Field techniques for dense corridors
A second advanced lesson for the team about to scope work in the corridors where locating fundamentals stop being sufficient — hospital podium decks, utility tunnels, dense urban rights-of-way. The lesson focuses on the geophysics you run when the surface does not give the answer: choosing between subsurface and surface methods, tuning ground-penetrating radar, electromagnetic induction, and acoustic emission, and fusing the four into a defensible report. The lessons here assume you have Module 01 — multi-utility sites already in hand.
Track each section as you finish it
Five sections — intro, conceptual framework, procedures, common errors, knowledge check. Sign in to mark sections complete; progress shows up on your dashboard and across sessions.
New to the advanced track? Start with Module 01 — multi-utility sites for the framework this lesson extends.
Why this lesson exists
A reflex sweep across a single frequency on a single instrument covers a sliver of the corridor — and on a podium deck or in a utility tunnel that sliver is the part of the site you can most afford to miss. This lesson is about the four-instrument method mix you run when the geometry stops being obvious and the answer does not survive a single-method read.
- Senior field technicians moving from multi-utility work to dense-corridor geophysics — the kind of site where the surface itself is the noise floor.
- Project engineers and owner-side oversight staff evaluating geophysics submittals, with an interest in what makes a method-fusion report defensible.
- Locators preparing for engagement on hospital campuses, university research buildings, or industrial facilities with post-tension slab construction.
By the end of the lesson you should be able to do the following without reaching for a checklist.
- 01Discriminate between subsurface and surface locating canvases, and pick the canvas that matches the question the site is asking.
- 02Choose ground-penetrating radar, electromagnetic induction, or acoustic emission by site material, depth, and access constraint — not by reflex.
- 03Tune antenna frequency and method pairing (e.g. high-frequency GPR with multi-frequency EM) so the rebar mat does not become the target map.
- 04Interpret a hyperbolic reflection signature to read depth and host-medium velocity without inventing either value.
- 05Design a dual-frequency EM pass that exposes a stainless-steel tracer wire the standard 9 kHz sweep attenuates away.
- 06Spot acoustic false positives — regulator venting, pump cycling, thermal expansion — and corroborate any hit with a non-acoustic method before writing it into the report.
- 07Fuse GPR, EMI, and acoustic time-series with aligned timestamps so the fusion sweep reads as one defensible dataset, not three splatter plots.
- 08Hand off a method-fusion report the project team can act on, including what was NOT found and the method that bounded each negative.
Subsurface vs. surface locating canvases
Every instrument on the truck operates in one of two canvases: it either sees beneath the surface (radar reflections, acoustic transients) or it energises what the surface couples to (induction, tracing, magnetometry). Picking the right canvas for the question the site is asking is the decision that decides every other instrument choice you make on that engagement.
A clean reflex on a hospital podium deck is to pull the GPR cart and run a profile. That same reflex on a copper tracer-wire lateral in a utility tunnel produces nothing — the surface couples to the wire well, the radar does not. The lesson that follows treats the canvas question as the first decision on every engagement, before any instrument is pulled.
The risk on dense-corridor sites is not technique — most field technicians can run each instrument competently — the risk is sequencing and fusion. A single-instrument read across a rebar deck confidently produces a target map that is mostly rebar. A single-frequency EM pass on a stainless-steel tracer-wire lateral confidently produces a non-detection that is wrong. The corrective move is method-fusion with alignment, not heroics on any one canvas.
A practical reading order: pick the canvas the question lives in, choose one primary method from that canvas, choose one corroborating method from the OTHER canvas (do not stack two subsurface reads and call it fusion), and align the timestamps before any mark-out. That ordering — canvas, primary, corroborating-canvas method, alignment — is what makes a method-mix defensible instead of decorative.
A method-fusion sweep in the field
Engagement vocabulary from Module 01 still applies — pre-field, field acquisition, interpretation & closeout — but in dense-corridor work each phase has a method-mix dimension that right-of-way work does not.
- Inventory what the surface couples to (metallic laterals, tracer wires, duct banks) and what it does not (PVC, HDPE storm, post-tension slab voids) — that decides the EMF pass list before it decides anything else.
- Pick antenna frequency from host material: high-frequency (1.5–2.0 GHz) GPR for shallow non-metallic runs in concrete; mid-frequency (400–900 MHz) GPR for soil profiles and deeper utility corridors.
- Plan the dual-frequency EM pass — 9 kHz for copper and steel geometry, a higher-frequency pass (≈30 – 80 kHz) for stainless tracer wire and thin conductors the low-frequency pass attenuates away.
- Decide which acoustic segments warrant corroboration-instrument pairing and which do not — a quiet segment does not justify acoustic instrumentation unless the project risk profile already warrants it.
- Run GPR on a transect grid calibrated against a known-depth target before any decision is made on profile inversion — the velocity assumption is where radargrams quietly move a target up or down by tens of centimetres.
- Pair a 9 kHz and a higher-frequency EM pass on every suspected segment; record both responses so the canonical EM signature for the segment is dual-frequency, not single-frequency.
- Mark acoustic instrumentation with a stationary reference probe on a known-good segment so threshold crossings can be sanity-checked against the reference signal — false positives on a quiet site without a reference are not deliverables.
