Soil electrical resistivity tomography, commonly called ERT, is a geophysical technique that uses electrode measurements to infer patterns of electrical resistivity below the ground surface. If you are trying to read time-lapse ERT soil moisture maps, apparent resistivity sections, electrode arrays, inverse models or resistivity-to-water-content calibrations, the first distinction is between the electrical data measured at electrodes and the coloured subsurface picture reconstructed from them.
Electrical resistivity depends on soil water, dissolved ions, temperature, texture, pore structure and other properties. During wetting, many soils become more electrically conductive and hence less resistive, but a blue patch on an ERT image is not automatically a uniquely measured water pocket. It is a model estimate constrained by voltage and current observations, electrode geometry, data error and an inversion procedure. Translating resistivity into water content adds further assumptions and usually benefits from independent soil moisture observations.
This article explains how to interpret soil ERT from first principles: current injection, voltage measurement, apparent resistivity, survey geometry, nonlinear inversion, regularisation, resolution, time-lapse differences, Archie’s relationship and site-specific moisture calibration. Its worked examples use hypothetical numbers, explicit units and stated simplifying assumptions. The aim is to teach readers what an ERT image shows, what remains uncertain and how it complements direct field probes or pore imaging.
Quick answer: what does an ERT colour map represent?
An ERT survey collects electrical measurements using electrodes placed in or near the soil. The instrument applies current through one electrode pair or configuration and measures potential differences at others. With known electrode positions, the observations can be expressed as apparent resistivities. Numerical inversion seeks a spatial distribution of subsurface electrical resistivity whose predicted measurements are consistent with the observations.
The displayed section is a constrained interpretation, not a photograph of underground moisture. Different subsurface models can produce similar electrical data, especially where electrode coverage or sensitivity is limited. Time-lapse ERT compares repeated surveys to examine changes, often related to soil wetting or drying when other influential factors are adequately controlled.
Resistivity and conductivity describe the same electrical property inversely
Electrical resistivity, commonly denoted rho, describes opposition to electric current and is often expressed in ohm metres. Conductivity, commonly sigma, is its reciprocal, with units siemens per metre. A resistivity of 200 ohm metres corresponds to conductivity 0.005 siemens per metre, because one divided by 200 equals 0.005.
This inverse relationship produces an easy verbal mistake. A fall in resistivity is a rise in conductivity, not a fall in every measure of electrical activity. When comparing papers, identify which quantity is plotted and whether the colour scale increases toward high or low values. The same physical change may appear with opposite directions in resistivity and conductivity graphics.
Why soil water changes electrical behaviour
Water containing dissolved ions can conduct electric current through pore spaces, while many dry mineral grains are comparatively resistive. Adding water can create more continuous conducting paths and change the fraction of pore space contributing to electrical conduction. Thus a moist soil frequently has lower bulk resistivity than the same soil under drier conditions.
However, dissolved ion concentration also matters. Fresh rainwater, saline irrigation water and pore solution altered by fertiliser can change conductivity even at similar water content. Clay surfaces may contribute to conduction beyond the simplest pore-water model. Any interpretation of water content should consider chemistry and soil texture, not only how wet the soil looks.
Electrical electrodes do not observe water molecules directly
ERT electrodes record electrical quantities such as potential difference and injected current. They are not miniature cameras transmitting a direct image of the water distribution. An inversion algorithm interprets many electrode configurations together to infer a resistivity field that can explain those measurements within the assumed physical model and error thresholds.
This distinction shapes the language of a report. Say that the ERT model shows a lower-resistivity zone after a rainfall event, consistent with increased water content under the supported calibration. Do not automatically say the electrodes counted water volume at every coloured pixel. The quantitative confidence of each part of the image depends on sensitivity and inversion uncertainty.
What a four-electrode arrangement accomplishes
Many resistivity measurements use electrodes for injecting current and electrodes for measuring a voltage difference. Separating current and potential measurements can help estimate subsurface electrical response while reducing some direct contact effects in the voltage measurement, although electrode and environmental problems can still affect data quality.
The separation and arrangement of electrodes control the regions of subsurface space to which a measurement is sensitive. Dipole–dipole, Wenner and other configurations have different geometric and sensitivity characteristics. A paper should identify its array, electrode spacing, installed depth and survey dimension before readers compare maps produced by different designs.
