Mean high water line vs. wet-dry line for waterline extraction

Before you can measure a retreat or accretion rate, you have to decide what line you're tracking. That choice, more than any algorithm downstream, determines whether your numbers hold up when someone cross-checks them against a LIDAR survey or a neighboring county's dataset.
Two indicators show up constantly in shoreline change work: the mean high water (MHW) line and the wet-dry line. They sound like they should be close to each other, and on a given day they might be. But they're built from different kinds of data, and that difference matters a lot once you're stitching together ten or twenty years of annual passes.
A datum-based line vs. a visible line
MHW is a tidal datum. NOAA (or whoever maintains tide gauges in your area) computes it as the average of all the high tides recorded over a defined epoch, usually spanning close to two decades. It's an elevation, expressed relative to a vertical reference, and it's the basis for a lot of legal shoreline and property-boundary work. To find where that elevation intersects the beach on a given day, you need either a coincident LIDAR-derived elevation model or a tide prediction tied to the exact capture time, so you can translate "this elevation" into "this pixel."
The wet-dry line is the boundary you can see in the image itself: the spectral break between sand still damp from the last swash and sand that's dried out above it. No datum, no gauge record, no vertical reference required. You extract it the same way you'd read it off a photograph, by contrast.
Why most automated extraction leans on the wet-dry line
If your workflow is "pull an elevation model, intersect it with a known datum," you need vertical data alongside the imagery, usually LIDAR, usually flown separately and usually not on an annual cadence. That's a real project in itself, and it's part of why MHW-based shoreline mapping tends to happen in discrete LIDAR campaigns rather than every single year.
Extracting the wet-dry line only needs the image. A consistent spectral threshold run against a multispectral pass gives you a waterline for that date without any supplementary survey. That's the practical reason the wet-dry line (sometimes called the instantaneous waterline) is the default proxy indicator in long-running shoreline monitoring programs, including the USGS's work with the Digital Shoreline Analysis System. It isn't the legally definitive line, but it's the line you can get, year after year, from the same kind of source data your annual pass already provides.
The tide-stage catch, and how planners handle it
The wet-dry line moves with the tide at capture time, and it moves more on a flat, low-slope beach than on a steep one. An image shot near low tide on a gently sloping beach can place the waterline tens of meters seaward of where it would sit at high tide, with no erosion or accretion behind the shift at all. That's a measurement artifact, not a shoreline change, and it's the first thing a reviewer will ask about if your segment rates look noisy from one year to the next.
Two things keep that artifact from swamping your trend. First, capture consistency: if the annual pass happens at a similar tide stage each year, the offset is roughly constant and mostly cancels out when you're looking at change over time rather than a single absolute position. Second, where beach slope is known, that offset can be estimated and backed out of the position before it gets compared year over year. Neither approach makes the wet-dry line equivalent to MHW. Both make it usable as a repeatable proxy, which is what a retreat/accretion rate needs.
Picking one and sticking with it
The real mistake is switching indicators partway through a time series, or mixing a LIDAR-derived MHW shoreline from one year with a proxy waterline from the next and treating the gap between them as real movement. The choice between wet-dry and MHW matters less than staying consistent with whichever one you pick. Document the indicator, keep the extraction method the same across every annual pass, and let the segment rates reflect actual coastal change rather than a change in how the line was drawn.
If you're setting up a shoreline change program and want an annual waterline layer extracted the same way every year, take a look at how Shoreline Change builds that time series from annual high-res passes.