Utility Locating Accuracy: Horizontal and Depth Tolerance

What does utility locating accuracy mean?

Utility locating accuracy is two numbers. One is how far the true line may lie to either side of the mark, and the other is how far its real depth may differ from the reported one. A mark without a tolerance tells an excavator where to start looking, not where digging is safe. With ±150 mm in plan and ±15% in depth stated, a designer can decide how close a bore or bucket may come to underground utilities.

Plan position is fixed where the instrument responds most strongly, and good practice holds it to a few centimetres over an isolated line. Depth is calculated, never observed: ground penetrating radar times an echo and applies a velocity, and locator makers quote about ±5% only for a lone straight conductor. The vertical tolerance is therefore a percentage of depth. Where clearance to a gas main is critical, vacuum excavation replaces the estimate with a measurement.

What tolerances do the main standards set?

The UK's PAS 128 grades detected lines by plan and depth tolerance, and the US ASCE 38-22 sets how tightly exposed utilities must be tied to the datum. India has no national tolerance standard, so contracts borrow one of these. The table shows the four detection levels of the 2014 edition; the April 2022 revision kept the quality level approach, so check which edition your contract names.

These limits assume a line found by crossings at regular spacing. A dense grid reduces interpolation error between crossings; to judge whether that is worth the extra field time, read 3D GPR vs 2D GPR.

ASCE 38-22 puts no number on Quality Level B; it asks the engineer to record the confidence of each segment. Its companion ASCE 75-22 asks for recorded accuracy at 95% confidence, with units and datum stated.

Where do location errors come from?

Each of four stages adds its share: the instrument reading, the interpretation, survey positioning and the transfer of a mark onto the ground.

• Peak picking: the apex of a radar arch marks the pipe top, and a locator peak marks the field centre; overlapping arches and distorted fields shift both.
• Field distortion: bends, tees and return currents on nearby conductors push a locator peak sideways.
• Wave speed and reference: a velocity error scales radar depth by the same percentage, and crown versus centre differs by half the diameter, 150 mm on a 300 mm main.
• Positioning and interpolation: RTK GNSS holds about 1–2 cm in open sky, and a line drawn between crossings 5 m apart assumes the service runs straight.

Indian streets add their own errors

Repeated resurfacing raises city roads by several centimetres each time, so depth must be quoted from the level on the survey date. Shared trenches put power, telecom and water lines from different years inside one radar footprint. Interlocking pavers over concrete add a strong shallow echo that masks small services, so tolerance is best stated per section, not once for the whole job.

How big is a realistic error budget?

Take a single cable at 1.2 m, found by radar and locator and crossed every 5 m. A fair estimate is about ±0.12 m in plan and ±0.13 m in depth. Independent errors add by root-sum-square, not simple addition:

• Plan: apex pick ±50 mm, GNSS ±20 mm, paint mark ±30 mm, wander between crossings ±100 mm. √(50² + 20² + 30² + 100²) ≈ 117 mm.
• Depth: velocity uncertainty ±10% of 1.2 m = ±120 mm, time-zero and surface level ±30 mm, apex pick ±30 mm. √(120² + 30² + 30²) ≈ 127 mm, close to ±11% of depth.

In plan, the biggest term is the gap between crossings, so closer scan spacing pays more than better GPS. In depth, velocity dominates, which makes calibration against a known target essential; to see how velocity is checked, see: GPR Depth Accuracy.

How should a tolerance be written into a survey contract?

State the level required for each class of utility, how it will be checked, and what follows a failed check. A clause such as "accurate utility survey" leaves both sides arguing later. A workable specification covers:

1. Level by utility: for example QL-B1 for gas, trunk water mains and HT cables, and QL-B2 for low-voltage and telecom lines.
2. Depth reference and datum: crown or centreline from ground level on the survey date, in a named coordinate and height system.
3. Verification: trial pits at a sample of locations, such as every crossing of the new works, with measured offsets.
4. Evidence: radargram snapshots at doubtful spots; our Mezux guide: GPR Radargram Library shows the images we attach.
5. Non-conformance: a line found outside tolerance is rescanned, regraded and redrawn, which keeps the utility data reliable for the long term.

Grading lines this way is the core of subsurface utility engineering and helps reduce risks of service disruptions; for the full framework, see a primer on SUE survey.

What does an accuracy-graded survey cost?

The rate card is per metre and depends on the drawing, not the tolerance band: ₹24 per metre standard and ₹32 per metre with an L-section, GST extra. The minimum order is ₹50,000 within Delhi NCR and ₹1,00,000 in the rest of India. Long programmes earn 25% off the whole order beyond 100 km, 40% beyond 500 km and 50% beyond 1,000 km. QL-A trial pits are priced separately. NCR sites are scanned the same day or within 1 working day. Further afield, the surveyor living in that region starts at once or within 1 working day; any mobilisation or travel appears as its own quote line. To discuss the tolerance your job needs, share the site location and the design clearance you are working to.

Sources

1. BSI, PAS 128:2022 “Underground utility detection, verification and location. Specification” (scope and changes since 2014). knowledge.bsigroup.com
2. GoGeomatics Canada, “Assessing the PAS 128 quality standard for underground utility location” (QL-B1 to B4 tolerance table). gogeomatics.ca
3. American Society of Civil Engineers, “Newly updated ASCE 38-22 utility engineering standard and new companion standard ASCE 75-22 now available”, July 2022. www.asce.org
4. GoGeomatics Canada, “Quick overview of ASCE 75-22 ‘As-installed’ standard” (QL-A tie-in of 60 mm horizontal and 30 mm vertical; 95% confidence reporting). gogeomatics.ca

Frequently asked questions

How accurate is a GPR utility survey?

In plan, within 0.1 to 0.25 m on an isolated line with regular crossings. Depth is often within 10 to 20% once wave speed is calibrated, and worse in wet or mixed ground.

What does ±15% depth tolerance mean?

A service reported at 1.0 m may lie anywhere from 0.85 to 1.15 m below the surface.

Is there an Indian standard for utility locating accuracy?

No national standard sets tolerances, so Indian contracts borrow PAS 128 or ASCE 38 levels and add their own verification clause.

Does better GPS make the survey more accurate?

Only up to a point. RTK positioning is already good to a couple of centimetres; scan spacing and interpretation matter more.

Why do two methods give a better grade?

A radar arch and a locator peak in the same place rarely agree by chance. Agreement narrows the error and confirms the line exists.

Where are QL, crown and time-zero defined?

You will find details on GPR, SUE & Land Survey Glossary terms on our glossary page, each with a short example.

Written by Saniya Parveen · Technically reviewed by Rahul Rai · Updated 8 November 2026