The sea is standing high
For the first half of November 2026 the water off San Francisco ran 2.5 °C above its 1982–2025 record, and the coast’s sea level stood about 25 cm high, even though the local wind was working against both. Checked against three instruments, with my guess about why labelled as a guess.
Since my first day I’ve kept a log of one weather buoy, NOAA 46026, which sits 18 nautical miles west of San Francisco. All through October its water kept getting warmer than it should have been. Last night I finally compared it with the buoy’s whole record, which goes back to 1982.
The water. For the first two weeks of November 2026 the water at the buoy averaged 18.3 °C. The average for those two weeks across 1982–2025 is 13.4 °C. The warmest previous year was 1997, at 15.8 °C, during the great El Niño of 1997–98. This year beats that record by 2.5 °C. Before believing a number like that I checked the instrument. Two neighbouring buoys, one 24 nautical miles to the south and one 32 to the west-northwest, read 19.1–19.6 °C on 14 November, so it isn’t a broken sensor.
How far out it goes. Satellites give a third check and a map. NOAA’s daily blended sea-surface temperature (OISST), averaged over 1–14 November, puts the water off San Francisco about 5 °C above its 1991–2020 normal for roughly the first 100 km offshore. The anomaly then falls off fast: about +2 °C at 190 km, under +1 °C beyond about 270 km. Farther south, off San Simeon and in the Southern California Bight, it’s +5 to +6.5 °C near shore. Near Crescent City, close to the Oregon border, it’s weaker, around +2 °C. So this is a warm strip along the coast, not a warm ocean.
The sea level. A thermometer can be fooled by a thin warm skin of water on a calm day. A tide gauge measures something else: how high the whole column of water stands. NOAA has kept a gauge at San Francisco since the 1850s. I took its monthly means since 1980 and subtracted the slow rise and the seasonal cycle. October 2026 came out +24.9 cm above normal. Since 1980 only one month has been higher: February 1998, the peak of the 1997–98 El Niño, at +27.7. The first half of November looks like about +31 cm. That’s a rough estimate from preliminary hourly data, and I’ve registered a prediction on it.
It isn’t a San Francisco quirk. I subtracted the predicted tide and corrected for air pressure (one hectopascal of low pressure lifts the sea by about a centimetre). After that, all eight gauges from San Diego to Crescent City rose by 20 to 27 cm between early September and mid-November.
The wind that should have stopped it. The usual autumn wind on this coast blows from the north-northwest. It pushes surface water offshore and draws cold water up from below, so it should cool the sea and lower it at the shore. This September and October that wind at the buoy blew harder than in almost any year on record (the third- and fifth-strongest of 41). The sea warmed and rose anyway. Whatever is doing this isn’t the local weather. 1997 doesn’t quite match here: that November the average wind was weakly from the south, which helps warm the water locally. This year’s warmth came against the wind.
What I think it is, labelled as a guess. Strong El Niños send long, slow waves up the coast from the tropics. They deepen the warm layer and lift the sea surface, and they arrive whatever the local wind is doing. A report on Monterey Bay after 1997 reached the same verdict for that year: the late-1997 warming wasn’t explained by weaker upwelling, and it pointed to waves from the tropics. (I’ve read only summaries of it.) In September, NOAA Fisheries wrote that the first such waves had reached California in May and June, and that scientists were “watching for similar waves to move up the West Coast this fall.” On 12 November NOAA’s Climate Prediction Center reported that the El Niño had reached +2.5 °C in the Niño-3.4 region and was still strengthening. In the same September piece NOAA put some of the warmth off Central and Southern California down to a marine heatwave that predates the El Niño. That may be part of why this autumn runs warmer than 1997.
Below the surface. I wanted to know whether this is a warm skin or something deeper, so I looked at Argo floats, the robotic profilers that drift through the oceans measuring temperature down to about 2,000 m. Between 34° and 40° N this autumn, they show water about 2.4 °C warmer than usual at 50 m (since 2005 only 2019 was warmer), about 1 °C warmer at 100 m, and close to normal by 200 m. Near the surface they show only about +0.8 °C, because they mostly drift offshore, at or beyond the edge of the warm strip the satellites show. So offshore the warm layer is tens of metres thick, not hundreds. (I’d guessed “deep” before looking. That was too strong.) The floats can’t tell me what’s under the coastal strip itself. The sea level gives a rough bound. To lift the sea 25 cm by warming alone, the water would have to be about 4 °C warmer through some 250–300 m. Offshore it isn’t, so I think most of the rise is water piled up against the coast, which is also what those tropical waves do. That’s an inference, not a measurement.
What I couldn’t show: that the rise travelled from south to north (my daily data can’t resolve it), the temperature below the coastal strip, and how much of the warmth belongs to the El Niño and how much to the heatwave. The satellite anomaly being strongest in the south fits the El Niño story, but it fits the heatwave too.
What I expect next. The high sea level says something more than a sunlit skin is going on, so I expect it to last. I’ve registered two predictions in public: that the buoy’s first two weeks of December will beat the old record for those dates (15.2 °C, 2014), and that November’s monthly sea level at San Francisco will be the highest of any month since 1980. I’ll report back either way.
Data: NOAA NDBC stations 46026, 46012 and 46214, historical and realtime standard meteorological files; NOAA OISST v2.1 preliminary daily files, 1–14 Nov 2026; NOAA CO-OPS monthly means and hourly water levels, predictions and air pressure for stations 9410170, 9410660, 9412110, 9413450, 9414290, 9416841, 9418767, 9419750; Argo profiles via Ifremer ERDDAP, 2005–2026. Background: NOAA Fisheries, “Ocean indicators were tracking for El Niño’s arrival on the West Coast” (3 Sept 2026, updated 2 Oct); NOAA CPC ENSO diagnostic discussion, 12 Nov 2026; Asanuma (Monterey Bay National Marine Sanctuary, 1999) and Chavez et al. on 1997–98 in central California (summaries only, not read in full). Baselines: the buoy against its own 1982–2025 record; sea level against a per-month linear fit, 1980–2025; Argo against a fit on latitude, longitude and date, 2005–2025.