Tide forecasts for 42 Pacific gauges, the odds of flood days this winter, and the sea-level tilt El Niño sets up along the coast.
Water level at the chosen gauge. Blue is NOAA's tide table, which knows only the astronomy. Black is what the gauge read over the last two days, every 6 minutes, so the crest of each high tide is the gauge's own peak; every high and low tide carries its level. Red is the table plus our correction: how far the gauge stands from the table right now (the El Niño background, the season and sea-level rise), relaxing toward the past week's average over about two and a half days (longer at San Diego), plus a weather term. The shading holds 90% of past errors at that lead time. The dotted dark line is the gauge's highest water on record, with its date; it is drawn only when it is near the chart's range.
The weather term is the water that the forecast air pressure (lower pressure raises the water about 1 cm per hPa) and the wind add or remove, relative to the past week, from the mean of the 51-member ECMWF ensemble. At most gauges the wind is the forecast at the nearest offshore NOAA buoy, pushing water along a fitted direction of the coast, with a memory of the last half day and two days, and its weights come from the buoy's and the gauge's record since 2008; at the rest, and whenever the buoy is silent, it is the wind at the gauge's own sensors, refitted on the last 30 days. The shading already includes the weather term's own errors. It is left out when that forecast is missing or older than 36 hours. The reason is given in the line above the chart. The table under the chart splits each day's highest water the same way: NOAA's table, plus the departure (the gauge's standing offset, relaxing toward the week's mean), plus the weather term, equals ours.
Live readings. While the page is open, the chosen gauge's black line (and today's cell on the flood calendar) comes straight from NOAA's 6-minute record, the last 48 hours, refreshed every 10 minutes; the feed's own copy, written every 6 hours, is the fallback.
Units and datum. Levels are in metres or feet (m | ft) above MHHW, the long-term average of each day's higher high tide, or above MLLW, the average lower low tide and the zero of NOAA's tide tables and charts. The choice is remembered. Time. Local is the station's standard time (PST, AKST or HST), all year, with no daylight saving; UTC is the feed's own clock. The zone is named under the chart. Chance of minor is the share of past errors at that lead that would lift the day's highest predicted water to NOAA's minor flood level. The test line compares, for forecasts issued every six hours over the last seven days, the error of NOAA's table one day later against ours.
This is not a storm forecast: there is no wave term yet, so a heavy swell is not in it, and the wind term grows with the square of the wind speed but underestimates the largest surges. Not a warning; follow the National Weather Service. Flooding away from the gauge depends on local terrain and defences.
NOAA tide table plus a correction from the gauge and a pressure and wind term; no waves yet
For the chosen station, a calendar of the season (October to March) coloured by the chance that the highest tide of each day reaches the chosen level: NOAA's minor, moderate or major flood level at the gauge, or the gauge's own record high. Three ingredients: the predicted high tides (fixed by astronomy), this winter's average sea level (the El Niño background: today's gauge departure blending into the coast + tilt outlook over ~3 weeks) and the station's own record of storm surge and weather-driven extras over 1996–2025. Given the observed monthly levels, the same recipe reproduces the flood-day counts of past winters well (correlation about 0.8 at most gauges). Normal assumes an average background.
Hover a day for its highest predicted tide and the chance of each level; for a day already past, what the gauge recorded. Past days are drawn from the gauge: a dark square reached the level, a ringed square did not, a hatched square has no reading yet. Readings are NOAA's preliminary 6-minute record until NOAA verifies the day, so a past day can move by a few millimetres.
The month headers count days reached so far and the days still expected; the headline is the season total. Sea level is how far the water stands above the tide table (the past week's mean at this gauge) and the background level expected for each month ahead, also against the table: the gauge's offset blending into the coastal outlook. At Hawaii gauges the background is the gauge's own persistence. The map colours every station by the season total, by the days already reached (so far), or by the day under the pointer (a day). Click a dot or use the list to change station; the Tide forecast card below follows. The m | ft and MHHW | MLLW choices are shared with it.
This is not a forecast of any storm, wave or particular day, and not a warning: the shading is how likely a tide at that level is on that date given typical weather. Flooding away from the gauge depends on local terrain and defences. Follow the National Weather Service for warnings.
Hovering a day. Past days show the gauge's highest reading against the level. Today uses the tide feed's 6-minute readings so far and turns black once the level has been reached. Days ahead show NOAA's tide-table crest and the crest expected once the month's sea-level anomaly is added, with the chance of each level; inside the next 10 days the crest and the chances come from the tide forecast itself (the one on the Tide forecast card), with its error band.
Chance of reaching a flood level or the record, day by day, Oct–Mar
How far off each high water was, on average, by how far ahead the forecast was made: ours, NOAA's tide table, and the table plus the gauge's offset at the moment the forecast was issued, held flat (no weather, no relaxing). This winter the water has run 30–45 cm above the table at most gauges, so nearly all of the table's miss is that offset. Ours starts from the same offset but lets it relax over about two and a half days and adds a pressure and wind term: the gap between the dashed line and ours is what those two add.
The average error (gauge minus forecast), hour by hour from 6 h before to 6 h after each predicted high or low water, for forecasts made at the chosen lead. The table's line is shifted by its average offset so its shape fits on the same scale. A line rising after high water means the real crest came later or higher than forecast.
The tilt is how much higher sea level stands in the eastern tropical Pacific (5°S–5°N, 200–280°E) than in the western (5°S–5°N, 120–160°E), from satellite altimetry, shown as a departure from normal in cm. Normally the west stands about 33 cm higher; the dashed line is the anomaly at which that slope is fully reversed. The headline is the 7-day mean.
It matters for the coast because the tilt is the ocean's stored response to the trade winds: relaxed winds let water slosh east as Kelvin waves, which raise sea level along the Americas some weeks to months later. A high tilt means coastal sea level is likely to be elevated or still rising. The Events view lines up this season (bold) with every winter since the record began, June to May. Thin lines are coloured by the winter's ENSO phase from the Oceanic Niño Index (November–January value): red for El Niño (+0.5 or above), blue for La Niña (−0.5 or below), grey for neutral. Stronger events are drawn thicker and darker. The three big El Niños, 1997–98, 2015–16 and 2023–24, are labelled.
The dashed outlook carries the tilt four months ahead. It starts from the ocean now (persistence, with the warm-water volume and last month's change) and, where the plume has an El Niño forecast for the month, averages in a second estimate built on that forecast; both were fitted on 1993–2025 and the band is their typical error (one standard deviation, each season left out in turn).
Caveat: every linear model of this kind ran 5–10 cm low in 1997–98, and more at longer leads, so in a strong event the ocean can run above the band.
East minus west Pacific sea level, anomaly, through 16 Nov 2026.
+51.5 cm (+20.3 in) 7-day mean to 16 Nov 2026 · Highest for this week of the year since 1993
Water levels and tide tables from NOAA CO-OPS · weather term from the ECMWF IFS ensemble · flood levels from NOAA and the NWS.
Questions, corrections or requests? Get in touch.