
Articles & Advice · 26 March 2026
How Accurate Is Ground Penetrating Radar in New Zealand Soil Conditions?
Ground penetrating radar is the tool people ask for when they want to see everything under a site, and it is the one most often oversold. How accurate it is depends almost entirely on what the ground is made of, and New Zealand ground varies from scoria that radar loves to saturated clay and peat that swallow the signal within half a metre. This guide sets out what GPR can realistically find in the soils around Auckland, Franklin, the Waikato and the Hauraki Plains, how accurate the positions and depths are, and how it fits with the other locating methods.

The honest answer to "how accurate is GPR" is: very, in the right ground and the right hands, and barely useful in the wrong ground. That is not a criticism of the technology. It is physics, and knowing where the limits sit is what separates a locate you can dig to from a pretty picture. Utility Works runs GPR alongside electromagnetic locating, sondes and hydro excavation on sites from the North Shore to Hamilton and across to Thames, so what follows comes from scanning this region's ground week in, week out.
How GPR works, briefly
A GPR unit sends short pulses of radio energy into the ground from an antenna and listens for the echoes. Wherever the pulse meets a change in material, from soil to a pipe, from soil to air in a void, from fill to natural ground, part of it reflects back. The unit records the time taken for each echo and, as it is pushed along the surface, builds a profile. A pipe crossing the scan line shows as a hyperbola, an inverted U, and the operator marks its apex on the ground. Because it responds to any change in material, GPR finds plastic pipe, concrete, timber, voids and tanks that electromagnetic locators cannot see, and it does so without putting a signal onto anything.
Two settings govern what you get. Lower frequency antennas, around 200 to 400 MHz, penetrate deeper but resolve less detail. Higher frequencies, around 900 MHz and up, show fine detail such as rebar and small conduits but only in the top half metre or so. Utility locating work is mostly done at 400 MHz or with dual-frequency units that record both at once.
What governs accuracy
Soil conductivity and moisture
Radar energy is absorbed by electrically conductive ground, and the more conductive the soil, the shallower the signal dies. Conductivity rises with clay content, with moisture and with salt. Dry sand, gravel, scoria and pumice are nearly transparent to radar. Wet clay is close to opaque. This single factor matters more than the equipment, the operator or anything else.
Depth
Every reflection gets weaker with depth, so a 100 mm PE water main that is obvious at 600 mm can vanish at 1.5 m in the same soil. Services in New Zealand are typically laid between 450 mm and 1.2 m, which is comfortably within GPR range in good ground and right at the limit in poor ground.
Target size and material
Larger targets reflect more. A 150 mm main is far easier to find than a 25 mm plastic lateral. Metal and water-filled pipes reflect strongly; empty plastic ducts and small conduits reflect weakly and can hide inside the clutter of a disturbed trench.
Surface and clutter
Reinforced concrete slabs return a dense pattern from the rebar that masks what lies beneath. Compacted road base, rubble fill and tree roots all scatter the signal. A smooth lawn or asphalt footpath is the easiest surface to scan.
Orientation
GPR sees a pipe clearly when the scan line crosses it at right angles and poorly when the pipe runs parallel to the scan line. That is why a proper locate is a grid, not a single pass, and why a pipe that runs directly below another one is routinely missed.
How New Zealand soils behave under radar
Our area has an unusually wide spread of ground types within a short drive, and GPR performance changes with them.
