Welcome to the new version of CaltechAUTHORS. Login is currently restricted to library staff. If you notice any issues, please email coda@library.caltech.edu
Published February 2017 | Supplemental Material
Journal Article Open

Improving the long-lead predictability of El Niño using a novel forecasting scheme based on a dynamic components model

Abstract

El Niño (EN) is a dominant feature of climate variability on inter-annual time scales driving changes in the climate throughout the globe, and having wide-spread natural and socio-economic consequences. In this sense, its forecast is an important task, and predictions are issued on a regular basis by a wide array of prediction schemes and climate centres around the world. This study explores a novel method for EN forecasting. In the state-of-the-art the advantageous statistical technique of unobserved components time series modeling, also known as structural time series modeling, has not been applied. Therefore, we have developed such a model where the statistical analysis, including parameter estimation and forecasting, is based on state space methods, and includes the celebrated Kalman filter. The distinguishing feature of this dynamic model is the decomposition of a time series into a range of stochastically time-varying components such as level (or trend), seasonal, cycles of different frequencies, irregular, and regression effects incorporated as explanatory covariates. These components are modeled separately and ultimately combined in a single forecasting scheme. Customary statistical models for EN prediction essentially use SST and wind stress in the equatorial Pacific. In addition to these, we introduce a new domain of regression variables accounting for the state of the subsurface ocean temperature in the western and central equatorial Pacific, motivated by our analysis, as well as by recent and classical research, showing that subsurface processes and heat accumulation there are fundamental for the genesis of EN. An important feature of the scheme is that different regression predictors are used at different lead months, thus capturing the dynamical evolution of the system and rendering more efficient forecasts. The new model has been tested with the prediction of all warm events that occurred in the period 1996–2015. Retrospective forecasts of these events were made for long lead times of at least two and a half years. Hence, the present study demonstrates that the theoretical limit of ENSO prediction should be sought much longer than the commonly accepted "Spring Barrier". The high correspondence between the forecasts and observations indicates that the proposed model outperforms all current operational statistical models, and behaves comparably to the best dynamical models used for EN prediction. Thus, the novel way in which the modeling scheme has been structured could also be used for improving other statistical and dynamical modeling systems.

Additional Information

© 2016 Springer-Verlag Berlin Heidelberg. Received: 17 November 2015; Accepted: 17 April 2016; Published online: 2 May 2016. J.B. gratefully acknowledges funding from the European Commission through a Marie Curie International Outgoing Fellowship (Project MEMENTO from the FP7-PEOPLE-2011-IOF call), and from the European Commission and the Catalan Government through a Marie Curie—Beatriu de Pinós Fellowship (Project 00068 from the BP-DGR-2014-B call). X.R. gratefully acknowledges funding from the Ministry of Science and Innovation, Spain (Project PANDORA CGL 2007-63053).

Attached Files

Supplemental Material - 382_2016_3139_MOESM1_ESM.pdf

Files

382_2016_3139_MOESM1_ESM.pdf
Files (257.5 kB)
Name Size Download all
md5:c2fd3d01f01f550462fffa2a9dc2f1c9
257.5 kB Preview Download

Additional details

Created:
August 22, 2023
Modified:
October 25, 2023