Pond Losing Water: Leak or Evaporation? How to Tell

A dropping pond in July doesn’t automatically mean a leak. In hot, dry, breezy weather, Clemson Extension puts normal evaporation at roughly a quarter to a half inch of surface drop per day — call it an inch and three-quarters to three and a half inches over a week, from sun and wind alone. That’s a lot of water to lose honestly. So before you price bentonite by the ton, figure out which problem you actually have. There are two good ways: a bucket and a day, or a sensor and a week.

The 24-hour bucket test — do this first

Extension offices have recommended versions of this test for decades, and it works. The idea is simple: a bucket of pond water sitting in the pond evaporates at the same rate as the pond, but it can’t leak. So any difference between the two is your leak.

1. Fill a bucket with pond water to an inch or two below the rim.
2. Set it on a step, shelf, or brick in the shallows so a few inches of the bucket sit below pond level — same sun, same wind as the pond surface.
3. Mark the water level inside the bucket with a permanent marker.
4. Mark the pond’s level on the outside of the bucket.
5. Come back in 24 to 48 hours — no pumping, no filling in between.

If both levels dropped about the same, you’re looking at evaporation. If the pond fell noticeably faster than the bucket, water is going somewhere besides the sky.

The test’s weakness is its length. One calm day and one windy day give different answers. A thunderstorm ruins the run. And a slow leak — the kind that costs you a foot a month — can hide inside the marker-line slop of a single day. That’s where logging earns its keep.

What a week of level data shows that one day can’t

Evaporation and leaks have different daily shapes, and a continuous level record makes the difference visible.

Evaporation is a daytime process. It concentrates in hot afternoons, picks up with wind and low humidity, and slows way down overnight when the air cools and the sun is gone. On a level chart, that draws a stair-step: a sag through each afternoon, then a nearly flat shelf through each night.

A leak doesn’t care what time it is. Water moving out through a rusted outlet pipe, a muskrat run, or a gravel seam under the pond drains at a near-constant rate around the clock. On the chart, that’s a steady ramp — and critically, the line keeps falling through the night.

So the question a week of data answers is the one the bucket can’t: what does your pond do between midnight and dawn? Flat overnight shelves point to evaporation. A line that falls all night, every night, at about the same rate as the afternoon, points to a leak. A mix — steep afternoons plus a modest but steady overnight fall — tells you you have some of both, and the overnight rate tells you roughly how big the leak’s share is.

A worked example. Say your pond dropped three inches over a hot week in a drought summer. At a quarter to a half inch a day of evaporation, three inches in seven days is entirely explainable by weather — if the overnight line is flat. The same three inches in a mild, overcast week, or three inches where the level keeps sliding through every night, is a different story. Same total; opposite conclusions. That’s why the number alone can’t settle it, and the shape can.

This is exactly what a Dam & Pond Level Monitor is built for: a submerged pressure sensor reads true water depth continuously and reports over NB-IoT cellular — LTE or low-orbit satellite where cellular is thin — with no power or Wi-Fi at the pond. It’s $499 with the first 3 months of connectivity and the dashboard included, and because the sensing element sits in the water rather than above it, it keeps reading through winter ice, when a surface sensor only sees the top of the ice. The same setup handles bigger water too — see pond and dam monitoring for how it’s used on reservoirs and irrigation storage.

Put rainfall on the same chart

One more thing a single test can’t do: account for rain. A wet week can mask a leak completely — the pond takes on two inches of runoff, loses two and a half to the leak, and the casual glance says “about the same as last Sunday.” Meanwhile the leak got a week older.

A rain gauge on the same dashboard closes that hole. When rainfall and pond level share a chart, you can see whether a bump in level matches the storm that should explain it — and whether the level resumes its overnight slide the moment the runoff stops. If you’re comparing conditions across pastures, the same gauges track rainfall across properties so you know which pond caught which storm.

When it’s just evaporation — and you can stop worrying

Plenty of “leaking” ponds are healthy ponds in a hot month. Penn State Extension notes that evaporation and plant water uptake together rarely exceed about 12 inches even during the driest month. If your losses are inside the normal daily range, slow down in cool or humid weather, nearly stop overnight, and stay under that monthly ceiling — that’s weather, not a defect. Watch the trend, plan your storage for dry years, and put the bentonite money back in your pocket.

If losses run past that, or the overnight line won’t flatten, it’s time to hunt.

The honest limit — a sensor finds that, not where

Level data tells you the pond is leaking, how fast, and when it started getting worse. It will not tell you where the hole is. Finding it is still boots-and-shovel work: inspect the outlet and drain pipe (old corrugated pipe rusts through), walk the dam for wet spots, soft ground, or a suspiciously green strip below it, and look for burrows and tree roots on the banks. Bottom seepage into sand, gravel, or fractured rock usually shows nothing on the surface at all.

And fixing it is dirt work, not electronics — compaction, clay, bentonite, or a liner, matched to your soils. That’s a conversation for your local NRCS office before it’s a purchase; they can characterize what’s under your pond and point you at the federal pond standard. The same overnight-slope reasoning applies to pressurized systems, by the way — if the suspect is a pipeline rather than a pond, see finding water leaks with flow data.

We’ve had sensors in the field for over four years — 2,000+ devices, from stock ponds to a municipal tank in Huntsville, Utah — and every one carries a 60-day money-back guarantee and a 1-year warranty. Tell us your setup — pond size, dam type, how far it sits from cell coverage — and we’ll map the right sensors for it.