| Ground type | Where you meet it | Typical useful GPR depth | Notes |
|---|---|---|---|
| Scoria, basalt ash and volcanic tuff | Central Auckland isthmus, Mangere, Otahuhu, Mt Wellington, Pukekohe Hill | 2 to 4 m when dry | Excellent penetration. Basalt boulders ("floaters") show as strong targets and can be mistaken for pipes |
| Pukekohe and Bombay volcanic loam | Pukekohe, Bombay, Patumahoe, Pukekawa | 1.5 to 2.5 m dry, around 1 m wet | Good when dry. Holds moisture after rain, which drops penetration noticeably |
| Waitemata Group clays and weathered siltstone | Much of urban Auckland, North Shore, Papakura, Drury, Karaka | 0.8 to 1.5 m, under 0.5 m when saturated | The common case. Services at standard depth are usually found; deep or small plastic pipes often are not |
| Heavy Franklin and North Waikato clays | Pokeno, Tuakau, Mercer, Te Kauwhata, rural Franklin | 0.6 to 1.2 m, less in winter | Winter scanning is markedly worse than summer on the same site |
| Peat and organic soils with high groundwater | Hauraki Plains, Whangamarino, parts of Huntly and Ngatea | Often under 0.5 m | Saturated and conductive. GPR is of limited use; electromagnetic methods and potholing carry the job |
| Coastal sand and reclaimed fill | Beach settlements, port and industrial reclamations | Sand 2 to 3 m; saline fill often under 0.5 m | Dry sand is ideal. Anything with salt water in it kills the signal |
| Engineered road base and compacted fill | Any road corridor or developed site | Variable | Clutter rather than absorption is the problem; old trenches show up, their contents may not |
The practical consequence is that a site in Mangere and a site in Ngatea with identical services can produce completely different radar results on the same day with the same operator. That is why we never quote GPR as a standalone product for a site we have not seen.
How accurate are the positions and depths
Horizontal position. Where the operator has a clear hyperbola and marks its apex, the horizontal position is typically within 100 to 150 mm of the pipe's true centreline. On a disturbed or cluttered site the operator will widen that tolerance and should say so on the mark-out.
Depth. GPR measures time, not distance, and converts one to the other using an assumed velocity for the ground. If that assumption is off, depth is off, and the error is typically 10 to 20 percent. The fix is calibration: expose one known target by potholing, measure its real depth, set the velocity to match, and depth accuracy on the rest of the site tightens to around 50 mm. A GPR depth that has not been calibrated against something seen should be treated as an estimate.
Identification. This is the limit people forget. GPR shows that something is there, roughly how big it is and how deep. It cannot tell you whether the target is a water main, a gas main, a comms duct or an abandoned pipe. Identification comes from records, from electromagnetic tracing of anything conductive, from sondes in ducts and drains, and ultimately from exposing it.
What GPR will and will not find
In the ground described above, GPR reliably finds:
- Water and wastewater mains, metallic or plastic, at standard depth in dry to moderately moist soil
- Concrete and clay drainage, chambers and culverts
- Septic tanks, old fuel tanks, soak pits and voids under slabs
- Ducts and conduits with nothing conductive in them, which electromagnetic locators cannot see at all
- Old trench lines, which tell you where to look even when the pipe itself is faint
- Rebar, post-tension cables and slab thickness ahead of coring
It struggles with, or misses:
- Small-diameter plastic pipe below about a metre in clay
- A pipe laid directly beneath another one
- Anything under heavily reinforced concrete or dense rubble fill
- Services in saturated peat, saline fill or wet heavy clay
- Telling one buried object from another without supporting methods
Why GPR is one tool, not the locate
The Australian and New Zealand standard AS 5488 classifies subsurface utility information by quality level. Records such as beforeUdig plans are QL-D. Surface features surveyed are QL-C. Services located by geophysical methods, which is where GPR and electromagnetic locating sit, are QL-B. Services physically exposed and measured are QL-A. A defensible locate for excavation uses all of them: plans to know what to look for, electromagnetic locating to trace and identify everything conductive, GPR to find what has no signal and to fill gaps, and potholing to confirm anything that will be crossed or worked beside. We explain the records side in our guide to what beforeUdig gives you and what it doesn't, and the exposure side in what hydro vac potholing is.
The operator matters as much as the method. Reading a radargram on a cluttered Auckland berm, deciding whether a hyperbola is a pipe or a boulder, and knowing that the strong target at 400 mm is the abandoned asbestos cement main and the faint one beside it is the live PE replacement, is experience, not software. Our technicians lay pipe and cable as well as find it, which is why they recognise how services were installed, not just that something is there.
