Underground utility records and unforeseen ground conditions

A trench opens and there is a duct in it that no drawing showed. Work stops. The utility owner is identified, eventually. A diversion is designed, priced and argued about. The programme moves.

The event is then recorded under a heading that appears in every set of conditions of contract in some form: unforeseeable physical conditions. The heading does useful work, because it allocates the consequence between the parties, which is what a contract is for.

What it does not do is explain anything. The duct was not unforeseeable in any ordinary sense. Somebody installed it, on a date, to a design, and in most cases recorded it. The condition was unforeseen because a record did not travel from the organisation that made it to the organisation that needed it. That is a failure of information transfer, and unlike ground itself, information transfer is something an owner can do something about.

What the standard already admits about records

The engineering profession has been explicit about this for longer than most owners realise. ASCE 38 grades utility information into four quality levels, and the definition of the lowest one is the part worth reading.

The levels rise from there. Quality Level C surveys visible surface features and correlates them with the records. Quality Level B applies surface geophysics to establish the existence and horizontal position of subsurface utilities. Quality Level A is physical verification by excavation, and gives the tightest tolerances available.

Set that scale beside how a great many projects are actually procured. The tender pack contains utility records obtained from the asset owners. Those records are, by the standard's own classification, Quality Level D, which the standard says is for route selection. The contractor prices detailed construction against them. Nobody has misrepresented anything, and nobody has broken a rule. The information has simply been used two levels above the confidence it carries.

Quality level	How the information is obtained, and what it carries
QL-A	Physical verification by excavation. The tightest tolerances available.
QL-B	Surface geophysics, establishing the existence and horizontal position of buried utilities.
QL-C	A survey of visible surface features, correlated with the records.
QL-D	Existing utility records or verbal recollections alone. Characterised by the standard as typically unreliable, and suitable primarily for planning and route selection.

A tender pack whose utility information is the records obtained from asset owners sits, by the standard’s own classification, on the bottom rung. Pricing detailed construction against it uses the information two levels above the confidence it carries, which is this publication’s reading of the scale rather than the standard’s claim.

FIG. 01The four utility information quality levels of ASCE 38, best first. The levels and their characterisations are the standard's own; the observation that a tender pack of owners' records sits on the bottom rung is this publication's reading, argued above.

Why the records are bad, structurally

The interesting question is not whether utility records are incomplete. They are, everywhere, and complaining about it has never improved one. The question is why they stay incomplete when so many parties are harmed by it, and the answer is a clean piece of economics that the UK government wrote down when it built the business case for fixing it.

That is the whole mechanism in one sentence. The organisation holding the record does not pay for the record being wrong. A telecommunications operator whose duct position is recorded to the nearest few metres suffers very little when a road contractor hits it. The contractor stops work, the road authority carries the programme consequence, the public carries the closure, and the operator receives a repair funded by somebody else.

An externality of that shape does not resolve through good intentions or through better contract drafting, because neither changes who bears the cost. It resolves when somebody with a wider view than any single asset owner intervenes, which is what the same document describes.

What it costs, measured

The UK numbers are worth quoting because they were produced to justify public spending and are therefore both conservative and checkable.

The Commission estimated the average direct cost of a utility strike at £3,371. Then it did the thing most cost discussions skip.

Twenty-nine times. The visible cost of a strike, the bit that generates an invoice and gets recorded, is around 3% of what the event actually costs. Everything else lands as programme time and disruption, distributed across parties who did not cause it and recorded, if at all, as something other than a utility strike.

The cost of one strike, drawn to proportion

FIG. 02One strike, drawn to proportion. Average direct cost £3,371; indirect costs, which the Commission identifies as including project delays and extended road closures, 29 times larger. NUAR Economic Case Summary.

Against roughly 60,000 strikes a year on buried services, the Commission put the economic cost at about £2.4 billion annually. It attributed around 30% of incidents to inadequate plans and on-site procedures for using data, taking that as its central estimate of the share avoidable through better information.

