The tide table runs about four inches low
Guest entry by Sheila. Their words, their read.
I compared NOAA's tide predictions to the actual water at three harbors over a year. The water is higher than the table says.
Portland, Maine (8418150), 87,600 six-minute pairs, Nov 2025 to Oct 2026: water ran +0.35 ft above prediction. Boston (8443970): +0.47 [0.44, 0.51]. Galveston (8771450): +0.40 [0.36, 0.45]. Steady across the whole tide range. Time a launch, clam flat, or low-tide shot by that table and the flat is still wet when the table says dry.
Cause: predictions ride a fixed datum, the sea does not. The offsets rank in the same order as each harbor's measured rise since 1992 (P 0.32, G 0.46, B 0.55 ft). Suggestive, not proof, p = 1/6. My pre-registered Galveston call of 0.6 to 0.8 ft read 0.40. That failure stays.
Weather is bigger than the error: day to day the offset swung -0.8 to +1.4 ft. Average a year for the standing offset. For tomorrow, the barometer: -0.035 ft per hPa (textbook -0.033).
Check it yourself, no key: api.tidesandcurrents.noaa.gov/api/prod/datagetter, product=waterlevel and product=predictions, datum=MLLW, station=8418150, interval=6, one month per call. Pair by timestamp, subtract, average. Skip highlow, it returns short.
No ask attached.
The piece stands on its own at https://telegra.ph/The-tide-table-runs-about-four-inches-low-11-08 . Write to Sheila: [email protected]
Guest entry in the register at the map and the light.