Chart datum vs tidal datum: what the difference actually means for survey planning
Ask three people in a hydrographic office to define "chart datum" and you'll get three slightly different answers, and most of them will also mention a tidal datum somewhere in the explanation. That's not sloppiness. The two concepts are related closely enough that people use the terms loosely, but they answer different questions, and mixing them up is how a least-depth figure ends up wrong on a published chart.
A tidal datum is a tide level, nothing more
A tidal datum is a vertical reference derived purely from tide observations at a station over a defined epoch, usually 19 years to average out the lunar nodal cycle. Mean Low Water, Mean Higher High Water, Mean Sea Level: these are all tidal datums. Each one describes an average or a specific tidal condition at that particular gauge. They're statistical constructs built from years of water-level records, and they live and die by the station that produced them.
None of that makes a tidal datum safe to chart against. Mean Low Water, by definition, gets exceeded by lower tides roughly half the time. A vessel with a draft set against MLW is going to touch bottom on a below-average low tide, and that's a problem nobody wants discovered by a keel.
Chart datum is chosen for safety, not for statistics
Chart datum is the reference surface to which charted depths and drying heights are reduced, and it's picked with a different goal: depths on the chart should rarely, ideally never, read deeper than what a mariner actually finds. That's a conservative design choice, not a statistical average.
The IHO's recommendation, and the datum most national hydrographic offices have converged on, is Lowest Astronomical Tide. LAT is the lowest water level that can be predicted to occur under average meteorological conditions, computed from the full set of harmonic tidal constituents rather than observed averages. Because it's near the floor of the tidal range rather than the middle of it, a depth charted against LAT is almost always a conservative figure. Storm surge and other meteorological effects can still push water below LAT, but astronomically, it's about as low as the tide goes.
Older charts and some regional practice still use Mean Lower Low Water or other local conventions instead of LAT. That's part of why this question keeps coming up: the datum on a chart isn't universal, and a survey planner moving between chart series or between countries has to check which one they're looking at before comparing depth figures across sheets.
Where sounding datum fits in
Sounding datum and chart datum are, in practice, the same thing for most modern work: it's the level to which raw soundings get reduced before they go on the chart. The term "sounding datum" just emphasizes the reduction step itself, the tide correction applied to each individual depth measurement taken during a survey so it reads consistently against chart datum rather than against whatever the tide happened to be doing at the moment the echo sounder pinged. That reduction is one of the more tedious parts of planning a hydrographic survey, because it requires tidal zoning, gauge records, and a defensible correction model for the survey area, especially somewhere without a long tide record nearby.
This matters directly for how a reconnaissance layer gets used before a boat ever leaves the dock. An image-derived depth estimate reflects the water level at the moment the satellite pass was captured, not chart datum, so it has to be read as a relative guide to where the shoals and deep channels sit rather than a datum-reduced figure. An annual depth estimate layer built from satellite imagery is useful for exactly that: scoping which parts of an area of interest are worth a survey boat's time, and which stretches are too shallow or too hazardous to risk the vessel at all, well before anyone runs the tidal reduction that turns raw soundings into a charted depth.
Once the area's been narrowed down, the actual reduction to chart datum, LAT or otherwise, still has to happen the way it always has: tide gauges, zoning, and a sounding survey tied to a known datum. Satellite imagery tells you where to point the boat. It doesn't replace the sounding.
If your office is scoping where to send that boat next season, take a look at what an annual depth estimate pass can rule in or out before you commit survey time.