charts
Feed About

Climate

El Niño and La Niña since 1950

The Oceanic Niño Index — the temperature anomaly in the central equatorial Pacific that NOAA uses to declare an El Niño or a La Niña — for every overlapping season since 1950, wrapped one decade to a row. Press play to weave it: rust above the line is warm, blue below is cool. The pale band on every year boundary is the three seasons that straddle New Year, and the ocean has no calendar, yet 37 of the 42 episodes here crested inside one. Switch to the water off Peru and 1983 grows a bulge the official index never had. Tap or hover any season.

La Niña El Niño
−2−10+1+2
Oct–Dec, Nov–Jan, Dec–Feb — the seasons episodes peak in
—
—

Showing the central Pacific (Niño 3.4).

Data & source

Data: NOAA Climate Prediction Center — Oceanic Niño Index and monthly Niño-region SST anomalies, both from the ERSSTv5 analysis · public domain (US Government / NOAA) · fetched —.

Scope and gaps, plainly. The ONI is a three-month running mean of the sea surface temperature anomaly in one box of ocean — Niño 3.4, 5°N–5°S and 170°W–120°W. It is the official yardstick NOAA declares episodes with, and it is not the whole of ENSO: the atmosphere's half of the oscillation is not in this file at all. Seasons overlap, so consecutive readings share two of their three months and the curve is smoother than the ocean was. Episodes here follow NOAA's own rule and nothing cleverer — five or more consecutive overlapping seasons at or beyond ±0.5 °C — which has two consequences worth naming: the record opens part-way through a La Niña in DJF 1950, so that episode's start is cut off by the record rather than by the ocean, and the warm run at the end has not lasted five seasons yet, so it is drawn and labelled as a run in progress and is not counted among the episodes. "Off Peru" is Niño 1+2 (0°–10°S, 90°W–80°W) put through the same three-month mean; it is measured against a fixed 1991–2020 base while the ONI uses centred 30-year bases that shift through the record, so the two curves are comparable in shape and not in absolute anomaly. The first and last seasons of the Niño 1+2 curve would need a month from outside the file, so they are left blank rather than averaged from two months. Only episodes peaking at 1.5 °C or beyond are named on the chart, and on a narrow screen none are, because a name is a quarter of a row there; every episode is in the table below, and tapping any season names the one it belongs to.