Pool Leak Detection: How to Find a Pool Leak Without Guessing

"I think my pool is leaking" is one of the most common calls a service business takes, and most of the time it is wrong. Water leaves a pool constantly through evaporation, splash-out and backwashing, and an owner who has never watched the level closely has no baseline to judge against. The first job is therefore not finding the leak. It is establishing whether there is one.

When there is a leak, the useful skill is not a clever trick for spotting cracks. It is working in an order that eliminates whole categories before you waste an afternoon on any one of them. A leak in a return line, a leak in the shell and a worn gasket sending water out of the backwash port all present as the same falling water level, and they are found by three completely different tests. Getting the order right is most of the job.

Most suspected leaks are not leaks

Before anything else, ask what changed. A pool that has always dropped half an inch a week and is now dropping two inches has a new problem. A pool an owner has simply started paying attention to in August, during the hottest and driest stretch of the year, usually does not. Evaporation is not a constant: it rises with air temperature, water temperature, wind across the surface and low humidity, and it falls with a cover, shade and still air. The same pool can lose three times as much in a dry August week as it does in a humid one.

Then rule out the ordinary drains on the volume. A pool that only drops after a weekend with children in it is losing water to splash-out. A sand or DE filter backwashed weekly sends a meaningful volume down the waste line each time, and nobody counts it. A heated pool or a spa left running at temperature evaporates far faster than a cold one. An automatic fill device complicates all of this in the other direction, because it quietly replaces whatever is lost: a pool with an autofill and a genuine leak shows a normal water level and an alarming water bill, which is why the owner's utility statement is sometimes better evidence than the pool is.

Evaporation rate

The depth of water a pool surface loses to the air per day, independent of any leak. A quarter of an inch a day is the figure most of the trade works to as a starting point for an uncovered residential pool in mild conditions. Heat, wind and dry air push it well above that; a cover cuts it sharply. It is a rule of thumb for deciding whether a loss is worth investigating, not a measurement, and it is the reason a leak is always diagnosed by comparison rather than by an absolute number.

The bucket test, done properly

The bucket test works because it gives you a control. A bucket of pool water sitting on a pool step is exposed to the same sun, air temperature, humidity and wind as the pool itself, so it evaporates at roughly the same rate. If both levels fall together, the pool is evaporating. If the pool falls faster than the bucket, the difference is the leak. It requires no equipment beyond a bucket and something to mark with, and it is the only test that can prove a negative, which is why it goes first.

1. 1

   Set the pool level and turn the autofill off

   Bring the pool to its normal operating level, roughly halfway up the skimmer throat. Then shut off any automatic fill device at its valve. A test run with an autofill still working measures nothing at all, and this is the step most often skipped.

2. 2

   Put the bucket on a step, not on the deck

   A bucket on hot concrete warms up and evaporates faster than the pool, which makes a real leak look like normal evaporation. Standing it on the second step keeps the bucket water at pool temperature. Weight it with a brick or a few pounds of water so it cannot tip or float.

3. 3

   Match the levels, then mark both

   Fill the bucket with pool water until the level inside matches the pool surface outside it, so the same water temperature and the same starting point apply to both. Mark the inside of the bucket and the outside of it, or mark the tile, with a grease pencil or a strip of tape. The two marks are the whole measurement.

4. 4

   Leave it 24 hours, pump running as normal

   Run the pump on its usual schedule so the test reflects how the pool actually operates. Keep swimmers out, and if rain is forecast, wait for a dry day: rain adds to both but not evenly, and it invalidates the comparison.

5. 5

   Read the difference, not the drop

   Both levels will have fallen. If they are within about an eighth of an inch of each other, the pool is evaporating and there is nothing to find. If the pool has dropped noticeably further than the bucket, the gap is the leak, and now the work starts.

Pump on, pump off: the test that splits plumbing from structure

Once a leak is confirmed, the next question is whether it is in the plumbing or in the pool itself, and the pump answers it for free. The plumbing on the far side of the pump is under pressure while the pump runs and at rest when it stops. The plumbing on the near side is under vacuum while it runs. The shell does not care either way. That difference in behaviour is enough to separate the three.

So measure the loss over a day with the pump on its normal schedule, then over a comparable day with the pump off entirely. A leak that gets dramatically worse with the pump running is on the pressure side: a return line, a fitting downstream of the filter, or the heater plumbing, where pressure forces water out through a gap that barely weeps at rest. A leak that is worse with the pump off is on the suction side, in the skimmer or main drain line, where running vacuum can pull air in rather than push water out. That is also why air in the pump basket, or bubbles blowing from the returns, is a suction-side symptom worth chasing even before the water level confirms anything.

Where the water is actually going

One more free clue is worth collecting before you do anything physical: let the pool keep dropping and note where it stops. Water only escapes down to the level of the hole. A pool that falls to just below the skimmer mouth and then holds is almost certainly leaking at the skimmer throat or its plumbing connection. A pool that stops an inch under a light is leaking at the niche or its conduit. A pool that keeps going past everything and empties toward the main drain is leaking in the floor or the main drain line. Owners find this frustrating because it means deliberately not adding water for a few days, but it converts a whole-pool search into a one-foot search.

