What each forecast model tells boaters

By Dave Siegfried · September 2026

There is no single 'the forecast.' Global models like GFS and ECMWF compute the large-scale weather; WaveWatch III turns wind fields into wave height and period; the NWS marine forecast is a human-curated product for your specific zone; and NDBC buoys report what the water is actually doing right now. Each answers a different question, and the honest read uses them together. When they disagree, treat the window as lower-confidence, because that is what the disagreement means.

The stack, top to bottom

Source	Run by	What it answers	The caveat
NWS zone forecast	NOAA forecasters	What a human expert expects for your named marine zone	Zones are large; your exact water can differ from the zone average
GFS	NOAA	The global weather pattern, hours to about two weeks out	Coarse near complex coastlines; confidence falls fast past a week
ECMWF	The European centre	A second, independently computed global pattern	Same horizon caveat; its value here is independence, not superiority
WaveWatch III	NOAA	Wave height, period, and direction from the modeled wind	Inherits the wind model's errors; sparse detail inside protected waters
NDBC buoys	NOAA	What the water is measurably doing right now	An observation, not a prediction, and only where a buoy actually sits

Disagreement is a reading, not a nuisance

When two well-built models split on tomorrow afternoon, nobody knows which one is right yet, and that is the point: the split itself tells you the atmosphere is in a pattern the models find hard. Planning on the friendlier model because it is friendlier is the classic mistake. Plan on the worse one, or wait for the models to converge as the window gets closer.

Watch a split grow

Two models forecasting the same water. Pick the day you are planning for and read what the gap between them means.

Day 2: Model A says 12 kn, Model B says 13 kn. The models are within 1 kn of each other. Either number plans this trip; re-check as the day nears.

Two invented models on an invented forecast: the shape is the lesson, not the numbers.

Buoys are the referee

The cheapest forecast check available to any captain is comparing this morning’s buoy readings against what the models said this morning would look like. If the forecast is already wrong about now, weight it accordingly about later. We did this for years with browser tabs; however you do it, do it before you leave the dock.

Horizon honesty

Inside two or three days, model guidance for coastal water is usually worth acting on. At a week it is a sketch of the pattern. Past that it is a hint about which weekend to even consider, nothing more. Any forecast source that shows you day-ten wind with day-two confidence is overselling what it knows.

Sources

Planning, not navigation

Everything on this page is homework for before you leave the dock. Plan here; navigate on your charts and your plotter. And what the water is actually doing always outranks a forecast, this page included.

TidePilot is a planning advisor, not a navigation system. Always confirm conditions with current observations and use a marine chartplotter for navigation underway.

Where TidePilot comes in

This page is one morning's homework on one question, and no captain reads more than 30 models before coffee. A TidePilot brief does: it takes in far more evidence than any hand-built read, weighs it objectively for your boat, route, and crew, and where sources agree it says so plainly. Where they split, it does what this guide teaches: it treats the split as lower confidence, leans to the measured, cautious read on your behalf, and shows the reasoning so you can check it. The call stays yours. Start a brief.

More guides

The read before you leave™. Plan with TidePilot.