Integrating Multiple Geophysical Methods

A decision framework for geophysical method selection works through three questions in sequence: what physical contrast is the target likely to produce in this soil, what will the noise environment look like, and what resources (time, equipment, personnel) are available? The answer to the first question comes from the soil and target characterisation; the answer to the second comes from the site history and local knowledge; the third is an operational constraint.

The framework is not a rigid algorithm. Site conditions are often mixed: a garden may have sandy topsoil over clay subsoil, or a concrete path adjacent to open lawn. In these cases the framework is applied zone by zone, with different method combinations for different parts of the search area. The Cheetham (2005) framework recommends documenting the zone-by-zone rationale explicitly, because this is precisely what a cross-examining barrister will ask about.

The value of a multi-method survey depends entirely on the ability to compare anomalies from different instruments at matching ground positions. This is only possible if all datasets share a common spatial reference. The standard workflow assigns a site-specific grid, tied to a surveyed datum point, before any instrument is deployed. All traverses are walked along marked grid lines, and the instrument's along-traverse position is recorded either by the instrument's internal odometer or by a connected GPS receiver.

In GIS software (QGIS, ArcGIS, or equivalent), the exported datasets from each method are loaded as separate layers. Each layer is georeferenced to the site datum coordinates. The layers are then visualised together, either as separate panels or as a composite image, and an analyst marks anomalies on a master annotation layer. An anomaly that appears in two or more layers at the same position is flagged as multi-method confirmed. An anomaly that appears in only one layer is flagged as single-method candidate.

Every geophysical method produces false positives, and the specific sources depend on the geological and land-use context of the search area. Knowing the local sources before the survey protects against misinterpretation that would misdirect excavation resources or, worse, lead to a false conclusion about the absence of a target.

Tree roots are among the most common GPR false positives in garden and woodland searches. A large root system produces multiple hyperbolic reflections at shallow depth that can mimic the gravity of a burial zone. The diagnostic is geometric: tree roots produce a radial or branching pattern in plan view, centred on the trunk position, rather than the sub-rectangular bounded zone expected from a grave cut. Overlaying the GPR anomaly map with a survey of surface vegetation usually resolves the ambiguity quickly.

Confidence ranking is a structured way of communicating uncertainty without abandoning precision. The three-tier system used in UK forensic geophysical practice, derived from Cheetham (2005) and widely applied by the Forensic Search Advisory Group (FSAG), is a practical standard that courts have accepted:

1. Confirmed
   The anomaly has been physically investigated and the target identified. The geophysical result is verified by excavation or probing. This tier is assigned retrospectively after ground-truthing.
2. Inferred
   Two or more independent methods detect an anomaly at the same position and consistent depth. The anomaly's geometry and dimensions are consistent with the expected target. This is the highest pre-excavation confidence tier, and it justifies directing excavation resources to the site.
3. Possible
   A single method detects an anomaly of appropriate position and shape, but no second method confirms it. The anomaly cannot be explained by known false-positive sources. This tier warrants further investigation (a second method, a control probe) before committing to excavation.

Below these three tiers sits a fourth implicit category: background variation, anomalies that do not rise above noise or that are fully explained by known infrastructure or geological sources. These are recorded on the site plan (so that the survey coverage is documented) but are not reported as candidates.

Published comparative studies from controlled burial experiments provide the empirical basis for the confidence model. The most systematic are those conducted by Pringle and colleagues (2008) at a simulated urban site in the UK using animal proxies and physiological saline at known positions, and the parallel North American work by Schultz (2008). These studies cross-validated geophysical methods against known ground truth and reported detection rates for each method and soil type combination.

Key findings from Pringle et al.: GPR achieved the highest detection rates in sandy and chalk sites (85-95% at depths to 1.5 m) but fell to less than 40% in clay-dominated sites. Magnetometry achieved consistent detection rates of 70-85% across soil types for recent burials but declined for burials older than approximately ten years at sites with undisturbed overburden. ERT provided confirmation of anomalies identified by primary methods but rarely detected targets that both GPR and magnetometry missed. The methods are complementary rather than substitutable.

Field case outcomes from UK missing-person investigations (published in redacted form through the FSAG and in academic case studies) consistently show that multi-method surveys directed excavation to the correct location in cases where single-method surveys had been inconclusive. The operational history supports the framework: start with the fastest method appropriate to the soil, add a second method to confirm candidates, and use excavation only when the multi-method evidence justifies it.

The final survey report must serve two audiences with different needs. Excavators need precise coordinates, depth estimates, anomaly dimensions, and a recommended excavation sequence. Legal teams and courts need a clear statement of what was done, what was found, what it means, and what it does not mean. These are not contradictory requirements, but they require deliberate structuring of the document.

• Site context section: soil type, geological setting, prior land use, and cultural features mapped before scanning. This section explains why the method selection was made and what false-positive sources were assessed.
• Methods section: each instrument used, its settings, traverse grid specification, processing steps, and calibration parameters. Sufficient detail for independent replication.
• Anomaly catalogue: a table listing every anomaly, its coordinates, depth, dimensions, detected methods, and confidence rank. Anomalies that were explained by known false-positive sources are listed separately with their explanation.
• Figures: site plan with grid, anomaly positions marked, GPR cross-sections annotated with depth scale and velocity, ERT sections, magnetometry greyscale map. All figures carry north arrows, scale bars, and datum references.
• Conclusions: factual findings (anomalies detected at these positions) clearly separated from expert opinion (these anomalies are consistent with, or inconsistent with, a burial). The expert does not determine what the anomaly is; they state what the data supports and at what confidence level.

After excavation, the report is updated with ground-truth outcomes. Correct detections and missed targets are documented alongside their geophysical signatures. This feedback is the professional development component of forensic geophysics: each confirmed prediction and each miss informs the practitioner's calibration of confidence levels for future surveys in similar terrain.