Subsurface & utility

You cannot verify what you did not dig up. Every subsurface method returns an interpretation, not a measurement, and the interpretation carries uncertainty that the deliverable has to be honest about.

Ground conditions dominate. Radar penetration depends on soil conductivity, moisture and material. Clay is difficult. Wet clay is worse. The same equipment performs very differently across one site.

And records are usually wrong. As-laid drawings, where they exist at all, were often drawn from memory after the fact. The record you inherit is a hypothesis, not a fact.

Establish what you are looking for and how deep. Antenna frequency is a trade between depth and resolution, and you cannot have both. A 500 MHz unit resolves shallow detail; a low frequency unit reaches deeper and sees less.

Grid the survey area properly. Radar returns a cross-section along a line, so coverage is a function of line spacing. Wide spacing means you can walk a linear service straight past it.

Fly the radar and the magnetometer as separate passes if both are in scope. They detect different things: radar returns interfaces, magnetometry returns ferrous anomalies.

Capture the surface with LiDAR or photogrammetry at the same time, so the subsurface interpretation sits on real topography.

Walk open trenches before backfill. That is the only moment the truth is visible.

Radar tells you what is under the ground. A scanner records the surface and every accessible void. Together that is a complete utility model.

Depth from radar is calculated from signal velocity, and velocity varies with ground conditions. Calibrate against a known target on site where you can, and state the assumption where you cannot.

Position in plan is generally more reliable than depth. Say so in the deliverable.

Every subsurface finding is an interpretation. The professional standard is to state confidence, state method, and state coverage, rather than present anomalies as located services. A drawing that shows a line without a confidence statement invites someone to dig on it.

Where the consequence of being wrong is high, verify with potholing.

A plan of interpreted features with confidence levels, a statement of method and coverage, and the raw radar sections so someone can check the interpretation.

Recipients are civil contractors, utility locators, geotechs and archaeologists, all of whom will act on it with machinery. They need to know what is confident and what is indicative.

What gets it questioned: anomalies presented as identified services, depths quoted without a velocity assumption, and coverage claims that do not match the line spacing actually walked or flown.

Line spacing too wide. A service running parallel to your survey lines can sit between them and never appear. Grid tighter than feels necessary, and grid in two directions where it matters.

The other one is over-claiming. An anomaly is an anomaly. Calling it a 150 mm water main because that is what the drawing says is how strikes happen.

And the missed opportunity: the trench that gets backfilled without anyone recording it. Ten minutes with a scanner would have produced a permanent record, and instead the next crew inherits the same uncertainty you did.