Apparent resistivity is an observation-derived quantity
Apparent resistivity is commonly calculated from measured voltage-to-current ratio multiplied by a geometric factor for the electrode configuration under the relevant assumptions. In a simplified notation, rho_a equals K times delta V divided by I. Here K has the appropriate length basis to convert measured resistance in ohms into apparent resistivity in ohm metres.
In heterogeneous soil, apparent resistivity is not necessarily the actual resistivity at one underground point. It is a way of reporting what a given electrode configuration observed, interpreted with the geometry of an ideal reference problem. An inversion combines many apparent resistivity observations to infer a more detailed subsurface distribution.
Worked example: electrode-derived apparent resistivity
Suppose a hypothetical electrode configuration has a valid geometric factor of 10 metres. The instrument measures a potential difference of 0.20 volt while injecting 0.010 ampere. The resistance ratio delta V divided by I is 20 ohms. Multiplying by 10 metres gives an apparent resistivity of 200 ohm metres.
The resulting number is associated with that electrode configuration and measurement. It should not be labelled the resistivity of a small soil voxel at a particular depth. Its spatial meaning arises from the current pathways and sensitivity pattern. The inversion uses many such observations to estimate spatial structure under a modelling framework.
The geometric factor matters as much as the voltage
Two electrode configurations can measure different voltages across the same ground because their spacings and layouts differ. Their geometric factors account for some expected differences under the reference physical setting. If an analyst uses the wrong K, the apparent resistivity can be wrong even though the voltmeter and current source were accurate.
A good report includes electrode positions, the configuration and how geometry was handled for non-ideal topography or buried electrodes. Measurements on a slope, irregular survey line or three-dimensional array may require numerical handling beyond a simple textbook geometric formula. Do not copy a factor from an unrelated layout merely because both surveys are called ERT.
Current pathways sample a broad volume
Electrical current does not follow a single straight ray between electrodes. It spreads through the conductive ground according to the distribution of conductivity, geometry and boundaries. A single measurement is therefore sensitive to a volume of soil, sometimes with complex patterns. It can be influenced by material outside the visual centre of the associated pseudosection point.
The common pseudosection display places measurements at approximate positions for visualisation, but those plotting locations are not direct measurements of cell resistivity. Inversion is necessary to seek a spatial distribution consistent with the entire survey. Confusing pseudosection points with actual subsurface voxels is a common error when reading electrical images.
What inversion does with a set of observations
An inversion starts with a model for subsurface resistivity and predicts what the electrodes would measure under that model. It compares the predictions with actual observations, then adjusts the model until the discrepancy and any regularisation penalties meet an acceptable criterion. The process is iterative because electrical response depends on the full resistivity distribution.
The final model is influenced by starting conditions, allowed parameter ranges, error weights and spatial smoothness assumptions. A low residual can mean the model explains the data well under the settings, not that every displayed boundary is uniquely known. The reader should ask both how well the data fit and how the inverse problem was constrained.
Why the inverse problem is not unique
Different subsurface distributions can generate similar surface electrical measurements. A small highly conductive zone and a larger moderately conductive region may be difficult to distinguish in some configurations. This nonuniqueness is a fundamental property of geophysical inversion rather than a software flaw that disappears with a colourful display.
To control nonuniqueness, inversion includes regularisation and prior assumptions. These stabilise the solution but also shape it. The result is a defensible model given data and assumptions, not a guaranteed one-to-one reconstruction of underground reality. Independent moisture probes, cores or geological observations help narrow the range of plausible interpretations.
Smoothness regularisation can hide sharp boundaries
Many inversion procedures penalise abrupt spatial changes unless the data strongly support them. This can make estimated boundaries broad and smooth even if the true soil contains a sharp horizon or narrow crack. Other methods encourage blockier structures or use structural constraints, leading to different-looking results from the same measurements.
A reader should inspect the regularisation choice and whether alternative settings materially change the key feature. A broad transition band may partly reflect algorithmic smoothing. It should not be measured with a ruler and described as a precisely observed 30-centimetre physical mixing layer without independent confirmation.
