Three NOAA gauges (San Diego, Los Angeles, San Francisco) read about a foot above NOAA's own tide predictions: 9.9 to 14.4 inches as a 24-hour mean on 2 October, and 11.7 to 14.5 inches in the 12:00 to 15:00 UTC window on 3 October. Crescent City read 2 to 4. The sources that address it call that warm water, with the Kelvin wave billed as extra. The gauges cannot confirm the split, and none of it is a flood forecast.
Reports say a pulse of warm water called a Kelvin wave is lifting sea level along the West Coast (our explainer covers the mechanism). Three numbers get blended in that coverage, and they are different things: the sea level the gauges already read above NOAA's own tide prediction, the extra rise the arriving pulse is forecast to add, and the height of the tides still to come. This piece keeps them apart, uses dated NOAA tide-gauge readings for the first, and does not add them together.
The short answer is that the gauges already show a large anomaly in southern and central California, and the two sources that address its cause, KQED and climate scientist Daniel Swain, attribute it to warm water along the coast and describe the Kelvin wave as an additional rise on top. These gauge readings cannot confirm that split, because the tide table leaves out everything except the astronomical tide. Measured against each gauge's own mean for 23 to 29 September, the mean for 30 September to 2 October is higher by 4.1 inches at San Diego, 3.7 at Los Angeles, 2.3 at San Francisco and 0.0 at Crescent City. The callout below sets a dated test for whether the reported extra rise appears at San Francisco and Crescent City.
NOAA publishes two series for each tide station: the water level its gauge actually measured every six minutes, and a predicted level computed from the astronomical tides. The difference, observed minus predicted, is everything the tide table does not include: weather, warm or cold water, currents, and slow changes in mean sea level. Taking both series from the same station, on the same datum (MLLW) and the same clock (UTC), and averaging the on-the-hour readings over each UTC day, the difference on 2 October, the latest complete day, was:
That is not an artifact of one afternoon. San Diego averaged 14.1, 14.0 and 14.4 inches on 30 September, 1 October and 2 October, Los Angeles 13.0, 12.9 and 12.5, and San Francisco 8.8, 9.4 and 9.9. In the three-hour window from 12:00 to 15:00 UTC on 3 October the difference was 14.4 (San Diego), 14.5 (Los Angeles), 11.7 (San Francisco) and 3.9 inches (Crescent City). A full 3 October UTC day is not available yet; the hours so far average 14.2, 14.1, 10.8 and 4.1 inches at the same four gauges.
| UTC day | San Diego | Los Angeles | San Francisco | Crescent City |
|---|---|---|---|---|
| 23 Sep | +9.2 | +8.4 | +8.4 | +2.9 |
| 24 Sep | +8.9 | +7.6 | +7.8 | +2.9 |
| 25 Sep | +8.6 | +7.9 | +6.9 | +3.0 |
| 26 Sep | +9.5 | +8.8 | +6.7 | +1.6 |
| 27 Sep | +10.1 | +9.4 | +6.8 | +3.2 |
| 28 Sep | +11.1 | +9.8 | +6.0 | +2.4 |
| 29 Sep | +12.9 | +11.7 | +6.9 | +1.7 |
| Mean, 23 to 29 Sep (baseline) | +10.0 | +9.1 | +7.1 | +2.5 |
| 30 Sep | +14.1 | +13.0 | +8.8 | +2.0 |
| 1 Oct | +14.0 | +12.9 | +9.4 | +3.5 |
| 2 Oct | +14.4 | +12.5 (23 h) | +9.9 | +2.2 |
| Mean, 30 Sep to 2 Oct | +14.2 | +12.8 | +9.4 | +2.6 |
| 3 Oct | +14.2 (21 h) | +14.1 (18 h) | +10.8 (21 h) | +4.1 (21 h) |
Inches above NOAA's prediction, mean of the on-the-hour readings for each UTC day, preliminary data. Where a day is incomplete the hours averaged are shown: NOAA reported no Los Angeles readings from 22:30 UTC on 2 October to 01:42 UTC on 3 October, so that gauge's 2 October value covers 23 hours and its 3 October value starts at 02:00 UTC, and 3 October runs only to about 20:00 UTC. The bold rows are means over every hourly reading in the window.
