Update as of December 2023.

After trying a variety of techniques - eventually the leak was identified by finding a patch of dampness on a path, which evaporated early in the morning, but which had no reason to be there. Just a slight dampness - and this was after accidentally leaving the system on all night. I then dug up the path and did find a leak. Another clue had been a dampness on dirt slope several meters away - which had strongly suggested to me the approximate location (to with perhaps 10m) of the leak.

I could not access radioactive tracers. Ground penetrating radar was through the roof expensive and several companies refused to turn up. This was similar to professional irrigation companies. The idea of the compressor was no good because the system had too many points where enough air could leak out, and real leak was one meter below the ground below a paved area - I doubt I would have seen or heard anything specific enough. There were not enough valves, and putting in new valves was limited because most of the exciting stuff was burried deap, and once underground, there was no guessing which way the pipes went. Moisture detectors did not work because there was moisture everywhere, and the leak was not a sufficiently specific point of moisture in the soil.

But, the real answer was ...

  1. I replaced the entire system and
  2. I put in extra valves

I also made sure that ALL underground elements of the system were straight and obvious lines so that the location of the pipe was obvious. This later saved me when there was actually a problem in the new system. A five minute walk showed me where the leak was immediately.

Do not try to cut corners by puting lines diagonally accross lawns or by minimizing the number of valves. I also suggest that in practice it is far easier to put in extra valves so that the pressure can be adjusted than trying to be fancy with the layout. It is a false economy and just trying to show off to get the pressure from the layout - and it is not always possible ... due to the topography. It makes things cheaper for the installer, and a lot worse for the user.

There is a large amount of pipe of mixed materials buried on my two-hectare hobby farm by the previous owner. I do not know exactly how much or of what kind. Recently, a large leak appeared as detected by the water meter spinning rapidly which I know it was not doing earlier.

My question is - how to detect a hole in a pipe from the surface?

So far the ideas are - radioactive tracer, dye, air, smoke, and a stethoscope. I will try the stethoscope. The smoke idea sound interesting, but is a larger project that I will try to organize.

My question is not - how to isolate a leak by turning valves on and off, or checking out places where heavy equipment was, and so on. I am looking for a method of physically detecting a leak, which might be under 100 meters of loose rock retaining wall.

I Include below the question as originally posed and some more details.

For the record, if you wish to get more involved.

I have two hectares of hobby farm, most of which is irrigated. Recently, I found that while the irrigation system is off there is still water coming through at about the rate of an open domestic tap or shower. I can shut off the valve to the exterior irrigation, and the flow stops. So, the problem is there and not in the house or gardens.

Note: I greatly appreciate all the efforts from contributors so far and hope that together we can solve this. I will post the pathology if I find out what it is.

With that much water flowing I would have thought that the location of the leak would be obvious. But, I have walked the whole property looking for gushers or swamps and not found any.

How could this much water not be obvious on the surface?

How can this problem be solved short of digging up the entire irrigation system or replacing it?

This feels like it would make an great lateral thinking puzzle, but I don't know the answer yet!

I am wondering about trying colored smoke pumped in with an air compressor.

More details.

I replaced several valves that were jammed open and closed them - but it made no difference. I repaired a small leak discovered through a damp spot on the ground. But it was small and this did not noticeably change the problem flow.

For the record - I am an engineer and computer scientist so, believe me, I have tried all the tree pruning tricks of the trade. Every valve is turned off but the main one - with no luck. I am about to go out and try cutting into pipes and insert new valves so I can shut off other parts of the system.

The reason that naïve tree pruning does not work in this case is - from a computer science point of view - that we have only partial information. We have some of the node and some of the edges but we do not know how they are connected up.

Consider the following - I am going to put a valve in the line that goes past the tap. I wanted to put it in upstream of the tap. When I dug up around the tap it turns out that the line does feed the tap, but just goes past it. The feed for the tap looks like it either goes under 20 meters of retaining wall, or plunged 2 meters down to the lower level.

The soil here was unexpectedly damp - suggesting that something might be happening in the next level up.

More digging showed that the line to the tap goes under the retaining wall, and there is a second line that I did not even suspect that is parallel to it. Again - emphasizing why turning on and off of valves is not a viable process to locate the leak.

a post by a pipe near a wall

  • 1
    Binary search…. Commented Feb 13, 2022 at 23:47
  • 1
    If you know where the lines are, a Mechanic's Stethoscope might help locate the leak. Leak should make some noise you could hear within a few feet.
    – crip659
    Commented Feb 13, 2022 at 23:56
  • 3
    If I were to tackle this problem, I'd try turning off the water and pressurize the system with an air compressor (hopefully you have a fitting to do that). It should be pretty noisy where the break is. ...just a thought. Commented Feb 14, 2022 at 0:29
  • 2
    How do you know there is a leak and how do you know the flow rate? "Like a shower" ... ??
    – jay613
    Commented Feb 14, 2022 at 5:01
  • 1
    @jay613 I measured the rate for a shower by observing the meter spin while the shower was on. With the shower off there is no water flow. With the irrigation valve open it spins at the same rate. The value is approximate. It is not a dribble but a clear flow of water. That is as far as I took that half of the issue. Commented Feb 14, 2022 at 6:09

4 Answers 4


Based on everything you've written here's what I would do next:

