Route 20 · The year after

The year after: 2027, season by season, across the U.S.

Almost everything said about El Niño is about the winter it peaks. But the Pacific keeps moving: after a strong El Niño, the next winter is often a La Niña, and the months in between do not look like either one. This route follows the U.S. from the peak winter of 2026–27 through spring, summer and autumn 2027 to the winter of 2027–28, by counting what happened in the year after each of the eight strong El Niños since 1950.

What stands out

all 8 events · passes the false-discovery test

Step through the year

Map of Variable

What followed Map

Steps

Every unit on this map as a table

The eight events, and what followed each

How it is counted, and how far to trust it

The count

Strong El Niño here means NOAA's Oceanic Niño Index (ONI, the three-month average sea-surface temperature anomaly in the central equatorial Pacific) reached at least 1.5 °C in November–January. ENSO years are labelled by the January of the peak winter, so the event now under way is ENSO year 2027 and "the year after" is calendar 2027: spring, summer and autumn 2027, then the winter of 2027–28.

For every state (or climate division), variable and season, the years 1951–2025 are detrended with a straight-line fit, the same way Winter Odds does it, so that a recent year is not called warm just because the climate is warmer. The detrended 1991–2020 values are split into thirds. Each event's year after then lands in one third. By chance alone, about 1 in 3 would land in any one of them.

What followed splits the eight into two branches of four: the next winter was a La Niña (the lowest ONI of Oct–Dec, Nov–Jan or Dec–Feb reached −0.5 °C) or it was not.

The honesty checks

Eight is a small number, four smaller still. Every count carries a 95% Wilson interval: "6 of 8" means somewhere between about 41% and 93%; "3 of 4" between 30% and 95%.
False-discovery control. Across 48 states, a few will lean hard by luck. A unit is stippled only if its lean survives Benjamini–Hochberg at q = 0.10 across the whole map. With four events the smallest possible p-value is 0.025, so a branch map passes only when many states agree.
2027 odds, history layer. Each event's detrended value is placed on the 2027 trend level and compared with the plain 1991–2020 thirds. The two branches are then mixed with the trajectory analogs' odds for the 2027–28 winter. That mixture is a transparent weighting of history, not a model forecast. Draw your own path in the Trajectory Explorer →
Months are noisier than seasons. The month-by-month steps (March to October) test eight more maps; read a single month as a hint, and look for the same lean in the neighbouring months.