El Niño has a neighbor: why we track two ocean signals, not one

Ask a commodity trader about the climate and you will hear about El Niño. Fair enough — ENSO is the loudest signal in the ocean. But the Indian Ocean runs its own switch, the Indian Ocean Dipole, and it is not El Niño's echo. It frequently runs alongside ENSO, sometimes runs entirely without it, and peaks about two months earlier. That is why we publish two ocean indices rather than one — and the data makes the case better than the theory does.

The assumption that trips everyone up The Setup

When we first sketched this work, a reasonable-sounding idea slipped in: that the Pacific and the Indian Ocean essentially take turns — that you watch one switch, and at any given time only one is meaningfully "on." It is an intuitive picture. It is also wrong, and the way it is wrong is the whole reason both indices earn their place.

Here is the test. For every year since 1982, plot the autumn state of ENSO against the autumn state of the Indian Ocean Dipole. If they were the same switch, the points would fall on a line. If they took turns, you would never see both lit at once. Neither is what the record shows.

Two numbers fall out of that chart, and both matter. The correlation is real — about r = 0.63 in the autumn — which is why the two so often share a sign. But it is nowhere near 1.0, and the literature puts numbers on the gap: roughly one-third of dipole events can be traced to ENSO forcing; the other two-thirds are internally generated in the Indian Ocean itself. The dipole is partly nudged by the Pacific and substantially its own animal.

And 2019 is the clean proof. The Pacific sat near neutral that autumn, yet the Indian Ocean threw one of the strongest positive dipoles on record — the event behind East Africa's catastrophic floods and Australia's Black Summer fire season. An El Niño-only watcher saw nothing coming. That single year is why "just watch ENSO" is not enough.

They peak in sequence, not in tandem Timing

There was a kernel of truth buried in the bad intuition, and it is worth rescuing because it is genuinely useful: the two do not peak at the same time. The dipole is sharply seasonally phase-locked — it develops through boreal summer, peaks in September–November, and collapses almost overnight when the monsoon turns. ENSO peaks later, in December–February. You can see both rhythms in their own month-to-month variability.

That offset is not trivia. It is the mechanism by which a developing El Niño can hand off to a positive dipole months before the Pacific itself peaks — and it means that for any crop season that hinges on autumn rainfall, the dipole is often the signal carrying the freshest information. "Take turns" was wrong. "Peak in sequence, and one can light the other" is right.

Why one index cannot stand in for the other The Case

If the two were redundant, we would publish the cheaper one and move on. They are not redundant, because they are different machines with different footprints.

Correlated cousins, not the same switch — and the regions they govern only partly overlap.

The footprints diverge most exactly where it matters for coffee and the monsoon. Over the East African arabica highlands, it is the dipole — not ENSO — that dominates the October–December short rains, with the Pacific's influence arriving second-hand, largely through the dipole. Carry only ENSO there and you are reading the weaker driver. Carry only the dipole and you lose the Americas. The honest answer is that you need both on the desk, each measured on its own terms — which is exactly what we built: a climate-adjusted ENSO read and a self-computed dipole, its two poles broken out so you can see whether a move is the real coupled signal or just transient noise off Sumatra.