GPR Depth Accuracy: Ground Penetrating Radar Depth Guide

Updated: 23 October 2026

What does GPR depth accuracy actually mean?

The first thing to grasp about GPR depth accuracy is that a ground penetrating radar (GPR) unit never reads depth off a scale. It sends a short radar signal, a burst of radio waves, into the ground and records how many nanoseconds pass before an echo returns from a cable, pipe or duct. That figure, the two-way travel time, becomes a depth only after it is multiplied by the speed of the wave through that particular soil and halved.

So the accuracy question is really a question about velocity. If the speed used in the calculation matches the ground, the depth is close. If it does not, every depth on the line is wrong by the same proportion. Two further limits sit on top of that:

• Vertical resolution – how finely the antenna can separate two targets stacked one above the other; it sharpens as antenna frequency rises.
• Detectability – whether the target returns a clear echo at all, which depends on its size, material and the soil around it.

A depth figure is therefore always an estimate with a margin, and a good report says how that margin was controlled.

Why does a depth error matter on site?

Plan position tells an excavator where a service runs; depth tells them how close the bucket or drill head can safely come. An HDD bore planned to pass below a water main, or a trench cut to a set level beside an HT cable, both rely on that vertical number.

Consider the arithmetic. Suppose the true wave speed in the clay soils of a road verge is 20% slower than the value assumed in the software. A duct actually lying 1.0 m down would then be reported at about 1.25 m. A designer trusting that figure might set a bore to pass just above it and strike it. The reverse error, reporting a service shallower than it is, is just as risky when a bore is planned to pass beneath it. Because either mistake can put a drill head into a live line, depth is treated with more caution than plan position.

Which methods produce a GPR depth figure?

Several techniques are used, sometimes together on one route.

Hyperbola fitting

When the antenna passes over a narrow target at right angles, the echo forms an inverted U on the radargram. The curvature of that shape depends on wave speed, so fitting a curve to it gives a velocity for the soil above the target. This is the most common field method and needs no digging.

Known-depth calibration

Where a manhole, an exposed pipe or a trial pit shows a service at a measured depth, the travel time to that service fixes the velocity directly. It is the most dependable check available.

Multi-offset soundings

Some systems can move transmitter and receiver apart in steps over one spot, a common-midpoint or wide-angle sounding. Velocity is read from how travel time grows with separation. It suits open, level ground more than busy streets.

Depth slices in 3D surveys

A dense 3D grid can be cut into horizontal slices at chosen depths. Such grids image the subsurface in plan view, but the slice depths still rest on the velocity chosen, so the same calibration rules apply.

EM locator depth readings

An electromagnetic locator estimates depth from the shape of the magnetic field around a conductive line. Nearby cables, bends and tee joints distort that field, so its readings are compared with the radar result rather than taken alone. Our guide on EM locators versus GPR explains the difference.

Which terms should you know before reading a depth report?

• Two-way travel time – nanoseconds for the pulse to reach a target and return.
• Velocity – speed of the wave in the ground, usually stated in metres per nanosecond. In air it is about 0.3 m/ns; in soil it is always slower.
• Dielectric permittivity – one of the dielectric properties of the ground, and the one that sets the wave speed. Water raises it sharply, which is why moisture content matters so much: a wet verge and a dry one beside it can give quite different speeds.
• Time-zero – the point on the trace taken as the ground surface. If it is set wrongly, every depth shifts up or down by a fixed amount.
• Apex – the top of a hyperbola, which marks the top of a pipe or cable, not its centre.
• Penetration depth – how far down usable echoes were obtained on a section; poor depth penetration is common in wet clay. Signal attenuation, the loss of GPR signal strength in conductive ground such as wet clay, limits it; that is a separate matter from how accurate the depths within that range are.
• Surface reference – the ground level the depth was measured from on the survey date.

For help reading the traces themselves, see how to read a GPR radargram.

What makes a depth reading trustworthy?

Not every number on a drawing deserves equal confidence. Readings carry more weight when:

• velocity was calibrated on that site, not left at an instrument default;
• the scan line crossed the target close to a right angle, giving a clean, symmetric hyperbola;
• the target is large compared with the antenna's resolution, such as a pipe rather than a thin earthing strip; metallic utilities also return stronger echoes than plastic ones;
• soil above it is reasonably uniform, without a wet layer or rubble fill distorting the wave path;
• radar and EM readings agree within a sensible margin;
• at least one service on the route was exposed and measured to check the conversion.

Readings with fewer of these supports should be flagged as lower confidence in the report.

How is depth calibrated and checked, step by step?

1. Set time-zero at the start of each data file so the surface is correctly placed on the trace.
2. Choose the antenna frequency for the job. Higher frequencies from a high-frequency antenna give higher resolution and sharper, high-resolution images of shallow, closely spaced services but loses signal sooner; lower frequencies see further at the cost of detail.
3. Find calibration targets early: hyperbolas from crossing services, visible chambers, or a pipe in an open trench nearby.
4. Derive velocity for each distinct ground type on the route, because a town road, a field edge and a canal bank can behave very differently.
5. Pick the apex of each target and convert travel time to depth using the velocity for that section.
6. Cross-check conductive lines with an EM locator and note any disagreement.
7. Prove key points with trial pits or boreholes where a bore or deep cut passes close to a service. Our comparison of GPR and trial pits covers when this pays off.
8. Report depth by section, stating the velocity used, depth reach achieved and confidence for each service.

Which mistakes make ground penetrating radar depth readings wrong?

• Trusting the factory velocity. Instrument defaults suit average soil conditions, which a real-world site rarely matches.
• Using one velocity for a whole corridor. Moisture and soil type change along a route; so should the calibration.
• Reading depth from an oblique pass. A line crossing a pipe at a shallow angle stretches the hyperbola and inflates the computed velocity.
• Confusing top and centre. For a large pipe the difference between crown and centreline can be significant.
• Forgetting the surface has changed. A road resurfaced or a verge filled after the survey adds cover that the old figure does not show.
• Treating silence as absence. No echo can mean a deep or small target in lossy ground, where attenuation swallows the return, not an empty corridor free of underground utilities. The soil limits guide explains why.

How should depth records be kept after the survey?

A depth figure loses value quickly if nobody can tell how it was produced. Keep, along with the drawing or report:

• the raw radar files and the velocity applied to each section;
• the survey date and notes on weather and ground moisture;
• the surface reference used, ideally tied to fixed features such as kerbs or chamber covers;
• locations and measured depths of every trial pit or exposed service;
• a confidence grade per service, in line with the SUE quality levels where the client uses them.

When the next contractor works on the same road, these records show which depths were proved and which were estimated.

Mezux Touch runs route and site surveys on 58 company-owned GPR systems with in-house crews, and each report states the depth reach achieved on every section. Service details are on our GPR survey page, and current rates, including the ₹32 per metre L-section that charts service depths along a route, are on the GPR survey price list.