Depth of investigation is a sensitivity question
ERT resolution generally decreases in regions where the electrode layout provides weak sensitivity, often at greater depth or away from the array. A published image may be coloured all the way to the lower edge of the plot even where the data poorly constrain the model. The plotting software’s ability to fill a rectangle is not evidence that the deepest cells were measured with equal precision.
Check depth-of-investigation or sensitivity information, electrode spacing and survey dimensions. A feature located below the reliable investigation zone should be described with greater uncertainty or excluded from a strong mechanistic claim. Colour saturation does not compensate for low electrical sensitivity.
Two-dimensional inversion may overlook three-dimensional structure
A two-dimensional ERT inversion often assumes that variation perpendicular to the survey line can be treated in a particular simplified way. Soil moisture, roots and infiltration pathways can be strongly three-dimensional. A conductive feature beside a survey line may influence measured voltages and appear as an artefact within the 2D section.
A 3D electrode grid can better represent complex structures under appropriate acquisition and computational conditions, though it also introduces modelling and data needs. When reading a 2D image, ask whether off-line structures might matter. Do not treat every vertical band in a cross-section as a physical wall extending uniformly out of the page.
Time-lapse ERT changes the question
Repeated ERT surveys at the same location can focus on how resistivity changes during rainfall, drying, irrigation or seasonal transitions. Fixed electrode positions can help reduce geometric differences between surveys. Inversion may use a reference model or explicit temporal regularisation so that changes are estimated consistently rather than as unrelated static images.
The main scientific question becomes whether the observed electrical changes are consistent with a physical process, often water movement. Time-lapse methods can reduce some ambiguity by exploiting a stable baseline, but they do not remove temperature, salinity or instrumentation effects. A temporal change is a measured electrical difference plus an inverse interpretation, not automatically a unique water-content trajectory.
Worked percentage change in resistivity
Suppose a model cell has resistivity 200 ohm metres at the reference time and 120 ohm metres later. The relative change is (120 minus 200) divided by 200, or negative 0.40, conventionally written a 40% decrease. Under a suitable moisture calibration, the decrease could be consistent with wetting.
The percent change describes the inverted resistivity estimate, not a 40% increase in soil water content. Resistivity and water content are nonlinearly related in many soils. Without a calibration and checks on pore-water conductivity or temperature, the size of the water-content change cannot be uniquely calculated from that percentage.
Conductivity percentage changes differently
For the same example, resistivity decreases from 200 to 120 ohm metres. Conductivity rises from 0.005 siemens per metre to approximately 0.00833 siemens per metre. That is an increase of roughly 66.7%, not 40%. Reciprocal quantities have different percentage changes even when they represent the same electrical state.
A chart labelled ‘conductivity increase’ can therefore show a different numerical percentage from one labelled ‘resistivity decrease’. Both may be mathematically correct. Readers should compute changes from a consistent reference and name the property. Confusing the two can make a field paper appear inconsistent when the issue is only a reciprocal transformation.
Temperature corrections need evidence
Electrical conductivity of pore water generally changes with temperature. Seasonal soil warming may lower measured resistivity even without the water-content change implied by a simplistic reading. Some studies correct resistivity to a reference temperature using a validated relationship, while others include temperature as a variable in interpretation.
Check which temperatures were actually measured and whether correction factors suit the soil. A single surface thermometer may not represent deeper zones, and an empirical temperature coefficient may differ with chemistry. Comparing a winter and summer ERT image without considering temperature can overstate a wetting trend.
Dissolved salt can imitate wetting
Soil solution conductivity responds to dissolved ions. Fertiliser additions, irrigation-water chemistry, salts mobilised by infiltration or other changes can modify resistivity. Under some conditions, a more conductive pore solution produces a lower resistivity even without a proportional increase in water content.
An ERT paper that aims to estimate soil moisture should therefore measure or reasonably constrain pore-water chemistry, especially where salinity changes are likely. Time-lapse conditions may justify assuming stable chemistry in some experiments but not all. State that assumption before converting electrical changes into hydrological claims.
Clay surface conduction complicates simple mixing laws
Some fine-textured soils have surface-associated conduction processes that are not captured by a pore-water-only model. Clay type, ion exchange, surface chemistry and water distribution may affect electrical behaviour. A relationship calibrated in a clean granular soil may perform poorly in a clay-rich horizon without modification.
