ONI vs RONI: how NOAA's two ENSO indices differ
Both indices read the same patch of the tropical Pacific. The difference is what they compare it against — and in a warming climate, that difference decides whether a winter reads as El Niño, La Niña, or Neutral.
The Oceanic Niño Index (ONI) is NOAA’s long-established ENSO indicator, comparing Niño 3.4 sea-surface temperatures to a fixed 30-year baseline. The Relative Oceanic Niño Index (RONI), which NOAA’s Climate Prediction Center adopted as its primary metric in February 2026, subtracts the all-tropics mean warming trend first, so the El Niño / La Niña signal stands out from background global warming.
ONI vs RONI: what each measures against
Both indices read the same patch of ocean, the Niño 3.4 region (5°N–5°S, 170°W–120°W), as an overlapping three-month running mean. They differ only in what they measure that temperature against.
Niño 3.4 SST compared to a fixed 30-year climatology (refreshed every five years). Long-established and simple, but as the entire tropical ocean warms the fixed baseline lags behind, so a growing share of “warm” readings is really just background global warming rather than a genuine El Niño signal.
Starts from the same Niño 3.4 anomaly, then subtracts the average anomaly across the whole tropical belt (20°S–20°N) and rescales so its variability matches ONI. What is left is how far the Pacific stands out from the tropics as a whole: the contrast that actually drives ENSO rainfall and teleconnections.
Because it strips out basin-wide warming, RONI is far less sensitive to which 30-year baseline is chosen, so the historical catalogue of events stays stable over time. In practice it damps El Niños and amplifies La Niñasrelative to ONI: the 2023–24 El Niño steps down from “strong” to “moderate” (about 0.6°C cooler), the 2020–23 “triple-dip” La Niña reads as one continuous three-year event, and several ONI-neutral winters (2024–25, 2025–26) register as weak La Niñas.
The ±0.5°C thresholds and the five-consecutive-season rule are unchanged between the two; only the reference frame moved. The shift follows L’Heureux et al. (2024), which introduced the relative index for classifying ENSO in a warming climate.
Other agencies frame the same phenomenon a little differently. Japan’s JMA monitors the NIÑO.3 region (5°N–5°S, 150°W–90°W) with a five-month running mean against a sliding 30-year climatology, and declares an event only once the anomaly holds past ±0.5°C for six consecutive months.
Why three-month running means?
ENSO is defined on overlapping three-month averages (DJF, JFM, FMA, … NDJ), not single months. Monthly sea-surface temperatures are noisy — weather, intraseasonal waves like the Madden–Julian Oscillation, and sampling gaps add scatter that has nothing to do with the slow ocean state ENSO actually represents. Averaging three consecutive months smooths that high-frequency noise while preserving the seasonal evolution, so a value only reflects a genuine basin-scale shift rather than a passing spike.
The windows overlap(each month is the center of its own season and also part of its neighbours’) so the index updates every month without losing the smoothing — and the official El Niño / La Niña definition requires the ±0.5°C threshold to persist across five consecutive overlapping seasons, which filters out brief excursions that never become a true event.