  1. Map the property with all known irrigation heads and valves, and draw likely paths for hoses to follow between them.
  2. Using that map, use a narrow trenching shovel to dig in the vicinity of where you think the hoses are, nearest to the soak well that you mentioned in a comment. I don't propose you dig up everything .... but that's a very strong clue as to why the water isn't percolating upwards. Go as far with that as you are willing ... stop when you feel you are no longer following clear clues like the soak well. You don't have to find the hoses, you're just looking for wet soil below the surface.
  3. Next, get a large compressor, like the ones used to blow out these systems for winter, and run it through the central part of the system with all zone valves closed. Do this on a quiet windless day. While leaking water may find its way down and away, rapidly leaking air will look for an escape and will find its way to the surface. Hopefully. And hopefully, you'll be able to hear it if you walk around the property. Use your map to determine roughly the areas where the unzoned feed hoses may be and focus on those areas.
  4. Getting a little desperate now. Drive around the property in a car or preferably a tractor or ATV. The soaked soil, even if it's well below the surface, should give under the weight of a car. Start by following the imagined hoses on your map, first near the soak well, then elsewhere, and eventually cover the whole property. We aware of the feel of the earth beneath you as you drive.
  • What about the idea of using colored smoke rather than just air? Commented Feb 15, 2022 at 22:48
  • 1
    1) I think the soil will filter the smoke but some smokeless air will hopefully make it to the surface with enough pressure to be heard. 2) I don't know if any compressor owner will let you put smoke or smoke-like powder through their machine, I doubt it's healthy for it, and I don't know how you'd introduce the smoke to the system anyway. Smoke generated in large volumes and delivered continuously at high pressure sounds, to me, like fun but no more practical than a radioactive "Fantastic Voyage"-style probe. Ok, a little more practical. :)
    – jay613
    Commented Feb 15, 2022 at 22:57

I have two hectares of hobby farm.

All of this is irrigated?

Recently I found that while the irrigation system is off there is still water coming through at about the rate of an open domestic tap or shower.

How many linear feet of water line is this? How many joints, valves, stems, etc...?

With that much water flowing I would have thought that the location of the leak would be obvious. But, I have walked the whole property looking for gushers or swamps and not found any.

How could this much water not be obvious on the surface?

Don't assume it's all leaking from a single point of failure. You could be dealing with 50 small leaks. It will be nice if this is just a single point of failure.

Did you let your irrigation system freeze over the winter?

Has any heavy machinery ran over your irrigation system?

When the system is irrigating do you notice worse flow at certain heads versus others?

How can this problem be solved short of digging up the entire irrigation system or replacing it?

You could add tracing dye inline to your irrigation system for a few days, leave it pressurized, and see if anything comes to the surface.

You'll obviously want to avoid running the irrigation system while the dye is being added.

  • I take it you did not read the question. I bought the property with the irrigation installed and buried, and do not know exactly how many feet of pipe there are, nor how many joints, nor even how many valves. Hence my interest in methods to detect the leak -- I intended that to mean fairly directly. I do know about things like turning valves on and off and checking pressure to find a leak. See in particular the details section of the question. The leak occurred suddenly a few days ago, so it is unlikely to be a large number of small leaks. Commented Feb 14, 2022 at 21:36
  • If my current conjecture is correct - the leak is under a loose-stone retaining wall, a meter under the ground, and near a soak well. That would give a reason why it was hard to find. Most likely dye would not work for this. Also in the post, I currently mention that the ground at one end of the retaining wall is unexpectedly damp. Just damp, but I would have expected it to be very dry at this time of year. Commented Feb 14, 2022 at 21:41
  • @PonderStibbons Yes, your question is quite arduous to read. When I answered it was not clear that you did not install this system. Anyways, your question can be boiled down to "Irrigation leak detection methods?" and the execution of each method would be up to you. There's no way for anyone here to say "Oh yeah, that's a leak in run 34.2, you should step 3 feet to your left and dig." Anyways, hopefully you've gained some new ideas from me because I'm not digging through your 7 large edits to figure out what's new to you and what isn't.
    – MonkeyZeus
    Commented Feb 14, 2022 at 23:31
  • 1
    get a friend to check for swampy vegetation, I have found leaks this way. Here we look for Juncus or Cyperus. Commented Feb 15, 2022 at 7:26
  • 1
    Detecting leaks depends a LOT upon soil conditions. If very loamy or gravelly, that water could just sink unnoticed. If hard pack or clay, yeah, it'd surface someplace. Not knowing the details of the system, I'd guess it's a failed connection. They do fail. I had one fail when my wife & I were on vacation and wondered why my power bill was so high when we got back, turned out there was a seriously failed connection and the pump was running constantly, flooding a garden area. DOH! Commented Feb 15, 2022 at 20:44

Idea no 1: The farmers way. Identify an suspicious pipe. Place an shutoff valve onto the suspected pipe. Repeat until the water flow stops.

Idea no 2: Locate the affected section with an clamp-on ultrasonic flowmeter instead of randomly placing shutoff valves.

Idea no 3: Check with an infrared imaging system for leaks. Expensive though.

  • Someone I was talking to mentioned the idea of using a hand held IR thermometer, which did not sound useful. But, I did wonder about a 2D infrared scan. Maybe just an IR camera, now that I think of it. Commented Feb 15, 2022 at 22:49
  • I have been putting shutoff valves into pretty much every pipe I can find! But the flow meter requires you to find the pipe. Commented Feb 15, 2022 at 22:51

I take care of a sprinkler system at a church; over a thousand feet of buried plastic pipe. Most commonly I find a wet spot at a leak when it is running. Alternatively ,in dry weather, after the sprinkler has run, I walk the line and look for damp ground or very lush grass. I apologize that this is not very scientific, but it works in my heavy clay soil with poor permeability. This will involve digging for verification. I find a 3/8 in. diameter steel rod with a good handle is necessary to probe for the exact position of the pipe.

  • Yes, I have been looking for that but could not find anything. This is why I have resorted to asking here. Thanks. Commented Feb 16, 2022 at 2:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.