The response was the National Underground Asset Register, a single digital map of buried pipes and cables assembled from the records of hundreds of separate public and private asset owners. The economic case put the total monetised benefit at £3.4 billion, or £347 million a year over ten years, of which £240 million a year was reduced strikes. A conservative estimate of the return was £30 of benefit for every pound invested.

The part that generalises

A reader outside the United Kingdom should not take the register as the transferable lesson. Registers are expensive, slow and specific to a jurisdiction's utility ownership.

What generalises is the diagnosis, and the diagnosis is uncomfortable for owners because it puts the problem on their side of the fence. The Commission's description of the pre-register situation is that data had to be sought from multiple organisations that delivered it in multiple formats and scales, on multiple base maps, with varying accuracy, possibly incomplete, and collected at different frequencies, making it challenging to reconcile all relevant data into a single plan.

Every clause of that describes a coordination failure rather than a data-collection failure. The records mostly existed. What did not exist was anybody whose job was to hold the composite.

That is a question an owner can ask about its own right of way today, without waiting for anybody to build anything. Who holds the combined picture of what is under this corridor? On most corridors, the honest answer is that nobody does, and that the nearest thing to it is an experienced engineer who has worked the area before and remembers.

The experienced engineer as an unmanaged control

That engineer is a real control and often a good one. Somebody who has opened the ground in a district before will look at a drawing and say that there is usually a water main closer to the kerb than that, and they will frequently be right.

Treating this as a happy accident rather than as a system is where owners get caught. It is a control with no specification, no coverage, no succession and no record. It works until the person is on another job, retires, or is simply not in the meeting where the alignment is fixed. Its failures are invisible, because a corridor that was never reviewed by somebody with local memory produces no signal at the time, only a trench six months later.

An owner relying on it should at least know it is relying on it. The usual reason it goes unnamed is that naming it would raise the question of what happens when it is absent, and that question has an expensive answer.

What this changes about the programme

The connection to the schedule is direct, and it is why this sits in delivery rather than in asset information.

A strike or an unrecorded obstruction does not produce a delay of the length of the diversion works. It produces the first shift in a sequence, and shifts compound, for reasons set out in a separate piece. The 29-times multiplier and the compounding argument are the same observation approached from two directions: the cost of a delay is mostly not where the delay is recorded.

Two practical consequences follow.

Buy the survey before the design, not after the strike. The choice is not between spending on subsurface investigation and not spending. It is between spending it before award, where it is a priced scope item, and spending it after, where it is a variation with a programme consequence attached and an argument about entitlement.

Ask what quality level the design was based on. This is a single question with a documented answer, and it is the fastest way to find out whether an owner is carrying a risk it thinks it has transferred. A design priced against records alone is a design priced against QL-D, whatever the contract says about who bears unforeseeable conditions.

Sources. ASCE 38, Standard Guideline for Investigating and Documenting Existing Utilities, for the four utility quality levels and for the characterisation of Quality Level D as information derived solely from existing utility records or verbal recollections, both typically unreliable, suitable primarily for planning and route selection. UK Geospatial Commission and Cabinet Office, National Underground Asset Register (NUAR): Economic Case Summary, published 4 November 2021 under the Open Government Licence, for the average direct cost per strike of £3,371, for indirect costs averaging 29 times direct costs, for the widely reported figure of 60,000 strikes a year and the resulting estimate of £2.4 billion in annual economic cost, for the attribution of around 30% of incidents to inadequate plans and on-site procedures for using data, for the monetised benefit of £3.4 billion or £347 million a year over ten years, for the £30 of benefit per pound invested, and for the stated absence of direct incentives for individual asset owners to improve data where the risks and costs fall on other organisations. The strike frequency figure is described by the Commission itself as a widely reported industry statistic rather than as a measurement of its own.

The ground was always there. The record was not.