This does not make ERT unusable in clay soils. It makes local validation important. A researcher can calibrate resistivity against independently measured water content for the soil domain of interest and assess residual errors. General-purpose formulas must be applied within their physical assumptions rather than as direct laws of every soil mineral assemblage.
Archie’s law is a model, not a conversion button
A common Archie-type formulation relates bulk resistivity to pore-water resistivity, effective porosity and saturation through empirical exponents. One illustrative form is bulk rho equals a times pore-water rho_w times porosity to the minus m power times saturation to the minus n power. Parameters a, m and n represent model choices or calibration; the form is best suited to conditions consistent with its assumptions.
Do not insert coefficients from one rock or soil into another without support. If the soil has significant surface conduction or pore-water chemistry changes, the simplified relationship may fail. An alpine hillslope time-lapse ERT study describes how researchers calibrated a resistivity–moisture relationship rather than treating the theoretical form as self-validating.
A simple Archie-style moisture illustration
Assume a fictional soil has constant porosity, pore-water resistivity and other parameters, with bulk resistivity inversely proportional to volumetric water content squared. At moisture theta equal to 0.20, let resistivity be 200 ohm metres. If resistivity later becomes 128, the model predicts new theta equals 0.20 times the square root of 200 divided by 128, equal to 0.25.
This illustrates a possible nonlinear calibration, not a universal soil-water law. The water content increased from 0.20 to 0.25 cubic metres per cubic metre, a 0.05 absolute change, while resistivity decreased 36%. If pore-water salinity had also changed, the same electrical observation could correspond to another moisture change.
Water content needs site-specific calibration
Volumetric water content describes water volume per bulk soil volume. ERT does not measure that quantity directly. A conversion requires a relationship between water content and electrical resistivity, supported by soil properties and calibration observations. Moisture probes, extracted samples or other independent measurements can help establish such a relationship for the relevant site and depth interval.
A calibration that works in the upper sandy horizon may not work unchanged in a deeper clay-rich layer. The soil structure and time-lapse ERT study combined ERT with direct moisture probes and soil profile observations, showing how both techniques support a better hydrological reading than the resistivity image alone.
Why volumetric and gravimetric water contents differ
Gravimetric water content is usually the mass of water relative to dry soil mass; volumetric water content is water volume relative to bulk soil volume. Their numerical values differ depending on dry bulk density and water density. An Archie-style model that uses volumetric saturation should not be fed gravimetric percentages without appropriate conversion.
For an illustrative soil with gravimetric water content 0.20 kilograms per kilogram, bulk density 1.30 grams per cubic centimetre and water density approximately 1 gram per cubic centimetre, volumetric water content is about 0.26 cubic metres per cubic metre. That conversion is needed before comparison with a calibration defined in volumetric units. A label reading simply ‘soil moisture 20%’ is not precise enough to choose the right equation.
Saturation must not be confused with water content
Degree of saturation represents the fraction of pore volume occupied by water, while volumetric water content represents the fraction of total bulk soil volume occupied by water. If total porosity is 0.50 and volumetric water content is 0.30, the approximate saturation is 0.30 divided by 0.50, or 0.60. The two values describe different denominators.
Some resistivity formulations use saturation rather than volumetric water content. Substituting 0.30 where the equation expects 0.60 changes the predicted resistivity substantially, particularly when the term is raised to a power. A reliable report names porosity and the quantity used for moisture before comparing model coefficients across studies.
Soil horizons can carry distinct electrical signatures
Topsoil, clay-accumulation zones, carbonate-rich horizons, saprolite and other layers may differ in water-holding properties, ion chemistry and texture. Their baseline resistivities can differ even before rainfall changes the water content. A static ERT image may therefore show a strong horizontal boundary that reflects soil material contrasts rather than a moving wetting front.
The 2021 soil ERT field study used independent soil profile information to help interpret static layers and changing resistivity through time. A good reader compares colour bands with known geology and soil horizons before treating them as water boundaries. Electrical variation and moisture variation are linked but not synonymous.
Time-lapse differences can reveal preferential flow candidates
During rainfall or irrigation, resistivity may decrease along zones where moisture moves more readily, including cracks, biopores or pathways through less permeable material. A repeated ERT survey can help identify such patterns when electrical contrasts and time sampling support them. However, an image alone rarely proves exactly which individual pore or root channel carried the water.
