The Ocean's Palette: Why El Niño Comes in Different Shades ✛

Original: "On the archetypal `flavours', indices and teleconnections of ENSO revealed by global sea surface temperatures"
arXiv:2406.08694 · 2024-06-12 · CC BY 4.0 · 1 min
El Niño has different flavors, and a new method helps tell them apart.

An artist painting the ocean doesn't settle for a single blue. There's ultramarine, azure, turquoise: many shades. Likewise, El Niño, the warm current off Peru, is not always the same. Sometimes warm water gathers in the central Pacific, sometimes in the east, and that affects rains and droughts in different countries.

Previously, scientists often divided El Niño into two types: eastern and central. But that's not enough. In the new study, they applied a method that looks not for average states but for the most extreme ones, like pure colors on a palette. These extreme states are called archetypes. Taking ocean temperature maps over 41 years, the algorithm found four main archetypes: warm and cold variants in the center and in the east.

When the number of archetypes was increased to eight, additional shades emerged, such as the onset and decay of El Niño. And sea level turned out to be an even better indicator than surface temperature.

Now, instead of a simple "El Niño or La Niña," we can talk in terms of probabilities: how much the current state resembles one type or another. This helps better predict weather and prepare for its consequences. The ocean's palette has become richer, and the artist can paint the climate picture more precisely.

One of the largest eastern Pacific El Niño events, in 1982–1983, was so unusual that traditional monitoring systems initially missed it, because the warm anomalies appeared where they were not expected.

\arg\min_{C,S} \|X - XCS\|_F^2
Find matrices C and S with nonnegative elements and row sums equal to one, such that the product XCS best approximates the data X.
\Delta_{i_{max}}^p(t) = 1 - \frac{1}{p-1} \left( \frac{1}{S_{i_{max} t}} - 1 \right)
The score is 1 if at a given time one archetype has probability 1 (pure state), and 0 if all p archetypes are equally probable (neutral or mixed state).
F^S_{s \times p} = \frac{\sum_t F_{st} S^T_{pt}}{\sum_t S_{pt}}
Time-weighted average of field F, where weights are the membership probabilities of each time point to archetype p from matrix S.
Scientists
Emile ClapeyronDavid CharbonneauR. A. RossiterD. B. McLaughlinMichel MayorDidier Queloz
Concepts
ENSO El Niño index Jin-Timmermann model El Niño Multivariate ENSO Index La Niña ERSST relative ONI Southern Oscillation NOAA CPC Teleconnection NOAA PMEL / TAO TAO/TRITON moorings spring predictability barrier zebiak cane model equatorial wave propagation OISST v2.1 Pacific decadal oscillation sea surface temperature anomaly UAH satellite record