For context, the table below shows three-hour samples from earlier dates and from a year ago, which show how the anomaly built through the late summer.
| Gauge (NOAA ID) | 10 Aug | 1 Sep | 15 Sep | 26 Sep | 3 Oct | 3 Oct 2025 |
|---|---|---|---|---|---|---|
| San Diego (9410170) | +4.8 | +9.5 | +11.8 | +9.6 | +14.4 | +1.1 |
| Los Angeles (9410660) | +4.7 | +7.1 | +11.0 | +8.9 | +14.5 | +0.9 |
| San Francisco (9414290) | +5.3 | +4.5 | +6.9 | +6.9 | +11.7 | +3.2 |
| Crescent City (9419750) | -0.1 | +2.2 | +2.8 | +2.6 | +3.9 | +0.7 |
Inches above NOAA's prediction. Each value is the mean of seven half-hour readings from 12:00 to 15:00 UTC, a three-hour sample and not a day's average. Within each sample the highest and lowest readings differed by 0.8 to 4.1 inches (the largest spread was at Crescent City on 10 August). At San Francisco the three-hour value ran 0.7 to 1.4 inches above that day's mean from 30 September through 3 October, so a three-hour value can overstate a day's level. The 10 August and 3 October 2025 columns are verified data; the rest are preliminary and can change.
Much of the elevation appears to predate the pulse. Each gauge's mean for 23 to 29 September, before the step described below, was 10.0 inches at San Diego, 9.1 at Los Angeles, 7.1 at San Francisco and 2.5 at Crescent City. San Diego read 9.5 inches above prediction in a three-hour sample on 1 September and 11.8 on 15 September, weeks before the reports expect the pulse in California. Live Science reported the wave reached South America's coast in late August, Swain wrote that the largest coastally trapped wave of the sequence so far arrives in October, and KQED put a second wave's arrival in the Bay Area as soon as mid-October. KQED also said sea levels were already about 6 inches above predicted levels because of warm water along the West Coast, and Swain said tides were already running a foot or more above average thanks to the extreme marine heat wave. This piece did not test that cause, but the readings are consistent with it. The baseline is not flat everywhere: San Diego's daily mean climbed from 8.6 inches on 25 September to 12.9 on 29 September and Los Angeles's from 7.9 to 11.7, so part of the rise at those two gauges began inside the baseline week. San Francisco's stayed between 6.0 and 8.4.
Last year looked very different. In the same three-hour window on 3 October 2025 the gauges read 1.1 (San Diego), 0.9 (Los Angeles), 3.2 (San Francisco) and 0.7 inches (Crescent City). So 2026 is far above a year earlier at the three southern and central gauges. That says the water is unusually high. It does not say why.
It does not rise in a straight line. In the three-hour samples San Diego fell from 11.8 inches on 15 September to 9.6 on 26 September before rising to 14.9 on 1 October. A single reading cannot be called the pulse.
San Francisco stepped up between 29 and 30 September. Its daily mean read 6.9, 8.8, 9.4 and 9.9 inches from 29 September through 2 October, and 10.8 for the hours available on 3 October. Pressure is one thing that can lift water, so I checked it: San Francisco's air pressure fell about 2 hPa between the 12:00 to 15:00 UTC windows on 29 and 30 September, which by the usual rule of thumb of roughly 0.4 inch per hPa explains under an inch of the step, and pressure was about 1 hPa higher in the 3 October window than on 29 September, which would lower the water, not raise it. Wind and waves were not checked. From 29 September to 2 October the daily mean rose about 3 inches, below the 6-inch low end of the range the reports describe for the pulse, and the step could include the pulse or other influences. KQED's figure of about 6 inches above predicted levels at the Bay Area is in line with San Francisco's 23 to 29 September mean of 7.1 inches.
The north-south pattern fits two stories. The 2 October means fall from 14.4 inches in San Diego and 12.5 in Los Angeles to 9.9 in San Francisco and 2.2 in Crescent City. A pulse moving north that had not yet reached Crescent City would look like that. So would a warm-water anomaly centered in the south. A rise at the northern gauges would be what a northward-moving pulse predicts, but a gauge rise only measures the water-level change. Calling it a traveling wave takes independent evidence of northward progression, such as the timing of rises from south to north or the sea surface height analysis in Amaya's tracker, described below. That is why the callout below tests the change at the northern gauges and says nothing about cause.
The Associated Press story of 3 October says sea levels along the West Coast would begin to rise by up to a foot this month as a series of pulses of warm water hit the shoreline, with a 6-inch to foot-tall surge moving north at about 6 mph, or 150 miles a day, reaching Oregon and Washington over the next week. It says experts predict the waves will continue to pass through from October until February. The 6 to 12 inch figure for Orange County comes from Alyssa Napuri, a deputy chief of staff to a county supervisor, who said forecasts show the wave could raise levels by that much. In the text I read, AP does not name the model or forecaster behind the figures.
