Using network theory and machine learning to predict El Niño

Abstract. The skill of current predictions of the warm phase of the El Niño Southern Oscillation (ENSO) reduces significantly beyond a lag time of 6 months. In this paper, we aim to increase this prediction skill at lag times of up to 1 year. The new method combines a classical autoregressive integrated moving average technique with a modern machine learning approach (through an artificial neural network). The attributes in such a neural network are derived from knowledge of physical processes and topological properties of climate networks, and they are tested using a Zebiak–Cane-type model and observations. For predictions up to 6 months ahead, the results of the hybrid model give a slightly better skill than the CFSv2 ensemble prediction by the National Centers for Environmental Prediction (NCEP). Interestingly, results for a 12-month lead time prediction have a similar skill as the shorter lead time predictions.

[1]  T. Evgeniou,et al.  To combine or not to combine: selecting among forecasts and their combinations , 2005 .

[2]  S. Havlin,et al.  Stability of Climate Networks with Time , 2011, Scientific Reports.

[3]  Andrew M. Moore,et al.  Stochastic forcing of ENSO by the intraseasonal oscillation , 1999 .

[4]  H. Dijkstra,et al.  Cold tongue/Warm pool and ENSO dynamics in the Pliocene , 2011 .

[5]  Constantine Dovrolis,et al.  ENSO in CMIP5 simulations: network connectivity from the recent past to the twenty-third century , 2014, Climate Dynamics.

[6]  Mehdi Khashei,et al.  A novel hybridization of artificial neural networks and ARIMA models for time series forecasting , 2011, Appl. Soft Comput..

[7]  J. Ignacio Deza,et al.  Distinguishing the effects of internal and forced atmospheric variability in climate networks , 2013, 1311.3089.

[8]  William W. Hsieh,et al.  Neural network forecasts of the tropical Pacific sea surface temperatures , 2006, Neural Networks.

[9]  J. Rissanen,et al.  Modeling By Shortest Data Description* , 1978, Autom..

[10]  A. J. Clarke,et al.  On the Warm Water Volume and Its Changing Relationship with ENSO , 2014 .

[11]  M. Latif,et al.  The Response of a Coupled Ocean-Atmosphere General Circulation Model to Wind Bursts. , 1988 .

[12]  B. Kirtman,et al.  El Niño in a changing climate , 2009, Nature.

[13]  Guoqiang Peter Zhang,et al.  Time series forecasting using a hybrid ARIMA and neural network model , 2003, Neurocomputing.

[14]  Çagdas Hakan Aladag,et al.  Forecasting nonlinear time series with a hybrid methodology , 2009, Appl. Math. Lett..

[15]  Elizabeth C. Kent,et al.  Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century , 2003 .

[16]  José Manuel Benítez,et al.  On the use of cross-validation for time series predictor evaluation , 2012, Inf. Sci..

[17]  Alexey Kaplan,et al.  Predictability of El Niño over the past 148 years , 2004, Nature.

[18]  F. Jin,et al.  The Pacific Cold Tongue and the ENSO mode: a unified theory within the Zebiak-Cane model , 2000 .

[19]  T. Delcroix,et al.  Observed equatorial Rossby waves and ENSO-related warm water volume changes in the equatorial Pacific Ocean - art. no. C06003 , 2008 .

[20]  Coherent Tropical Indo-Pacific Interannual Climate Variability , 2016 .

[21]  Isabelle Guyon,et al.  An Introduction to Variable and Feature Selection , 2003, J. Mach. Learn. Res..

[22]  Andrew T. Wittenberg,et al.  How Predictable is El Niño , 2003 .

[23]  Paul J. Roebber,et al.  What Do Networks Have to Do with Climate , 2006 .

[24]  Ping-Feng Pai,et al.  A hybrid ARIMA and support vector machines model in stock price forecasting , 2005 .

[25]  Thomas M. Smith,et al.  Extended Reconstructed Sea Surface Temperature Version 4 (ERSST.v4). Part I: Upgrades and Intercomparisons , 2014 .