Preferential flow is a mechanistic interpretation that benefits from tracer tests, soil observations and direct probes. The geophysical preferential-flow monitoring study provides a concrete example of repeated electrical observations during water movement. The general lesson is to distinguish an electrically responsive zone from a fully mapped microscopic water pathway.
A wetting front is not necessarily a crisp colour boundary
In physical hydrology, a wetting front describes a change in soil moisture moving through a profile. An ERT inversion may produce a diffuse colour transition because the true water gradient is broad, because the instrument has limited resolution, or because regularisation smooths sharp changes. The apparent front thickness is therefore not automatically a direct hydrological measurement.
A cautious description states the range where inverted resistivity changed and identifies the resolution limitations. A study can compare the timing of changes at several depths without claiming centimetre-perfect wetting-front locations. If a sharp front is central to the conclusion, supporting soil moisture profiles or tracer evidence becomes especially valuable.
Repeated electrode positions reduce one source of error
Time-lapse surveys work best when comparisons use consistent electrode geometry. Moving an electrode or changing its contact conditions can alter the measured electrical response. Permanently installed arrays can help maintain repeatability, although weather, surface condition and corrosion may still affect electrode contact over time.
Check whether the researchers reoccupied precisely the same locations or reconstructed the geometry for each survey. A nominally identical line with several shifted electrodes may not be equivalent. The time-lapse inversion should account for or test such changes before interpreting small electrical differences as real shifts in soil moisture.
Contact resistance can compromise measurements
Current electrodes must couple electrically to the ground, and contact conditions can vary with soil dryness, compaction and surface material. Poor contact may limit injected current or increase noise. Potential measurements can also be affected by environmental and instrument conditions. Quality controls often include current levels, contact checks and repeat or reciprocal observations.
An unusually high apparent resistivity in very dry soil might be physically plausible, but it could also coincide with poor electrode contact. Inspect acquisition diagnostics rather than accepting every extreme value. Data filtering is legitimate when based on documented criteria, not when values are removed merely because they complicate a preferred hydrological story.
Reciprocal measurements are a quality-control tool
Under suitable physical and instrumentation conditions, exchanging the current and potential electrode pairs can provide a reciprocal measurement that should agree within appropriate uncertainty. Differences can reveal problematic contact, unstable conditions or measurement noise. The method has assumptions and should be interpreted in the context of the acquisition system.
A low reciprocal error supports data repeatability but does not by itself prove that the inverted model is unique or that the moisture conversion is correct. It validates an earlier link in the chain: the quality of electrical observations. Geophysical confidence requires separate consideration of inversion and hydrological calibration.
Data misfit must be understood with its noise model
An inversion reports how closely predicted electrode responses match measured data under an error model. A small residual can indicate successful fitting, but overly flexible models may fit noise or artefacts. Conversely, a slightly larger misfit can sometimes accompany a more robust model if its regularisation avoids implausible detail.
Readers should ask how errors were weighted and whether bad data were removed by objective quality criteria. A figure caption that says the model ‘converged’ does not establish that every contour reflects a real boundary. Convergence means an optimisation procedure reached a stopping condition under its objective, not that the entire subsurface has been observed directly.
A smooth image can hide true uncertainty
A polished resistivity colour map often contains no visible sign of the many possible subsurface models compatible with the data. Spatial smoothing can create continuous bands even when electrode sensitivity is low or data sparse. Such continuity is partly a product of the inverse model’s prior assumptions.
Useful studies may show sensitivity maps, alternative inversions, uncertainty envelopes or tests of model stability. A 2025 ensemble-ERT research approach explores multiple models to quantify variation in inferred water content. It should be treated according to its publication status and actual study conditions, not as a universal guarantee that one ensemble method resolves nonuniqueness.
Colour maps require careful legends
A red region may indicate high resistivity in one figure and low resistivity in another. Some maps use logarithmic scales, clipped colour limits or percent differences rather than absolute ohm-metre values. A similar shade across two panels therefore need not represent the same quantity. Always read the legend and its units before translating visual contrast into a scientific statement.