- Sync timestamps across methods at sweep start and every calibration milestone so the fusion read at closeout is one walk, three streams, not three reads glued together in post.
- Mark every segment with the method stack that confirmed it — e.g. "GPR + 33 kHz EM + acoustic reference probe" — so the consumer does not have to infer whether the depth estimate came from radar, a hyperbola apex, or an acoustic-transient threshold.
- Document host-medium velocity on every GPR-derived depth estimate, including when the velocity was picked from a known-depth target versus the radargram fingerprint.
- Pair every acoustic hit with the corroborating-method result that promoted it from "spike" to "target" — a single-method acoustic report invites a forensic re-read on every subsequent engagement.
- List the negative results and the method that bounded each (e.g. "no metallic conductor detected on 9 kHz + 33 kHz EMI sweep, bounded by trace length 240 m and tie-lines every 4 m") — the absence of evidence is evidence only when bounded.
The two-frequency EM pass is the difference between a defensible surface canvas and a confident miss. A 9 kHz pass gives you the geometry of every copper and steel conductor at depth; a higher-frequency pass tightens the footprint and is what exposes the stainless tracer wire the low-frequency pass attenuates away. Picking only one of those is not a method mix — it is one of the two readings the other would have caught.
Phase 1 — method selection — is the SCOPE precondition. Phase 2 — fusion sweep — produces the field markings, findings map, and field report that DELIVERABLES names. Phase 3 — interpretation & report-out — is the recommendations memo and closeout briefing the framework promises, with each segment annotated by method stack rather than by instrument used in isolation.
Open the engagement frameworkSix geophysics mistakes that show up on dense-corridor work
The lesson collects the calls we have flagged in dense-corridor debriefs across hospital podiums, utility tunnels, and congested urban rights-of-way. Each mistake has a clear field signal — the kind a careful technician can spot before closeout — and a corrective move the team can run without a second mobilisation.
Radar returns go black — no reflectors, no apex, just attenuated signal where a clean profile was expected.
Switch to mid-frequency (200–400 MHz) for wet-clay soil profiles and document the host-medium limitation on every interpretation; do not invert an attenuated radargram into a depth estimate.
Strong continuous signal across the entire deck; geometry aligns with what the structural drawings show for the slab, not with the conduit run set.
Reconcile the EM response with the structural drawings before any mark-out; isolate the target stream by running high-frequency GPR to separate the structure from the conduit runs underneath.
Weak or absent EM response on what the GIS labels as a traced gas service; project team insists the wire is on site.
Run a higher-frequency induction pass (≈30 – 80 kHz) or direct-couple a sonde onto a conductor access point; never report a stainless tracer wire as absent on a single low-frequency pass.
Instrument flags transient spikes on a quiet site with no active excavation; no way to sanity-check whether the spikes are real events or instrument-noise patterning.
Keep a reference probe planted on a known-good segment for the duration of the sweep; any hit written into the report must beat the reference threshold, not just the instrument default.
Two clean datasets, but the apparent target on one method does not align with the apparent target on the other at any single moment of the sweep.
Sync timestamps at sweep start and every calibration checkpoint; re-base before you trust any fusion read — unsynchronised data is three readings, not one.
Clean radargram, no obvious targets; the project team is told the corridor is clear and excavation begins.
Bound every negative with the method that produced it — frequency, sweep length, calibration point, and the reason a reflector at this depth would have appeared; absence of evidence is only evidence of absence when bounded.
An acoustic emission profile reads as a time-series: amplitude on the vertical axis, time on the horizontal axis. A real leak produces a transient spike rising above the noise floor and into a corroboration band; a regulator venting produces sustained sub-threshold oscillations tied to a cycle. Any hit written into the report must beat the reference threshold — not just the instrument default.
Self-check — four questions, reveal as you go
A short authored self-check covering the method-mistakes that show up on day one of a dense-corridor engagement. Reveal an answer to read the rationale; reset between engagements with Hide all.
This is a self-check, not a graded exam. Pick a question, reveal the answer, and use the rationale to decide whether the topic earns another pass before closeout.
0 of 4 answers revealed. State is kept on this device only.
- 01
You arrive at a hospital podium deck with corrugated metal decking and PVC conduits in the slab below. How do you choose between ground-penetrating radar and electromagnetic induction?
- 02
You are reading a GPR radargram and see a clear hyperbolic reflection signature crossing the transect. What does the geometry of that hyperbola tell you about depth and velocity?
- 03
A client asks you to trace a stainless-steel tracer wire on a gas service lateral. Why is a 9 kHz induction sweep the wrong starting point, and what should you reach for instead?
- 04
On a quiet site with no active excavation, your acoustic emission instrument spikes repeatedly on what reads as a live gas service. What is your first diagnostic move before you call it a target?
Coming in the advanced track
Module 02 stops at method-fusion fieldwork. The modules below extend the track into SUE quality levels, records reconciliation at scale, and induced-signal environments — each one authored as a standalone lesson when the curriculum is ready.
Each advanced module ships with its own lesson page, knowledge check, and a stand-alone diagram set so the track stays useful even if you arrive mid-sequence.