Soil Resistivity Surveys: Wenner vs Schlumberger

An earthing study is only as good as the ground data feeding it. Model the wrong soil and every number downstream — electrode resistance, earth potential rise, touch and step voltages — is wrong with it. The soil resistivity survey is where that data comes from, and it's worth understanding before you commission one.

The short version: a four-probe survey measures how the ground's resistivity changes with depth. Take it to a maximum spacing as wide as your earth grid, in more than one direction, and record the weather — then it inverts into the layered soil model the earthing software actually needs.

Why Soil Resistivity Matters

Soil resistivity (measured in ohm-metres) sets how readily current spreads from an earth electrode into the ground. It directly drives the electrode resistance and therefore the earth potential rise during a fault. Real ground is layered — a dry, high-resistivity topsoil over a wetter, lower-resistivity layer, or the reverse — and those layers change the answer, so the survey has to reveal them. A small rod electrode "feels" mostly the shallow soil; a large grid is dominated by the deep layers. That's why the same site can need a very different electrode depending on which layer governs.

The Wenner Method (Four-Probe)

The workhorse. Four probes are driven into the ground in a straight line at equal spacing a. Test current is injected through the two outer probes and the voltage measured across the two inner probes. The apparent resistivity is ρ = 2πa·R, where R is the measured resistance. Increasing the spacing a pushes the current deeper, so a set of readings at growing spacings builds a picture of resistivity with depth.

One geometric factor, two ways of standing the same four probes. The difference between the arrays is where you put them, not what you calculate.

Worked example: at a probe spacing of a = 10 m the tester reads R = 1.2 Ω. The apparent resistivity is ρ = 2π × 10 × 1.2 ≈ 75 Ω·m at that depth. Repeat at 2, 4, 8, 16 m and you have the curve the soil model is built from.

The Schlumberger Method

A variation where the inner potential probes stay close together while the outer current probes are moved progressively outward. It's faster for deep soundings — you move two probes instead of four, and because the potential pair never moves, whatever is in the ground directly under it stays the same from reading to reading, which makes for a smoother curve. What it does not buy you is signal: at the same array length it reads a smaller voltage than a Wenner array, and the gap widens the further you spread. For deep or large sites it's often the more practical choice; for routine work the Wenner array's simplicity wins.

	Wenner	Schlumberger
Probe spacing	All four move, kept equal	Inner fixed, outer moved out
Speed (deep survey)	Slower — move all probes	Faster — move only outer
Signal at large spacing	Stronger	Weaker
Calculation	Simple (ρ = 2πaR)	Slightly more involved
Probes to move each step	Four	Two
Best for	Routine sites	Large / deep sites

The voltage an instrument has to resolve, plotted against AB/2 on logarithmic axes for one ampere injected into 100 ohm metre ground. Both curves fall as the array is widened, but the Schlumberger curve with a one metre potential spacing falls much faster than the Wenner curve: Wenner delivers roughly 7.5 times more voltage at AB/2 of ten metres and 75 times more at one hundred metres, because Schlumberger geometric factor grows as L squared over two l while Wenner grows only as two pi a. — Schlumberger reads the SMALLER voltage, not the larger. Widening MN as the array grows is how field crews claw the signal back.

A correction, and the reason for it. This page used to say Schlumberger holds the better signal at large spacings. It does not, and the geometric factor says so: the measured voltage is ρ_a I / k, and Schlumberger's k grows as L²/2l where Wenner's grows only as 2πa. At the same array half-length and the same injected current a Wenner array reads roughly 7.5× the voltage at AB/2 = 10 m and 75× at 100 m. Schlumberger is chosen because you move two probes instead of four and the potential pair stays put — and widening MN as the array grows is exactly how a deep sounding claws its signal back.

From Readings to a Soil Model

The raw output is apparent resistivity plotted against probe spacing. That curve is then inverted into a layered soil model — typically two or three layers with a resistivity and thickness each. It's this model, not the raw readings, that goes into the earthing software (SES MultiFields / the CDEGS suite) to compute electrode resistance and EPR. A good inversion is a genuine fit to the data, not a guess: if the field curve can't be matched by the assumed number of layers, that's telling you the ground is more complex and the survey may need extending.

Apparent resistivity plotted against AB/2, half the separation of the current electrodes, on a logarithmic horizontal axis, over a ground of 40 ohm metres topsoil four metres deep on 300 ohm metres beneath. Both the Wenner and the Schlumberger curves start at the top layer value of 40 ohm metres at small spacings, rise through a knee as the array widens, and approach the lower layer value of 300 ohm metres at large spacings. Plotted against this axis the two arrays agree to within about six per cent across the whole sweep. — Both arrays on AB/2, which is the only fair comparison. They agree within about 6 % — the choice between them is a field decision, not a different answer.

One Reading Is Not a Soil Model

This is the part worth sitting with. A single four-probe reading returns one number, and that number is an average weighted over everything the current reached — so over layered ground it is usually neither layer. Take 40 Ω·m of topsoil 4 m deep over 300 Ω·m beneath: a Wenner reading at a = 10 m comes back at about 100 Ω·m, a value the ground never has at any depth.

