NASA’s New El Niño Maps Resemble 1997. What That Does and Does Not Tell Us About Winter.

In June, NASA’s sea-level satellite showed the eastern Pacific lagging 1997, with fewer Kelvin waves than at the same date. On 23 October NASA’s Earth Observatory published a three-way comparison of October 1997, 2015 and 2026 maps, and says the 2026 pattern now closely resembles 1997, with elevated water extending to South America, more closely than 2015 did. That change is the news. Our 13 June post asked whether the east would catch up. According to NASA, it has.

This post asks what the resemblance can support. The weekly ocean data back the east-loaded part of it. The resemblance provides evidence relevant to winter’s odds, but it cannot specify California’s rainfall or guarantee a repeat of 1997-98. The uncertainty sits alongside the finding, not in place of it.

What NASA published

The maps are from NASA Earth Observatory, written by Kathryn Hansen with maps by Lauren Dauphin. They show sea surface height anomalies, meaning how far the sea surface sits above or below its normal level, with the seasonal cycle and long-term trend removed. Red marks higher than average, white normal and blue lower. The image includes a color bar in centimeters, with endpoints labelled ≤−15 and ≥+15. Those open-ended limits prevent reading peak amplitudes from the colors alone. The article text does not specify the baseline period. A 2023 NASA caption using the same satellite sequence gave numbers for the earlier years: by early October 1997 and 2015, large areas of the central and eastern Pacific were more than 18 centimeters above normal.

Josh Willis, sea-level project scientist at NASA’s Jet Propulsion Laboratory, is quoted saying no El Niño is truly typical, that the 35-year altimetry record allows a direct comparison with 1997, and that the current event may grow larger. He expects impacts could resemble 1997-98, which brought heavy rain to California and the southern U.S., and he cautions that the atmosphere is chaotic, that the signal can blur outside the tropics, and that resembling 1997 does not guarantee identical outcomes. NASA describes the exact winter impacts as uncertain, and says the 2026 event arrived much more suddenly than 1997.

The change from June is specific. NASA’s 18 June article said fewer Kelvin waves had built up in the eastern Pacific by 8 June than at the same point in 1997, and JPL said in May that this event started later than 1997 and 2015. By October, NASA describes warm water pressing to South America the way it did in 1997.

What satellite sea level measures, and what it does not

Radar altimeters measure the height of the sea surface. Because water expands as it warms, a higher surface indicates warmer water. NASA’s Earth Observatory explained in June that the measurement reflects heat stored beneath the surface, not just surface temperature, and that a large subsurface heat reservoir can matter more for climate and weather than a shallow warm layer. JPL added in May that Sentinel-6 maps ocean height about every 10 days, that it continues a record started in 1992 by TOPEX/Poseidon, and that it detected a Kelvin wave raising sea level more than 15 centimeters above long-term averages off Peru by mid-May.

Three limits matter for the comparison. A height anomaly is not a sea surface temperature map, and it does not by itself show how the heat is distributed with depth. The three maps come from three different satellites, though the record is built to be continuous. And the NASA article we read gives no similarity statistic, so “resembles” is a scientific description, not a computed score. CPC’s 8 October discussion adds the subsurface view: the subsurface temperature index was well above average, reflecting a deeper-than-average thermocline.

Test one: is the structure the same?

We cannot reproduce NASA’s maps, because we did not have the gridded altimetry fields. We can check the east-west structure of the warming on NOAA’s relative weekly Niño index, the index this site tracks, using the same file for all three years and the weeks closest to NASA’s map dates.

These are relative sea surface temperature anomalies, with the tropical mean subtracted, so they compare structure and are not comparisons of absolute temperature. They separate 2015 from the other two. In 1997 and 2026, Niño 1+2, the box off Ecuador and Peru, carries the largest anomaly (+3.4 and +4.4), above Niño 3.4 by 1.3 and 2.0. In 2015 the anomalies are flatter, with Niño 1+2 at +1.6, below Niño 3 at +2.0 and Niño 3.4 at +1.8. That agrees with NASA’s description that in 2015 the elevated water stayed farther offshore. Niño 4 is near zero in all three years (+0.3, +0.2 and −0.1), so the far-western box does not tell the events apart. Between 1997 and 2026 the 2026 relative anomalies are larger in Niño 1+2, 3 and 3.4, by 1.0, 0.7 and 0.3, and lower in Niño 4 by 0.4. These are regional averages, not a map correlation. They support the east-loaded structure NASA describes, not an identical ocean.