KQED reported on 1 October that a deep-ocean wave could raise Bay Area sea levels by up to a foot by the end of November, and that this is on top of sea levels already about 6 inches above predicted levels because of warm water along the West Coast. It put the wave's arrival in the Bay Area as soon as mid-October. Dillon Amaya of North Carolina State University urged people to temper expectations and said the wave was still weeks from the region. Swain's own 27 September post on Weather West says tides are already running a foot or more above average thanks to the extreme marine heat wave, and that additional sea level elevation may begin soon as the largest coastally trapped Kelvin wave of the El Niño sequence so far arrives in October. Live Science (30 September) says up to 1 foot for months.
The nearest thing to a published wave tracker is Amaya's coastal trapped wave page. It says such waves can raise background sea levels by 6 to 12 inches depending on the strength of the wave, and it posts daily updated sea surface height analyses and forecasts from the GLORYS ocean system as waves move up the coasts of South America, Central America, Mexico and the United States. It is a model-based analysis, not a NOAA tide-gauge forecast, and in the text I read it does not assign that range to a particular station and date. None of the press reports I read names the model behind its figures or cites a NOAA gauge forecast.
KQED's and Swain's wording is additive: the foot is an increase on top of the existing anomaly, not the anomaly itself. AP's 6 to 12 inch figure and Live Science's foot do not say. Neither KQED nor Swain shows that none of the 30 September to 3 October increase at the gauges is already part of a pulse. I have not added any reported rise to the gauge readings above. A forecast range with no stated gauge, datum or time cannot be summed with a reading at a named gauge, and the callout instead tests the change at named gauges.
The sources disagree, and the disagreement is easy to hide behind a date range. AP describes a series of pulses and says experts predict waves passing from October until February. KQED, quoting its sources, says as many as six Kelvin waves can occur in an El Niño year, that temporarily high seas can last three to six months, and calls the one now moving north a second wave, travelling about 6 mph from the Gulf of California. Gizmodo quotes Swain calling this the largest coastally trapped wave of the sequence so far, and writes that El Niño tends to produce more than one. Amaya told ABC7, as Gizmodo reports, that he does not expect Bay Area sea levels to return to normal until spring 2027. Live Science quotes David Webb of the UK's National Oceanography Centre saying he doubts there will be more than one full wave this time.
So “through February” can mean several pulses arriving or the lingering high seas after one, and the sources do not settle which. A date range is not a wave count. My own reading, not a source's: a second, separate rise at San Francisco after the first has faded would look like another pulse, while a single rise and a slow decline would fit one.
NOAA's predicted high tides at San Francisco, in feet above MLLW and local time, include 6.08 ft on 11 October (12:18 PDT), 6.67 ft on 28 October (12:56 PDT) and 7.03 ft on 25 November (10:49 PST), the highest in a check of 20 to 30 November. NOAA's San Francisco datums put mean higher high water at 5.84 ft above MLLW, so the 28 October predicted high is 0.83 ft above it and the 25 November one is 1.19 ft above it. KQED quotes USGS's Sean Vitousek giving spring tides above 6 feet around 11 October, a peak of 6.68 feet on 27 October and king tides over 7 feet in late November, on top of the Kelvin wave. NOAA's high and low table agrees on 11 October and late November; for late October it has 6.65 ft on 27 October and 6.67 ft on 28 October, a small difference this piece resolves in favor of the table.
These are tide-table values. They do not contain the anomaly in the first number, which changes from hour to hour, so adding today's reading to a tide four weeks away would not be a forecast. They show only that the astronomical tides are already building toward late October and late November.
A higher baseline does not flood anything by itself. In AP's story NOAA Fisheries' Michael Jacox said the Kelvin wave would not come and flood anything on its own, and Amaya said actual damage will depend on storms and tides. The precedent is in California's own record: a December 2015 memo from the state's Ocean Protection Council describes the 1982 to 83 and 1997 to 98 El Niños as bringing 6 to 10 inches of elevated sea level from fall until late spring, with peak water levels of 1.5 to 3 feet above predicted levels during winter storms. The baseline is the first layer; tide, swell and storm surge are the layers that decide whether water reaches a street or a house. This piece does not forecast any of those. Our California emergency post, two-events sea level comparison and NASA 1997 comparison cover the surrounding ground.
“Sea level is already up a foot” and “a foot more is coming” are different sentences, and the coverage has used both. The measured part is that three California gauges averaged about 10 to 14 inches above NOAA's tide predictions on 2 October and the northernmost one averaged 2.2. The forecast part is a reported range: the press reports name no model, and Amaya's tracker, the closest published analysis, assigns no range to a gauge or date. Whether the pulse accounts for the measured part, and whether any of it becomes flooding, will show up in the weekly gauge gap, independent evidence of the wave's progression, the tides and the winter storms, not in a single number.