Ride the Tide
A tide is not an event. It is a wave, and a wave has to get where it is going. High water touches the whole open beach within a few minutes, then spends hours working its way up the inlet, around the bends and into the creek you actually fish.
Pick a port and watch it happen. Each frame below is one minute.
Why it is late upriver
A tide is a shallow water wave, and a shallow water wave travels at the square root of gravity times depth. Deep water hands it along quickly. Shallow water drags on it. That one fact explains almost everything you see above.
Out on the open coast the water is deep and the wave is moving fast enough that a whole shoreline lights up at once. Turn into an inlet and the depth collapses. The St Johns runs from seventeen metres at the jetties to a couple of metres past Palatka, and the tide slows to match. High water reaches the Mayport jetties and then takes nearly eight hours to get to Welaka, a hundred and thirty kilometres upriver.
Height fades with it. The M2 tide at Mayport is 2.19 feet. At the I-295 bridge it is down to 0.44. Then, oddly, it climbs back to 0.63 at Palatka before dying out. That bump is the wave reflecting off the shallowing river and meeting itself coming back, which is the same trick that makes the Bay of Fundy what it is, just very much smaller.
The Gulf runs on a different clock
Most of the US coast gets two high tides a day. Charleston had two on twenty eight days out of thirty last month. Mobile Bay had one, on twenty five days out of thirty, because the twice daily part of the tide at Dauphin Island has an amplitude of six tenths of an inch. What you feel there is a once a day wave.
Texas is the interesting case, and it is why the Galveston panel is the least trustworthy one here. Galveston had one high tide on ten days last month and two on the other eighteen. Corpus was thirteen and fifteen. The tide there switches character with the moon's declination: high declination gives you one tide a day, and near zero gives you two. Same moon, different number of tides, inside the same month.
A co-tidal line describes one wave. Where two are taking turns, it is the wrong picture, and the held-out error says so: Corpus Christi is the least accurate panel on this page and Galveston is close behind, while Mobile, which is genuinely once a day, sits with the good ones.
This is worth knowing on the water and not just on a chart. If you fish the Texas coast on a rhythm you learned in Carolina, some weeks you will be wrong by half a day.
How the picture is built
Every gauge is pinned to the arrival time NOAA publishes for it. In between the gauges the water itself decides, through the bathymetry: a deep channel carries the signal ahead, a wide shoal holds it back. It is one model covering the whole coast. A port is just a window onto it.
The honest test is to hide a gauge, solve again without it, and see how far off the answer lands at the one place the model cannot see. Those numbers are in the panel under each map, and they are not uniformly good. Rivers behave. Broad open bays are harder, because the wave sloshes rather than travels, and a travel time is the wrong way to describe sloshing.
– Greg Barnes
Depths from SRTM15+ and NOAA ENC soundings. Water extent from USGS NHD. Arrival times from NOAA published harmonic constituents and subordinate station offsets. Not for navigation.