Nine feet.

Bench notes

Why it is not a chart

A line on a tide chart encodes one number, height, and it does it in a way that quietly implies the number is the point. It is not. The point is whether the water is over the bar, which way it is running and how hard, how much of the beach is left, and what all of that will be in two hours. Draw the section instead and every one of those is legible at a glance without a single label being read, because they are the same picture. The chart is not wrong. It is just an abstraction that arrived so early nobody remembers choosing it.

The moon pulls about a foot

This is the fact worth taking away. The gauge in the top corner is at exactly the same scale as the water beside it: the raw tide-generating force here, computed from the actual positions of the sun and the moon, is a swing of about a foot on an average day, and never more than about eighteen inches. Nine feet turn up. The difference is not astronomy at all, it is the shape of the Gulf of Maine, a basin whose natural sloshing period sits close enough to twelve hours and twenty-five minutes that it rings like a struck glass. The same moon over a different coastline moves a few inches. That gap between the input and the output is the most under-appreciated thing in the whole subject, and it is why the Bay of Fundy exists.

The sun is a switch

Turn the sun off and the comb along the bottom stops breathing: every tide becomes the same size, because springs and neaps are nothing but the solar tide walking into step with the lunar one and back out again over a fortnight. Turn the moon off instead and what is left is the sun's own tide, about two and a half feet, arriving on the clock at almost exactly the same time every day, because a solar day is what a clock is made of. Neither of those is a display filter. The water genuinely stops answering to the body you switched off, which is the fastest way I know to see which one is responsible for what.

Slack is not halfway

Almost everyone assumes the current is slowest at high and low water because that is when the tide is doing the least. It is the opposite. A river mouth is a neck feeding a basin, so the water going through it is whatever volume the basin is gaining or losing, which is the tide's rate rather than its height. The rate is zero at the turns and greatest exactly midway between them. So it goes slack at high water, runs hardest three hours later, and the Merrimack adds a river to all of it, which is why the ebb is always the stronger of the two and why the bar is a different place on a falling tide than a rising one.

Drawn the way a section draws a flow

The current here is running perpendicular to the view: in and out of the page, not across it. Drawing it as arrows sliding sideways would have been prettier and would have been a lie about the geometry. So it uses the notation every engineering section has used for a century: a dot for a vector coming toward you, a cross for one going away. Flood is dots, ebb is crosses, and both grow and thicken with the speed. It took about three minutes for that to stop looking like a diagram convention and start looking like water.

Where the numbers come from

Nothing is fetched and nothing is a lookup table. The sun and moon are computed from standard abridged series, the tide-generating potential is derived from their positions rather than fitted, and the local response is four constants: how much this coast amplifies the moon, how much it amplifies the sun, and the two lags. Those four are set so the model reproduces the tidal datums NOAA publishes for station 8440452, Plum Island, Merrimack River entrance, and it hits all five of them within half a centimetre. It also lands spring tides about twenty-nine hours after new and full moon on its own, which is the age of the tide on this coast and was not fitted to anything.

What it is not

A prediction to navigate on. The section is schematic rather than surveyed, the bar is a control you set rather than a sounding, and the model carries no nodal correction, so it will drift from NOAA's official tables by minutes now and more over years. It is honest about the physics and approximate about this particular entrance, which is the right way round for a piece meant to explain a tide. For crossing the bar, use NOAA.

The manners

Everything runs in your browser and nothing is stored. The canvas claims no vertical gestures, so scrolling past it on a phone always works, and on a narrow screen the annotations that would collide simply do not draw rather than being shrunk under twelve pixels. The table under the stage is not a fallback: it is the same instrument, with every turn of the tide reachable and the whole month scrubable from the keyboard. With reduced motion on, the water stops rippling and the grass stops moving; the tide, the run button and every control stay exactly where they were.