8 min read1,836 words

GNSS Accuracy Requirements for Utility Field Capture

Positional accuracy in a utility estate is not a single number, and treating it as one produces two opposite failures at the same time: crews carrying survey-grade equipment to photograph poles, and buried tees located with a phone. The useful specification is per asset class, derived from what the position has to support — finding a valve during an isolation, avoiding a main during an excavation, identifying which pole an inspection refers to. This guide sets out how to derive those budgets, which correction method meets each, which error sources corrections actually remove, and what instruction has to reach the crew so that an unsuitable site produces a record rather than a bad position. It supplies the accuracy half of the capture pipeline described in field data capture and mobile sync.

Environment Prerequisites

  • A receiver capable of reporting its own accuracy estimate, and a capture application that stores that estimate with the point. Without it, no downstream validation is possible.
  • A correction service appropriate to the tightest class the crew will capture — a network real-time kinematic subscription for buried assets and control work.
  • A published accuracy budget per asset class, held with the schema so the staging validator can read it.
  • A recorded capture method per point — uncorrected, augmented, corrected, offset — because the method explains the accuracy when the estimate alone does not.
  • A defined offset procedure for sites that cannot support a direct observation, with its own accuracy expectation.
  • Known control — surveyed tie points or benchmarks in the authoritative frame, aligned per CRS alignment and geodetic transformations — so equipment can be verified rather than trusted.

Schema-Aware Validation Protocol — Run Before Setting a Budget

  1. Start from the use, not from the equipment. Ask what decision the position supports and how wrong it can be before that decision changes. A budget derived from what the crew already carries is a description, not a requirement.
  2. Separate horizontal from vertical. Depth of cover and invert elevation have their own budgets and their own error sources, and a horizontal specification says nothing about either.
  3. Check the frame before the accuracy. A centimetre-accurate position in the wrong realisation is metres wrong in the enterprise frame, which is why control verification precedes any accuracy claim.
  4. Verify the receiver against control, on a schedule. A receiver that has drifted or been misconfigured reports confident accuracies that are wrong, and only a known point reveals it.
  5. Confirm the reported accuracy is being stored. Many capture configurations discard it. A point with no accuracy attached cannot be validated, and it silently degrades the estate’s overall claim.

Deriving the Budgets

Four classes cover most utility estates, and stating each one’s justification is what makes the budget defensible when someone questions the equipment cost.

Buried assets — mains, tees, service connections, valves below grade. The position must allow an excavation to find the asset without striking it, and it must survive being the reference for everything else surveyed nearby. Two to five centimetres, real-time corrected.

Surface devices and structures — poles, cabinets, hydrants, above-grade valves. The position must dispatch a crew to the correct object and support isolation analysis. Submetre is adequate, and the practical constraint is usually that adjacent objects must not be confusable.

Presence and condition records — inspections, photographs, meter reads. The position only has to identify which asset the record belongs to, which an uncorrected receiver achieves whenever the assets are more than a few metres apart.

Control and tie points. These are the reference for everything else and need the tightest available accuracy, plus a documented occupation time and a verification against known control.

Positioning methods against what they cost and what they can be trusted for An uncorrected receiver is free and produces a position good enough to identify which asset is being described and nothing more. A satellite-based augmentation improves that to roughly a metre in the open. A network correction service brings a survey-grade receiver to a few centimetres in real time, which is what buried assets and control work need. Post-processing reaches the same accuracy without a live link, at the cost of not knowing in the field whether the observation succeeded — which matters more than it sounds, because a failed observation discovered next week means a second visit. Method Typical accuracy Field feedback Fit for Uncorrected receiver 3–10 m immediate identifying an asset SBAS-corrected ~1 m immediate presence and condition RTK network correction 2–5 cm immediate buried assets, control Post-processed 2–5 cm days later planned survey only Immediate feedback is worth as much as accuracy: a failed fix you know about is fixable.

What Corrections Do and Do Not Fix

A correction service removes the error sources shared between the receiver and a nearby reference station: satellite orbit and clock error, ionospheric and tropospheric delay. Those dominate an uncorrected position, which is why corrections improve accuracy by two orders of magnitude.

