surface temperatures (SST), equatorial Pacific warm water volumes, and western Pacific zonal winds as predictors, individually and in combinations, on monthly and daily resolutions, from 1- to 11-month leads. By using these predictors, many characteristic time scales are encompassed—from days-to-weeks in the atmosphere, to months-to-seasons in the coupled system, and interseasonal-to-interannual in the subsurface ocean. Results show, with monthly input, predictions from LSTM were like predictions from LR. However, with daily SST at longer leads, LSTM exhibited some advantage over LR in terms of the correlation coefficient. This suggests that daily SST may contain some nonlinear element that improves LSTM predictability compared to LR. In addition, this suggests that more information, such as gridded data and additional variables, would likely improve predictability using LSTM, but results would be more difficult to interpret. Overall, LSTM may be appealing because once the computationally expensive training of LSTM is complete, the predictions employing the trained model can be relatively cheap to perform thereafter.

AB - El Niño–Southern Oscillation (ENSO) can have global impacts, affecting daily temperature and precipitation, and extreme weather, such as hurricanes and tornadoes. Because of its importance, scientists strive to understand the processes that govern ENSO and develop models to predict its evolution and changes in variability. Here long-short-term-memory models (LSTMs) were compared to linear regression models (LR) to explore the benefits of simple, deep neural networks in predicting ENSO, in addition to quantifying the relative importance of the sources of ENSO's predictability. The models use central Pacific sea surface temperatures (SST), equatorial Pacific warm water volumes, and western Pacific zonal winds as predictors, individually and in combinations, on monthly and daily resolutions, from 1- to 11-month leads. By using these predictors, many characteristic time scales are encompassed—from days-to-weeks in the atmosphere, to months-to-seasons in the coupled system, and interseasonal-to-interannual in the subsurface ocean. Results show, with monthly input, predictions from LSTM were like predictions from LR. However, with daily SST at longer leads, LSTM exhibited some advantage over LR in terms of the correlation coefficient. This suggests that daily SST may contain some nonlinear element that improves LSTM predictability compared to LR. In addition, this suggests that more information, such as gridded data and additional variables, would likely improve predictability using LSTM, but results would be more difficult to interpret. Overall, LSTM may be appealing because once the computationally expensive training of LSTM is complete, the predictions employing the trained model can be relatively cheap to perform thereafter.

KW - fuzzy logic

KW - model verification and validation

KW - modeling

KW - neural networks

UR - http://www.scopus.com/inward/record.url?scp=85061842551&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85061842551&partnerID=8YFLogxK

U2 - 10.1029/2018EA000423

DO - 10.1029/2018EA000423

M3 - Article

AN - SCOPUS:85061842551

VL - 6

SP - 212

EP - 221

JO - Earth and Space Science

JF - Earth and Space Science

SN - 2333-5084

IS - 2

ER -