The bucket is a control, and that is the whole idea
A pool loses water two ways at once and only one of them is a problem. Evaporation takes water off the surface every day whether the shell is sound or not, at a rate that depends on air temperature, humidity, wind and how warm the water is. A leak takes water out of the bottom. Looking at the level in the morning tells you the sum of the two and nothing about the split.
The bucket fixes that. Fill a bucket with pool water, sit it on a step so it is at the same temperature, and leave it open to the same sun and the same wind. It loses evaporation and, being a bucket, nothing else. Whatever the pool loses over and above the bucket is going somewhere the bucket cannot go.
Leak rate = pool drop − bucket drop
It is a differential measurement, which is why it works at all. The weather during the test does not have to be typical, or even known, because both vessels get the same weather and it cancels.
Turning inches into gallons
An inch of depth across a surface is that area in square feet, divided by twelve, times 7.48052 gallons per cubic foot. For a 16 by 32 pool:
512 ÷ 12 × 7.48052 = 319 gallons per inch
Which is the number that makes small readings alarming and large ones survivable.
| Net drop per day | 16 x 32 pool (512 sq ft) | Per 30 days |
|---|---|---|
| 1/16 in | 20 gal/day | 598 gal |
| 1/8 in | 40 gal/day | 1,197 gal |
| 1/4 in | 80 gal/day | 2,394 gal |
| 1/2 in | 160 gal/day | 4,788 gal |
| 1 in | 319 gal/day | 9,575 gal |
Two consequences fall out of that table. The first is that a leak small enough to be invisible on the tile line is expensive: an eighth of an inch a day, which nobody would notice, is over a thousand gallons a month. The second is that your measurement resolution sets the smallest leak the test can find. If you cannot read the level to a sixteenth of an inch, you cannot resolve twenty gallons a day, and the answer to that is a longer test rather than a sharper eye — a 48 hour test halves the rate error of a 24 hour one.
What is normal evaporation
Between a quarter and a half inch a day covers most residential pools in summer, and the range is wide because the drivers are. A heated pool in dry windy weather can pass an inch a day on its own, with nothing wrong with it. A cool sheltered pool under a cover loses almost nothing. This is exactly why an absolute number for what a pool should lose is useless and the bucket comparison is not.
If you want to know why your evaporation is what it is, rather than just how much of the loss it accounts for, the pool heat loss calculator works it out from air temperature, humidity and wind, and prices the heat that leaves with the water. That page predicts evaporation from the weather; this one measures it with a bucket. When the two disagree badly it is usually the wind estimate.
Two tests, and what the difference points at
Running the test twice — once with the pump running, once with it off for the whole period — splits the plumbing in a useful way, because the pressure in the pipes is different in each state.
- Worse with the pump running. The return plumbing is only under pressure when the pump is on, so a pressure-side leak shows up then and largely stops when it does not.
- Worse with the pump off. Suction lines run under vacuum while the pump works and under static head when it stops, so a suction-side leak can behave the opposite way. Structural losses often look like this too, since the water level is higher when the pump is not circulating.
- The same either way. That points away from the circulation loop entirely: the shell, the skimmer throat, a light niche, a tile line crack, a fitting.
This narrows the search. It does not locate anything, and the distinction is worth respecting. Finding a leak is done with pressure tests on individual lines, dye at suspected points and acoustic gear, by people who do it every day. Handing them two numbers and a pump state is genuinely useful; guessing from the same two numbers and opening a deck is not.
What ruins the test
An autofill that nobody switched off is the commonest, and it makes a leaking pool look perfect. A hose left trickling does the same. Rain during the test period makes the pool gain against the bucket and can turn a real leak into an apparent surplus, since the bucket catches rain over a much smaller mouth relative to its depth. Swimmers carry water out on themselves and splash more out again. Backwashing a filter dumps a large slug in one go.
All five are visible if you write down what happened during the test, which is the argument for doing it deliberately rather than noticing a low level and reasoning backwards.
An empty pool can float out of the ground. Groundwater pushing up under an unweighted shell has lifted finished pools clear of their beds, and the season, the water table and the soil all decide whether that is a real risk on a given day. Whether a pool gets emptied, when, and what happens first is a decision for a licensed pool contractor, not something to work out from a web page.
Anything electrical within reach of pool water is licensed work, and the reason is that people are standing in the circuit. Bonding, grounding, receptacle placement, clearances and equipment disconnects are set by the code your jurisdiction has adopted and enforced by its inspector, and none of it appears on this page. If a pump, light, heater or panel needs more than plugging into a working receptacle, that is a job for a licensed electrician.
Questions people ask
How do I do a bucket test on a pool?
Fill a bucket with pool water and stand it on a step so it sits at the same temperature as the pool and gets the same sun and wind. Mark the water level inside the bucket and mark the pool level on the tile or skimmer at the same moment. Turn the autofill off, disconnect any hose, keep everyone out of the water and do not backwash. Come back 24 hours later and measure how far each level has dropped. The bucket drop is evaporation. Anything the pool dropped beyond that is going somewhere else. Enter both numbers here with the pool surface area and you get the loss in gallons a day.
How much water should a pool lose to evaporation?
A quarter to a half inch a day covers most residential pools in summer, but the range around that is so wide that the figure is close to useless on its own. Heated water evaporates much faster than cool. Dry air pulls more than humid. Wind across the surface is the biggest single driver and a windbreak changes a pool measurably. An uncovered heated pool in a dry breezy climate passing an inch a day can be perfectly sound. That is exactly why the bucket exists: it experiences the same weather as the pool, so the comparison works without anyone having to know what normal is for your site this week.
My pool loses an inch a day. Is that a leak?
On a 512 square foot pool that is 319 gallons a day, which is a lot of water, but whether it is a leak depends entirely on what the bucket does over the same period. In hot dry windy weather on heated water, an inch a day can be evaporation alone. In humid still weather it almost certainly is not. Run the bucket test before drawing a conclusion, and run it for 24 hours rather than an afternoon. If the bucket drops half an inch and the pool drops an inch, you have found roughly 160 gallons a day of loss that is not evaporation, and that is worth a call.
Does the size of the bucket matter?
Not for the measurement, since evaporation depth per day does not depend on the area of the surface it is coming off. What matters is that the bucket water is at the same temperature as the pool water and exposed to the same air, which is why it goes on a step in the pool rather than on the deck in the shade. A bucket on hot concrete in the sun will run warmer and evaporate faster than the pool, which makes a leaking pool look sound. A tall narrow container sheltered from wind will do the reverse. Straight-sided is better than tapered because the level drop is then the same everywhere in the bucket, and a mark you can read to a sixteenth of an inch matters more than the volume.
Where is the leak likely to be?
This page will not guess, and neither should anyone who has not looked. What it can do is narrow the field, by comparing a test with the pump running against one with it off — pressure-side plumbing behaves differently from suction-side plumbing, and structural losses often behave the same in both states. Beyond that, locating a leak is a trade. It involves isolating and pressure testing individual lines, dye testing at suspected points, and acoustic listening equipment, and the people who do it find leaks in an afternoon that an owner would spend a season chasing. The gallons a day figure from this page is exactly the information that makes that call productive.