When comparing time points, check whether both panels use the same colour scale. Independently scaled maps can exaggerate or conceal the apparent amount of change. If the figure uses log resistivity, a difference in plotted colour relates multiplicatively to actual resistivity. Avoid using ordinary linear interpretations without checking the scale.
A log-resistivity example
Suppose two ERT cells are modelled at 100 and 1,000 ohm metres. On a base-10 logarithmic scale, these correspond to values 2 and 3, a difference of one log unit. That is a tenfold resistivity difference, not a one-ohm-metre change. On a linear colour scale, the visual distribution can look quite different.
Logs can help display broad value ranges but make verbal interpretation more demanding. A report should specify whether colour limits represent raw resistivity or its logarithm. If a model estimates log conductivity rather than resistivity, converting the values may also reverse the direction of the colour relationship. Units and transformations are part of the measurement.
Electrical sensitivity differs across the image
Electrode configurations do not constrain all underground cells equally. Near-surface regions and central portions of an array may be better resolved than distant or deep areas depending on the acquisition. A deeply coloured anomaly can still have substantial uncertainty if its region contributes weakly to the observed voltages.
Depth of investigation should be treated as a range supported by survey geometry and sensitivity tests, not as the bottom of the displayed rectangle. Scientists should avoid interpreting fine substructures in poorly constrained regions. When reading a paper, look for tests in which the inversion is repeated with different starting or reference assumptions to see which major features persist.
What 3D surveys change
A three-dimensional array can capture variation in two horizontal directions and depth, helping where water moves through irregular paths. It requires enough electrode coverage, measurements and computational effort to constrain that extra dimensionality. A 3D model is not inherently accurate merely because it uses more coloured cubes.
The 2021 seasonal field ERT study used a three-dimensional array and direct moisture probes to interpret hydrological pathways. The benefit was not the label ‘3D’ by itself but the combination of acquisition geometry, repeated measurements and independent field context. Always assess coverage and validation alongside dimensionality.
ERT resolution and soil micro-CT resolution differ greatly
X-ray micro-computed tomography can visualise small pore structures within a sampled core under its imaging limits. Soil ERT surveys a much larger volume using electrical observations and inversion. Their spatial resolution, physical targets and assumptions differ. A feature visible at micro-CT scale is not necessarily resolvable by field ERT.
The soil X-ray micro-CT guide helps explain pore-network images. ERT instead captures effective electrical properties distributed across measurement sensitivities. The two techniques can complement each other in understanding connected pores, but neither image should be treated as a pixel-for-pixel substitute.
ERT and gas diffusivity describe different transport dimensions
ERT is sensitive to electrical conduction pathways that often involve pore water and ions. Gas diffusivity concerns movement of gas molecules through available air-filled paths. Wetting can lower resistivity while reducing air-filled gas transport, but the magnitudes are not linked by one universal equation.
The soil gas diffusivity guide explains Dp/D0, air-filled porosity and tortuosity. When combining data, first estimate or measure water content and pore geometry, then consider how each affects electric and gas pathways. A resistivity model cannot directly replace an oxygen transport measurement.
Salinity change is a different hypothesis from moisture change
Imagine irrigation increases pore-water electrical conductivity by adding dissolved salts. Soil resistivity may decrease even if water content returns quickly to its earlier value. Without solution chemistry information, a persistent low-resistivity anomaly could be mistaken for a permanent wet region. The observation is electrical; the moisture interpretation depends on assumptions about ions.
A stronger experiment measures pore-water conductivity or uses a calibrated model that includes salinity effects. Time-lapse ERT can also be used deliberately to follow saline tracers under a separate transport question. The same electrical signal can serve different scientific purposes, but the analyst must name which quantity was manipulated and which mechanisms could produce it.
A dry-down experiment requires several clocks
After a rainfall event, different soil layers may dry at different rates. Time-lapse ERT can show resistivity changes over hours, days or seasons. The sampling interval determines which events are visible. A weekly measurement may miss a brief infiltration pulse, while rapid surveys may capture transient electrical states but require stable electrode and temperature conditions.
A report should state when each survey occurred relative to rain or irrigation and whether other observations were synchronous. Avoid presenting a time series as though it continuously observed every moment between surveys. Interpolation between images is a modelled construction, not an independent measurement.