They do nothing about the two error sources a utility crew meets most often. Multipath — the signal arriving twice, once directly and once after bouncing off a substation wall, a truck or a building face — is local to the receiver’s position and is not shared with the reference station. Canopy attenuation is the same: no correction recovers a signal the receiver never heard.

The practical consequence is that a crew working beside a metal structure or under heavy canopy will see a degraded accuracy estimate that no amount of subscription fixes, and the correct response is a procedure rather than more equipment: move the antenna, extend the occupation, or use a documented offset from a clear observation. What must not happen is capturing the degraded position anyway because the form requires one.

The four error sources a utility crew actually meets, and which corrections remove them Satellite geometry and atmospheric delay are common-mode errors that a correction service removes almost entirely, because the reference station experiences the same conditions. Multipath — the signal arriving twice after bouncing off a substation wall or a truck — is local to the receiver and is not removed by any correction. Canopy attenuation is the same: no correction service can recover a signal the receiver never heard. The practical consequence is that corrections fix the errors a crew cannot see and leave the ones they can do something about. Satellites Atmosphere corrected Geometry corrected Multipath not corrected Canopy not corrected Receiver reported accuracy removed by correction local — move the receiver Corrections fix what the crew cannot see; the rest is site craft.

Vertical Accuracy and the Assets That Need It

Depth of cover, invert elevation and clearance all depend on a vertical position, and vertical accuracy from a satellite receiver is roughly half as good as horizontal accuracy from the same observation. That relationship is fixed by geometry — the satellites are all above the receiver, so the vertical component is poorly constrained — and no amount of correction changes it.

Three consequences follow for utility capture. A class whose vertical budget matters needs a tighter horizontal specification than its horizontal use alone would justify, simply to bring the vertical component inside its own budget. The vertical datum has to be stated: an ellipsoidal height and an orthometric height differ by tens of metres, and a depth computed against the wrong one is not slightly wrong. And where the vertical requirement is genuinely tight — a gravity main whose grade is the whole design — levelling from known control is still the right method, and recording that the position came from levelling rather than from a satellite is what lets a later reviewer trust it.

Production Deployment Pattern

  1. Publish the budget with the schema, not in a document. The validator reads it, the capture application displays it, and a single source keeps them consistent.
  2. Put the acceptance threshold on the device. A crew that sees “needs 0.5 m, currently 1.4 m” can act; one that discovers the rejection next week cannot.
  3. Specify the fallback explicitly. Offset from a clear observation, extended occupation, or return with different equipment — and make recording “site unsuitable” a valid, expected outcome rather than a failure.
  4. Verify receivers against control monthly. Log the check. A drifted receiver produces confident, wrong positions across every asset it touched since the last verification.
  5. Store the method alongside the accuracy. When a position is later questioned, the method is what distinguishes a genuine survey from an optimistic estimate.
  6. Report accuracy compliance by class. The share of captures meeting their class budget is a better data-quality metric than any error count, because it is the number a regulator’s accuracy claim rests on.
From an asset class to a capture instruction a crew can follow The accuracy budget comes from what the asset is used for: an isolation trace needs to find the valve, an excavation needs to avoid the main. The method follows from the budget, and the instruction that reaches the crew names the method, the acceptance threshold and what to do when the site will not support it. The last element is the one usually missing, and it is why crews capture unusable positions rather than recording that the site was unsuitable. USE what the position has to support BUDGET accuracy required per asset class METHOD receiver and correction INSTRUCTION threshold and the fallback the missing instruction SITE WILL NOT SUPPORT IT record that, and return with an offset method — never capture a position you know is bad Without a stated fallback, a crew under canopy will capture something rather than nothing.

Conclusion

Accuracy requirements derived from use rather than from equipment let an estate spend correction subscriptions where they change decisions and skip them where they do not. Recording the reported accuracy and the capture method with every point turns the estate’s accuracy from an assertion into a measurement, and giving crews an acceptance threshold and an explicit fallback is what stops an unsuitable site from producing a confident, wrong position. With budgets published and enforced, the validator has something to check against and the reconciler has grounds to prefer one observation over another.

For authoritative reference, consult the NGS guidelines for real-time GNSS positioning and the OGC standards portal.

Related pages