The Gold Standard of Archaeology | Self-Potential, Redox Gradients and Groundwater Flow

1. Begin with a voltage difference, not a picture of the past

Back to contents

A self-potential reading is the difference between a moving electrode and a reference electrode. The instrument does not directly measure water velocity, redox potential or an archaeological feature. It measures an electrical consequence produced by one or more subsurface processes. Sign and magnitude depend on the reference, wiring convention, electrode contact and survey geometry.

Start every interpretation with the observable: where, when and relative to which reference was the voltage recorded? Preserve raw millivolt values before filtering or gridding. A colourful anomaly map is a derived view, not a photograph of what lies below.

2. Separate the main source mechanisms

Back to contents

Electrokinetic or streaming potentials can arise when groundwater moves through charged mineral surfaces. Redox gradients can support electrochemical source currents, sometimes described through a geobattery model. Sulphide bodies, corrosion, ionic concentration gradients and temperature contrasts may contribute too. More than one mechanism can operate at the same place.

Archaeological interpretation must therefore proceed through process hypotheses. A negative anomaly beside a ditch might reflect focused seepage, reduced organic fill, corroding metal or an electrode problem. The voltage pattern becomes evidence only after independent observations constrain these alternatives.

3. Design a stable reference network

Back to contents

One electrode is usually kept at a reference station while the other visits grid points, or several electrodes are monitored as an array. The reference location should be geologically quiet, accessible and protected from changing drainage, vehicles and buried services. Closing loops and repeat occupations help reveal drift.

Record electrode serial numbers, contact medium, contact resistance, cable route, polarity, time and weather. Swap electrodes or make reciprocity checks according to the protocol. If the reference drifts by 18 millivolts during a survey whose target anomaly is 20 millivolts, apparent spatial structure may actually be temporal change.

4. Treat electrode behaviour as part of the measurement

Back to contents

Non-polarising electrodes reduce electrode polarisation but are not perfect. Differences in electrolyte concentration, porous-pot condition, soil contact and temperature can create offsets. Dry crust, stones and disturbed topsoil make contact unstable. Operators may also introduce noise by moving cables or handling electrodes before readings settle.

Use a settling rule, log repeat variability and avoid smoothing away poor contact. A field sheet should distinguish a true reoccupation from a replacement reading after contact repair. Quality flags belong with each datum so that later maps do not give every value equal credibility.

5. Connect topography, hydrology and redox carefully

Back to contents

Groundwater flow often follows hydraulic gradients, but self-potential polarity is not a universal downhill arrow. Soil texture, saturation, surface charge and boundary conditions affect the coupling. Redox-related anomalies depend on electron donors, acceptors, conductive pathways and the geometry of oxidising and reducing zones. Rainfall can move water tables and shift both processes.

Survey levelling points, drainage, standing water, soil moisture and accessible piezometers. Where preservation questions involve anoxia, pair voltage with dissolved oxygen, redox-sensitive chemistry or direct redox measurements. Correlation can guide a mechanism; it does not make the electrical signal a substitute for geochemistry.

6. Worked example: a wetland settlement grid

Back to contents

In an invented 30 by 20 metre grid, readings relative to one reference range from minus 42 to plus 18 millivolts. A linear negative band crosses a timber platform and seems to mark an ancient channel. Repeat measurements after overnight rain shift the whole eastern half by about 15 millivolts. Piezometers show the water table rose there, while the western reference area remained stable.

After drift correction and a dry-weather repeat, only part of the band persists. Auger samples in that segment contain organic channel fill and reduced iron; the disappearing portion follows a modern drain. The responsible conclusion is two processes: an archaeological palaeochannel signature interacting with current hydrology. One survey alone would have overdrawn the feature.

7. Build interpretation from converging methods

Back to contents

Electrical resistance or resistivity tomography can test conductivity structure; electromagnetic induction can map conductivity more rapidly; ground-penetrating radar can examine interfaces; magnetometry may detect fired or infilled features. Cores and soil chemistry can establish stratigraphy, moisture and redox conditions.

Agreement should be process-based, not merely visual. A self-potential line and low-resistance line may share a cause—wet organic fill—or may overlap by chance. Co-register uncertainty, compare depths where methods permit and test predictions at selected points. A small number of well-placed cores can be more valuable than another attractive map.

8. Changed-context transfer check: from wetland to tell site

Back to contents

Move the method from a saturated wetland to a dry, layered tell. Streaming potentials may weaken, electrode contact may deteriorate and buried metal or modern services may dominate. The same 20-millivolt threshold cannot be carried across. Survey timing, electrode preparation and mechanism hypotheses must be redesigned.

The transferable principle is not a favourite anomaly shape. It is the chain from voltage, through field controls, to plausible source processes and targeted validation. When hydrogeology changes, both expected signal and confounding structure change with it.

9. Know the limits and common misconceptions

Back to contents

Self-potential inversion is non-unique: many source distributions can reproduce a surface field. Reference choice changes the zero level. Cultural metal, fences, cathodic protection, power infrastructure, roots, rainfall and topography can complicate maps. A sharp anomaly is not automatically shallow, and a large anomaly is not automatically archaeological.

Do not convert voltage directly into groundwater discharge without a calibrated model and supporting hydraulic data. Do not call every redox-associated pattern a microbial battery. When mechanisms cannot be separated, report a zone for investigation rather than a named feature.

10. Audit the survey from electrode to interpretation

Back to contents

Archive raw readings, coordinates, elevations, timestamps, reference occupations, loop closures, electrode checks, contact notes, weather and processing code. Keep uncorrected and corrected datasets with the exact drift model. Record interpolation settings and map the distance to observations so blank areas do not look measured.

For interpretation, state each source hypothesis and the observation that would weaken it. Show repeat-survey differences and independent hydrological or geochemical evidence. If excavation follows, compare predictions with exposed stratigraphy and retain negative results; validation is part of the method, not an illustration added afterward.

11. Frequently asked questions

Back to contents

Does the method inject electricity? No; it passively measures natural voltage differences. Can it find a wall? It may detect processes affected by a wall, but the voltage does not identify masonry by itself. Is a fixed reference always fixed electrically? No; its environment and electrode can drift. Does a negative value mean reducing conditions? Not universally; polarity depends on source geometry and convention.

Can rain ruin a survey? Rain can change the signal, which may be nuisance or useful hydrological evidence. Must every anomaly be excavated? No; repeat measurements and complementary methods can prioritise tests.

12. The gold standard explains the source before naming the feature

Back to contents

Self-potential surveying is powerful where archaeology, groundwater and biogeochemistry meet. Its strength is sensitivity to active subsurface processes; that is also its interpretive difficulty. A defensible study stabilises the reference, monitors electrodes and weather, separates plausible mechanisms and tests them with independent evidence.

The final map should not pretend that voltage is an archaeological label. It should show where an electrical process is supported, how stable the pattern is and what observation would decide between explanations. That makes the method useful for discovery, preservation and honest uncertainty.