The equipment pad leaks you can see

A surprising share of leaks never involve the pool at all, and they are the ones you can find in five minutes standing on dry concrete with the pump running. Water escaping on the pad evaporates off the slab between visits, so there is often nothing to see by the time anyone looks, which is why this walk-round belongs at the start of the visit rather than at the end of it. Look for a trail rather than a puddle: a thin darkened line running down a housing, a rust stain under a union, a patch of slab that is damp when everything around it is dry.

The one worth naming on its own is the multiport valve on a sand or DE filter. When its internal gasket wears, water passes continuously from the filter to the waste port, so the pool loses water at a steady rate with the pump running and there is nothing whatsoever wrong with the pool. The tell is water trickling from the backwash line when the valve is set to filter. It is a cheap part, it is a common failure, and it has sent more than one technician crawling around a light niche for an afternoon.

Dye testing the shell

If the loss is the same with the pump running and stopped, the water is leaving through the pool itself, and dye is how you find where. The principle is simple: water moving through a gap will carry dye with it, and dye released next to a gap that is not moving water will just hang there and disperse. What makes the test work or fail is stillness. The pump must be off, long enough that the water has stopped drifting, and it needs to be a calm day. Any current at all moves the dye and you will read it as a leak that is not there.

Work the suspects in order of likelihood rather than sweeping the whole shell: the skimmer throat and its plumbing seam, light niches and their conduit, return and cleaner-line fittings, the main drain cover seam, steps and the corners where a step meets the wall, tile grout along the waterline, and any visible crack. Release a thin ribbon of dye an inch or two from the feature and watch it, without disturbing the water with your own arm. A leak draws it in as a definite thread. Anything ambiguous is not a result, and repeating it once the water has settled again is quicker than acting on a maybe.

Two limits are worth being honest with the customer about. Dye finds leaks you can reach and see, so a crack behind a light, under a bench or beneath the main drain cover can defeat it entirely. And a vinyl liner pool is a different problem: the water can travel a long way between the liner and the wall before it leaves, so the place where dye is drawn in and the place the pool is losing water may be the same spot or may be nowhere near each other.

When to stop and bring in a leak detection specialist

Everything above is within reach of a service technician with a bucket, a bottle of dye and some patience, and it resolves most calls. What comes next is not. Pressure testing a line means plugging it at both ends and pressurising it with a rig; finding the point of failure along a buried line means acoustic listening equipment and an operator who uses it weekly. Buying the gear to do that a handful of times a year is poor economics, and guessing at it with a shovel is worse, because an excavation in the wrong place is a cost the customer can see.

Quoting a leak call, and who pays for what

Leak work does not fit a maintenance visit and should never be absorbed into one. The diagnosis alone is billable time: a bucket test spans two visits, a pump-on against pump-off comparison spans two more, and a careful dye test is an hour on its own. Quote the investigation separately from any repair, and say plainly that the investigation may conclude the pool is not leaking, because that is a real and frequent outcome that the customer still has to pay for. How you price that hour is the same question as any other non-routine work, and the pool service pricing guide covers how to set it.

The awkward conversations are all about authority rather than money, and they are avoidable in advance. Who approves a repair, at what value, and what happens when a specialist has to be subcontracted, are all things a service agreement can settle before anybody needs them. If yours does not address repair approval and third-party work, what belongs in a pool service contract is the place to start. Water loss during a leak investigation is worth a line of its own, since the customer is being asked not to top the pool up for several days and will want to know why.

Fitting leak work into a route week

The operational difficulty with leaks is that the tests are separated by days, not minutes. A bucket test needs setting up on one visit and reading on the next; a pump-off comparison needs a schedule change the owner has to agree to; a dye test needs the pump to have been off long enough that you cannot simply arrive and start. None of that fits a route built on a fixed weekly circuit, which is why leak calls so often get half-done and restarted. The fix is to treat the investigation as a sequence of short scheduled tasks against the property rather than one job, so the next step is attached to the next visit whoever happens to make it.

It also pays to carry the kit so a suspected leak never becomes a return trip: a spare bucket, dye, a grease pencil, spare lid and valve o-rings and a multiport gasket for the filters you actually service. All of that is small and cheap, and how a service truck is laid out decides whether you reach for it or promise to come back. One more thing worth ruling out while you are at the pad: a filter overdue for service raises pressure, and higher pressure makes an existing pressure-side weep noticeably worse, so cleaning the filter properly occasionally makes a "leak" much smaller on its own.

The thread running through all of it is that a leak is found by comparison across days, and comparison needs records. Water level, the weather that day, whether the pump ran, what was tested and what the result was all belong against the property rather than in somebody's memory of last Tuesday, particularly once more than one person services the route. Keeping that note at the stop rather than writing it up later is the difference between a diagnosis and starting again, and the app technicians use at the stop is where it belongs.