GPR can guide where to investigate, but it does not itself determine structural capacity, ground bearing capacity, repair quantities or treatment design.
1. Decide whether verification is needed
The required certainty depends on consequence. Screening a broad area for follow-up may tolerate an interpreted result. Drilling through a slab, reopening a production line, defining a structural repair or estimating treatment quantities may require direct confirmation or complementary evidence.
Ask which unresolved question would change the decision: target identity, exact location, depth, lateral extent, active water pathway, material condition or structural significance.
2. Select a method that tests the hypothesis
- Records and visual correlation: compare drawings, chambers, joints, repairs, levels and drainage alignments.
- Known-depth correlation: improve time-zero or velocity assessment using a reliably surveyed feature.
- Repeat or denser GPR: test repeatability and geometry where original coverage is insufficient.
- Complementary geophysics: use a method sensitive to a different physical property where appropriate.
- Probing, coring or drilling: inspect local material and support conditions under controlled permits.
- Borescope or camera inspection: observe an accessible opening or confirmed cavity locally.
- Excavation or opening-up: expose the feature where consequences and controls justify it.
- Drain testing or inspection: test a suspected water or washout mechanism independently.
A verification point confirms local conditions. It does not automatically validate every interpolated area.
3. Choose locations without bias
Include points within the strongest or most repeatable area of interest, boundary or transition points where extent matters, and reference points outside the interpreted zone. Coordinate them against the same site control used for the radar data.
Plan for utilities, reinforcement, contamination, process hygiene, buried services and structural safety. Verification is subject to the client’s permits and engineering controls.
4. Preserve the evidence chain
Record the verification coordinate, method, depth, recovered or observed condition, date, photographs, samples and any deviation. Then reconcile the result with the specific radar line and recorded coordinate. A mismatch may reflect positioning, calibration, spatial variability or an incorrect interpretation; it should be analysed, not hidden.
5. Structure the report for audit
- Document control, status and revision.
- Client-focused summary and decision context.
- Objective, scope and exclusions.
- Site context and available records.
- Survey geometry, coverage and inaccessible areas.
- Equipment and acquisition settings.
- Data inventory and integrity findings.
- Processing sequence and display controls.
- Results by dataset, line and coordinate.
- Integrated interpretation and competing hypotheses.
- Separate confidence grades.
- Verification results or suggested further assessment.
- Limitations and conclusion.
- Traceable plans, profiles and technical appendices.
6. Make every figure reconstructable
Plans need scale, orientation, line identifiers, origins, directions, surveyed coverage and inaccessible areas. Radar profiles need line name, distance axis, time or qualified depth axis, band, processing description and marked reference coordinates. Captions should call interpreted zones “areas of interest,” not confirmed cavity boundaries.
7. Use bounded language
Prefer “repeatable radar-response change,” “consistent with,” “may be influenced by,” “not uniquely diagnostic” and “requires correlation or physical verification.” If a void has been physically confirmed, state the method and point of confirmation; do not extend that confirmation beyond the evidence.
Suggested next steps should be conditional and within scope. Do not let the GPR report silently become a repair design.
8. Complete issue checks
- Coordinates use one stated origin and direction.
- Dataset and line identifiers match the figures and source files.
- Acquisition and display parameters are not confused.
- Depth wording reflects calibration status.
- Survey gaps and non-intersecting coverage are disclosed.
- No unsupported area, volume, grout quantity or safe load is stated.
- Dates, project names, revision, page numbering and PDF metadata are current.
- The rendered final document has been visually checked page by page.
Official application references
FHWA InfoTechnology — GPR for pavement layer-thickness assessment