Test two: is the trajectory the same?

NOAA’s RONI series gives the build-up for all three years on one scale.

By July to September, 2026 sits at +1.69, below 1997 at +1.84 and above 2015 at +1.54. But it got there differently. 2026 started the spring at about zero and climbed +1.73 from the March to May season to July to September, against +1.37 in 1997 and +0.82 in 2015. That numerically supports NASA’s line that this event came on faster, and JPL’s May note that it started later than 1997 and 2015 and was catching up. It also means “resembles 1997” is true of the shape and the stage, but not of the history. A map taken on a single October day does not show that.

The peak is the open part. 1997-98 peaked at +2.28 and 2015-16 at +2.25, both in November to January. CPC’s 8 October discussion gives chances of a 3-month RONI of +2.5 or more of 54% for September to November, 83% for October to December and 70% for November to January, and says a strong to very strong event is likely through January to March 2027 at greater than 83%. If that verifies, 2026 will have exceeded both events, and “resembles 1997” will have understated it. It is a forecast, not an observation.

Why strong El Niños have produced different winters

This is where the maps stop. Scripps researchers, in analysis reported by UC San Diego in 2017, compared three strong events with eastern equatorial Pacific temperatures more than 2°C above average. California received 150% of average precipitation or more in 1982-83 and 1997-98. In 2015-16 it received less than the average of non-El Niño years.

The Scripps explanation points outside the eastern Pacific. In 2015-16, sea surface temperature and rainfall anomalies in the Indian Ocean and western Pacific were unlike earlier events, shifting storms that would normally reach California farther north, and the Pacific Northwest received record precipitation. Most models did not predict that pattern. Random atmospheric variability also contributed. A separate account, from Benjamin Cash of George Mason University and COLA in a NOAA PSL webinar in 2020, finds El Niño explains at most about one-third of observed West Coast rainfall variance, and that unpredicted variations in the upper-level height field largely caused the dry 2015-16 winter. These are two explanations of the same winter, not a settled one. NOAA’s Mike Halpert wrote the same caution in October 2015: El Niño tilts the odds but does not determine outcomes, even for strong events.

NASA’s own article carries the same split. It says 2015-16 brought more rain to Central America and the southeastern U.S., with significant flooding in Mexico. A map of tropical sea level cannot show which way the jet stream will go.

What this means for 2026

The resemblance is evidence, and it points one way. NASA’s Willis considers 1997-like impacts plausible, and CPC says a historic event would make impacts consistent with El Niño more likely, though not guaranteed. That is the odds language our California winter outlook already uses. What the resemblance cannot do is specify California’s rainfall or guarantee a repeat of 1997-98. The factors that turned 2015-16 dry, such as Indian Ocean conditions and mid-latitude atmospheric variability, are not visible on a Pacific sea-level map, and we have not assessed them for 2026. For our earlier look at the two benchmark years, see the 1997-98 refresher and the 2015-16 comparison.

What a fuller test of “resembles 1997” would need. A pattern correlation of height anomalies over a stated box and baseline, at matched dates for several years, with the same done for upper-ocean heat content and for atmospheric fields such as tropical rainfall and 200-hPa height. We did not have the gridded data to compute that, and the NASA article we read does not report one.

What we cannot say

We cannot say how closely the maps match, because the NASA article we read gives no similarity measure or baseline period, and the color bar’s open-ended limits at ≤−15 and ≥+15 centimeters prevent reading peak amplitudes. We did not read the Scripps paper itself, only the UC San Diego account. We have not assessed the Indian Ocean, the MJO or the Atlantic for 2026, all of which the cited work suggests can matter for the winter.

Our reading. The news is the shift: in June the east lagged 1997, and in October NASA says it resembles it, and the index data agree on the east-loaded structure and the stage reached. 2026 got here faster and, per CPC, is on track to peak higher. The resemblance bears on the odds and cannot specify California’s rainfall. Strong El Niño events produced sharply different California winters, with differences in ocean conditions and atmospheric variability contributing to the outcomes.