[26]  A. Al-Zaben,et al.  ARMA Model Order Determination Using Edge Detection: A New Perspective , 2005 .

[27]  Mark Newman,et al.  Networks: An Introduction , 2010 .

[28]  Yang Wang,et al.  Oceanic El-Niño wave dynamics and climate networks , 2015, 1505.07220.

[29]  H. Akaike.  A new look at the statistical model identification , 1974 .

[30]  Henk A. Dijkstra,et al.  The ENSO phenomenon: theory and mechanisms , 2006 .

[31]  Q. Feng.  A Complex Network Approach to Understand Climate Variability , 2015 .

[32]  Michael Ghil,et al.  El Ni�o on the Devil's Staircase: Annual Subharmonic Steps to Chaos , 1994, Science.

[33]  C. Ropelewski,et al.  Current approaches to seasonal to interannual climate predictions , 2001 .

[34]  W. Drosdowsky.  Statistical prediction of ENSO (Nino 3) using sub‐surface temperature data , 2006 .

[35]  P. R. Julian,et al.  Observations of the 40-50-day tropical oscillation - a review , 1994 .

[36]  J. Bjerknes.  ATMOSPHERIC TELECONNECTIONS FROM THE EQUATORIAL PACIFIC1 , 1969 .

[37]  S. Havlin,et al.  Emergence of El Niño as an autonomous component in the climate network. , 2010, Physical review letters.

[38]  Yang Wang,et al.  ClimateLearn: A machine-learning approach for climate prediction using network measures , 2016 .

[39]  Jean-René Donguy,et al.  Relations Between Sea Level, Thermocline Depth, Heat Content, and Dynamic Height in the Tropical Pacific Ocean , 1985 .

[40]  Eli Tziperman,et al.  El Ni�o Chaos: Overlapping of Resonances Between the Seasonal Cycle and the Pacific Ocean-Atmosphere Oscillator , 1994, Science.

[41]  A. E. Gill.  Some simple solutions for heat‐induced tropical circulation , 1980 .

[42]  Shlomo Havlin,et al.  Percolation framework to describe El Niño conditions. , 2016, Chaos.

[43]  Mark A. Hall,et al.  Correlation-based Feature Selection for Machine Learning , 2003 .

[44]  Fei-Fei Jin,et al.  An Equatorial Ocean Recharge Paradigm for ENSO. Part II: A Stripped-Down Coupled Model , 1997 .

[45]  Jürgen Kurths,et al.  Topology and seasonal evolution of the network of extreme precipitation over the Indian subcontinent and Sri Lanka , 2014 .

[46]  L. Goddarda,et al.  CURRENT APPROACHES TO SEASONAL-TO-INTERANNUAL CLIMATE PREDICTIONS , 2000 .

[47]  S. Havlin,et al.  Pattern of climate network blinking links follows El Niño events , 2008 .

[48]  Emilio Hernández-García,et al.  Percolation-based precursors of transitions in extended systems , 2016, Scientific Reports.

[49]  Héctor Pomares,et al.  Hybridization of intelligent techniques and ARIMA models for time series prediction , 2008, Fuzzy Sets Syst..

[50]  Qing Yi Feng,et al.  Climate network stability measures of El Niño variability. , 2017, Chaos.

[51]  K. P. Singh,et al.  Support vector machines in water quality management. , 2011, Analytica chimica acta.

[52]  Michael R. Hush,et al.  Machine learning for quantum physics , 2017, Science.

[53]  Jiuyong Li,et al.  Using causal discovery for feature selection in multivariate numerical time series , 2015, Machine Learning.

[54]  Demis Hassabis,et al.  Mastering the game of Go with deep neural networks and tree search , 2016, Nature.

[55]  Shlomo Havlin,et al.  Very early warning of next El Niño , 2014, Proceedings of the National Academy of Sciences.

[56]  Norbert Marwan,et al.  Characterizing the evolution of climate networks , 2014 .

[57]  Nitesh V. Chawla,et al.  Multivariate and multiscale dependence in the global climate system revealed through complex networks , 2012, Climate Dynamics.