Numerical practice: water content from a local calibration
Suppose a locally validated illustrative model follows rho proportional to theta raised to the power minus two, with other properties fixed. At theta 0.20, resistivity is 200 ohm metres; later, resistivity is 128. The ratio gives theta later equal to 0.20 times the square root of 200 divided by 128, or 0.25. The absolute water-content increase is 0.05 cubic metres per cubic metre.
The relative water-content increase is 0.05 divided by 0.20, equal to 25%. That is different from the resistivity decrease of 36%. This calculation is valid only within the explicitly assumed local model. If pore-water salinity changed, the same ERT data might need an alternative explanation.
A second transfer: two soils with the same resistivity
Suppose soil A and soil B each show 150 ohm metres in an ERT inversion, but B has much more saline pore water and a different clay content. Can the reader conclude they have identical volumetric water content? No. Different combinations of water, ions, porosity and surface conduction can produce similar bulk resistivity. A single scalar electrical property does not uniquely fix moisture.
The correction is to calibrate or constrain each soil domain and report uncertainty. A good explanation distinguishes the electrical equivalence from the hydrological nonuniqueness. This skill transfers to many inverse measurements: one observed response can have more than one underlying cause.
A checklist for reading a published ERT figure
Identify array geometry and whether the displayed image is raw apparent resistivity, a pseudosection or an inversion. Read the colour scale, units, reference time and any logarithmic transformations. Check the data quality controls, inversion regularisation and sensitivity at depth. Then locate independent moisture, temperature or pore-water chemistry observations.
Only after completing this chain should a reader interpret a conductive band as possible wetting, preferential flow or a geological horizon. If a key detail is absent, state the uncertainty instead of reconstructing a method from the artwork. Good technical translation explains why a conclusion follows, not merely what the image looks like.
Common interpretation errors and how to repair them
One error is calling a colour map direct moisture imaging. Repair it by distinguishing electrode measurements, inversion and calibration. Another is equating apparent resistivity with the resistivity of a single underground point. Repair it by explaining electrode sensitivity. A third is assuming that lower resistivity always means more water. Repair it by considering pore-water salts, temperature and clay.
Further mistakes include comparing maps with different colour scales, treating the deepest pixels as equally resolved, interpreting smooth inversion boundaries as precisely measured fronts and converting resistivity percentages into water-content percentages without a model. Each error can be corrected by tracing the plotted number back through its measurement and mathematical transformations.
Questions readers often ask
Does ERT measure water content directly? No. Are apparent resistivity and inverted resistivity identical? No. Does wetting often reduce resistivity? Yes, under common conditions, but chemistry and texture also matter. Can temperature affect the result? Yes. Does a low inversion residual prove a unique underground model? No. Are 2D and 3D acquisitions equivalent? No.
Can ERT reveal likely preferential flow zones? It can support such interpretations under suitable time-lapse design and validation. Is a colourful boundary a precisely measured wetting front? Not necessarily. Does a 40% resistivity drop imply a 40% water increase? No. Why measure moisture probes too? They can calibrate and test the relation between electrical properties and water.
Primary research and connected study routes
The time-lapse soil ERT and moisture study in Catena provides an original soil survey integrating ERT with probes and soil-profile observations. The 2026 alpine hillslope ERT study gives an example of calibrated relationships between soil moisture and resistivity. A vadose-zone preferential-flow investigation illustrates why repeated geophysical surveys can be informative about hydrologic pathways.
For related soil measurements, read soil X-ray micro-CT segmentation, soil NMR pore-water signals, soil gas diffusivity and cross-linguistic meaning. They address different observable quantities and their limits.
The final rule for ERT interpretation
Electrical resistivity tomography reconstructs subsurface electrical property patterns from electrode observations under a numerical model. Time-lapse changes can help investigate soil wetting, drying and flow pathways, but their hydrological meaning depends on soil chemistry, temperature, inversion resolution and a suitable resistivity–water calibration.
A careful reader preserves the entire chain: electrode geometry, voltage and current, apparent resistivity, inverse model, uncertainty, temporal comparison and independently justified water-content interpretation. The resulting explanation is readable precisely because it does not mistake an attractive image for a complete photograph of underground water.