How sedimentation changes a reservoir's area-capacity curve over time
Every reservoir model you work from, whether it's a utility's operating manual or a basin-wide dispatch model, leans on an area-capacity curve: a lookup table built once from a bathymetric survey that says "at this elevation, there's this much stored volume." Traders treat it like a constant. It isn't.
Sediment settles, the curve shifts
Rivers carry silt and sand into every impoundment, and the reservoir's whole job, slowing the current, is also what makes it drop that load. Sediment settles fastest near the inflow delta and in the dead-storage pool, but over decades it works its way across the basin floor. The effect on the curve is simple to state and easy to miss in practice: at a given elevation, the reservoir now holds less water than the curve says, because part of the space that used to be water is now mud.
That bias compounds over the years instead of holding steady. A curve built from a 1998 survey might be reasonably close in 2005 and meaningfully wrong by 2026, especially on high-sediment-load rivers draining erodible terrain, or on reservoirs downstream of recent wildfire or land-use change in the catchment. Industry sedimentation studies have put global average capacity loss on the order of 0.5 to 1 percent of original storage per year, though the number swings hard by basin, by upstream erosion control, and by whether the dam passes sediment during high flows or just accumulates it.
For anyone converting a water-surface elevation into a storage number, that drift matters in a specific way: the error isn't random noise, it's a one-directional bias that gets worse the longer the curve goes unrevised. A reservoir reading "70% full" off an old curve could be sitting meaningfully lower in actual stored energy than that number implies.
Why resurveys lag the problem
The fix, in principle, is a new bathymetric resurvey: send a boat out with sonar, remap the basin floor, rebuild the curve. In practice, resurveys are expensive, slow, and triggered by budget cycles or regulatory requirements, not by how fast a particular reservoir is silting up. Many dam owners resurvey every 10 to 20 years if at all, and smaller or less critical reservoirs can go far longer. Even when a resurvey happens, the updated curve doesn't always make it into the disclosures a counterparty or a grid operator can actually see, and it rarely shows up before the next settlement period.
That gap is where a stale curve costs you money. Estimating storage across a basin or a multi-dam portfolio for trading or balancing decisions means working off whatever curve is on file. That curve may be years out of date by the time anyone uses it. The silt loss itself shows up gradually, and the number it corrupts, your storage estimate, reads a little high every week until someone finally pays for a resurvey.
Reading storage without waiting for the next survey
None of this means the area-capacity curve is useless. It means the curve is one input, and the other input, the actual current water-surface extent, needs to be checked against it regularly rather than assumed. Reservoir Monitoring takes that second half seriously: it measures each reservoir's water-surface polygon from satellite imagery on a weekly cadence and runs it against the area-capacity curve on file to produce a storage-level series, so you're not relying on a single stale snapshot, or on waiting for a utility to publish its own number.
It won't tell you when the curve itself needs a bathymetric resurvey. That's still a physical survey question. But it gives you a current, independent read on surface extent to pair against whatever capacity curve you're using, which is often the difference between catching a drift early and finding out about it in a settlement dispute.
If you manage a watchlist of reservoirs and you're tired of storage estimates that are only as fresh as the last disclosure, it's worth seeing what a weekly surface read looks like against your own curves.