Two jobs from this year show the combination at work. In September 2026, at one of Auckland's most critical operations, other contractors had spent months failing to find a leak; our crew used aerial infrared, GPR and acoustic equipment together, traced the pipeline, pinpointed the leak and had it excavated, repaired and reinstated within days. In May 2026, live gas pipes two metres down in boulder-strewn volcanic overburden were detected and traced in from over 100 metres away by electromagnetic methods before a careful benched excavation reached them, on a site where no hydrovac could get in. Radar found what it could; tracing identified it; the dig confirmed it.
When GPR is worth paying for
- Plans are missing, old or obviously wrong
- Non-metallic services are known or likely: plastic water, drainage, empty ducts
- You are about to drill, pile, core or directionally bore and a strike would be serious
- You need to find a tank, a void or an old foundation rather than a pipe
- You are designing a new service route through a congested corridor
- You want a survey-grade map of a site for a subdivision, campus or asset register
It is less worthwhile as a stand-alone scan of saturated clay or peat in winter, where an honest operator will tell you that electromagnetic locating and potholing will carry the job and GPR will add little.
What a GPR survey costs
Leak detection and service detection jobs can start from $300 +GST. Every job is different, so call 0800 779 588 and we'll talk through your situation and give you an estimate. What moves a GPR job is the area to grid, the ground on the day, whether you need survey-grade GPS capture and a drawing, and how many targets need potholing to calibrate depth and confirm identity. Tell us the site and we will tell you what the ground there will let radar do before you book it.
Working with Utility Works
We bring GPR, electromagnetic locators, sondes and the hydro excavation truck to the same site, so a target found by radar can be traced, identified and exposed in one visit rather than three. We have been mapping one network provider's assets throughout its territory for over ten years, substations included, and on a single South Auckland mapping project in 2025 tracked and traced more than 200 individual pipes, cables, ducts and drains. The mark-out tells you how confident we are in each line and which depths have been calibrated against a pothole. If you need the result on a plan, we capture everything by survey-grade GPS and supply the drawing or shapefile. The service itself is on our ground penetrating radar investigations page, and sits inside our wider underground service detection and mapping work across Auckland, Franklin, North Waikato and the Hauraki Plains.
Frequently Asked Questions
Can GPR detect plastic pipes?
Yes, and it is the main method for doing so when the pipe has no tracer wire. Detection depends on pipe size, depth and ground. A 100 mm PE main at 700 mm in dry Auckland clay is routinely found; a 20 mm lateral at 1.2 m in wet clay often is not.
Does clay stop GPR from working?
Clay reduces penetration rather than stopping it. In the Waitemata and Franklin clays around Auckland, useful depth is typically 0.8 to 1.5 m when the ground is dry and under half a metre when saturated. Services at standard depth are usually found; deep or small targets may not be, and winter is noticeably worse than summer.
How deep can GPR see in New Zealand?
From under half a metre in saturated peat or saline fill to 3 or 4 metres in dry scoria and sand. Most of the region's clay soils give 0.6 to 1.5 m, which covers typical service depths.
How accurate is GPR depth?
Uncalibrated, typically within 10 to 20 percent. Once one target has been exposed by potholing and the ground velocity set to match, depth accuracy on the rest of the site tightens to around 50 mm. Treat any uncalibrated GPR depth as an estimate.
Is GPR better than beforeUdig plans?
They do different jobs. Plans tell you what registered asset owners recorded, with no depth and no private services. GPR shows what is actually in the ground on the day, including unrecorded and non-metallic services, but cannot identify what each target is. A proper locate uses both, plus electromagnetic tracing and potholing.
Should GPR be used before directional drilling or piling?
Yes. Drilling and piling are blind, so the consequences of a strike are serious. A gridded GPR and electromagnetic locate with potholes at every crossing is the minimum we recommend before a bore or a pile goes in.
Can GPR tell the difference between a water pipe and a gas pipe?
No. GPR shows size, depth and position, not identity. Identification comes from records, electromagnetic tracing of conductive pipes, sondes in non-metallic ones and, where it matters, exposing the service. We locate gas services as part of every site sweep so they can be avoided.
Let's discuss your underground project needs. Contact us today for a free, no-obligation quote.
Free, no-obligation quote · Mon–Fri 7am–5pm · Auckland, Franklin & North Waikato
