Recent forecast skill for the El Niño/Southern Oscillation

Abstract

Abstract. We outline a relationship between three slowly varying characteristics of the coupled ocean-atmosphere system in the tropical Pacific: (i) quasi-periodicity, (ii) extended predictability, and (iii) approximate low dimensionality. The Southern Oscillation Index (SOI) and Nifio-3 sea surface temperatures characterize climatic variations in the tropical Pacific; these two time series are usually anticorrelated. This low-dimensional characterization suggests that much of the system's seasonal-to-interannual predictability depends on the regular behavior of the two scalar time series under consideration. The predictive skill of two idealized models is studied, showing the strong connection between regularity and predictability. E1-Nifio/Southern-Oscillation (ENSO) predictability is then assessed for current forecast models. When the periodic component of the ENSO signal is strong, it results in higher forecast skill. This skill decreases when the anticorrelation between SOI and Nifio-3 temperature anomalies is lost, as it has been in the first half of this decade.

Extreme events: dynamics, statistics and prediction

“…As shown by Ghil and Jiang (1998), these predictions are quite competitive with, or even better than other statistical methods in operational use, in the standard range of 6-12 months ahead. The same is true when comparing them to methods based on the use of deterministic models, including highly sophisticated ones, up to and including coupled GCMs.…”

A chaotically driven model climate: extreme events and snapshot attractors

“…The coupling of low-order models has already been used in the climate context, see e.g. the works of Palmer (1993) and Ghil and Jiang (1998). We focus on changes in the variability of the system, relative to either the undriven or the noisy case.…”

Relationships between the El-Niño Southern Oscillation and spate flows in southern Africa and Australia

“…With streamflow representing the complex integration of climatic and landscape responses that arguable amplify ENSO effects, a practical method for forecasting available water resources is to provide hydrological models with ENSO input. Ghil and Jiang (1998) showed that recent progress in understanding the dynamics of ENSO has produced techniques that have useful forecasting skill over periods of up to 12 months. ENSO prediction techniques include dynamical approaches such as the use of general circulation models (GCMs) and purely statistical approaches such as the use of canonical correlation analysis (Barnston and Ropelewski, 1992).…”

“…ENSO prediction techniques include dynamical approaches such as the use of general circulation models (GCMs) and purely statistical approaches such as the use of canonical correlation analysis (Barnston and Ropelewski, 1992). Ghil and Jiang (1998) showed the predictive skill for both dynamical and statistical approaches to be comparable, and forecasts of SSTs over the NINO3 regions can achieve a correlation skill against observed anomalies of up to 0.75 (Barnston et al, 1994). More recently, Berliner et al (2000) used Bayesian dynamic modelling for predictions of Pacific SSTs at lead times of up to seven months, achieving comparable results to larger deterministic models.…”