A two-layer ground of 40 ohm metres topsoil four metres deep over 300 ohm metres. One Wenner reading at ten metre probe spacing over it returns about 100 ohm metres, a number that is neither layer. Beneath, the same 625 square metre earth grid is sized on each of the three figures in turn: on the 40 ohm metre topsoil it would be 0.80 ohms, on the 100 ohm metre reading 2.00 ohms, and on the 300 ohm metre lower layer 6.00 ohms. — The reading is real, but it is an average the ground never had. Everything downstream then treats it as if the soil were uniform.

And it does not stop there, because every electrode formula downstream assumes uniform soil. The same 625 m² earth grid sized on that reading comes out at 2.00 Ω; on the topsoil alone it would be 0.80 Ω and on the ground beneath it 6.00 Ω. That is a factor of 7.5, and it propagates straight into the earth potential rise — 400 V, 1,000 V or 3,000 V for the same fault current on the same site.

Which is why the answer is the curve, not the reading — and why the model that goes into the software has layers in it rather than one figure. Note too that the fit is not unique: more than one combination of layer resistivities and depths can reproduce the same set of readings, so an inversion is an interpretation and it is worth knowing which one you were handed.

What a Good Survey Delivers

Readings taken to a maximum spacing at least as large as the electrode system you're modelling — too short and you miss the deep layer that dominates a large earth grid.
More than one traverse, ideally in different directions, to catch lateral variation.
Recorded probe spacings, instrument, date and recent weather (soil moisture strongly affects results — a survey after a drought reads very differently from one after rain).
Notes on buried services and fences that could distort readings, kept clear of the array.

Common Pitfalls

Too-short maximum spacing on a large grid, so the governing deep layer is never sensed.
A single traverse treated as representative of a varied site.
Surveying along buried metalwork (pipes, fences, other electrodes), which short-circuits the reading.
Ignoring the season — a summer survey used for a design that must be safe in dry conditions.

What BS 7671 Says About This

Less than you might expect, and the little it does say is pointed. Regulation 643.7.2 requires that where the earthing system incorporates an earth electrode, “the electrode resistance to Earth shall be measured” — measured, not calculated. Its NOTE allows the external earth fault loop impedance to stand in where a measurement of R_A is not practicable.

Regulation 542.2.4 is the one that earns this article its place: the type and embedded depth of an electrode “shall be such that soil drying and freezing will not increase its resistance above the required value”. That is BS 7671 naming the seasonal problem a survey exists to characterise — and it is a question one fair-weather reading at one spacing cannot answer. On TT, Regulation 411.5.3 caps the product R_A × I_Δn at 50 V, and Table 41.5 NOTE 2 warns that an electrode resistance above 200 Ω “may not be stable”.

What BS 7671 does not supply is any method for the soil measurement itself. The NOTE to Regulation 542.2.2 points at BS 7430 for further information on earth electrodes, and the four-probe field procedure lives in IEEE Std 81. Neither document is held here, so both are named and neither is described.

One trap worth flagging: BS 7671 does contain a table of “soil resistivity” — Table 4B3, the rating factor C_s — and it is a different quantity entirely. That one is thermal resistivity in K·m/W, used for the current-carrying capacity of buried cables. Two properties of the same ground, two units, two parts of the standard.

A left-to-right chain: a Wenner survey reading of about 100 ohm metres feeds a 625 square metre earth grid with 440 metres of buried conductor, which gives an earthing resistance of 2.00 ohms to remote earth, which multiplied by 500 amperes returning through the soil gives a 1,000 volt earth potential rise. Beneath, the same chain is run on the three candidate soil figures, giving 400 volts, 1,000 volts and 3,000 volts. Same fault current, same grid, same site. — The earthing study's headline number is one survey figure with three more calculations stacked on it. None of them carries the uncertainty forward.

What This Means for Your Study

If you're commissioning an earthing study, get the soil survey specified properly up front — it's the cheapest way to avoid a re-visit. We can advise on the survey scope before it's done and interpret the data into a layered model afterwards. For background on what the model then produces, see what is earth potential rise.

Frequently Asked Questions

What is the formula for Wenner soil resistivity?

For an equally spaced four-probe (Wenner) array the apparent resistivity is ρ = 2πa·R, where a is the probe spacing in metres and R is the resistance the tester measures. Repeating at increasing spacings gives apparent resistivity versus depth.

How large should the maximum probe spacing be?

At least as large as the earthing system you are modelling — roughly the diagonal extent of the earth grid. Wider spacings sense deeper soil, and a large grid is dominated by the deep layers, so a survey that only goes to a few metres will miss the layer that governs the result.

Does the weather affect a soil resistivity survey?

Strongly. Soil resistivity depends on moisture, so a survey after a long dry spell reads much higher than one taken after rain. Record the date and recent weather, and where the design is sensitive, consider the seasonal worst case rather than a single fair-weather reading.

Wenner or Schlumberger — which should I use?

Wenner (equal spacing) is simplest and is the routine choice. Schlumberger (inner probes fixed, outer probes moved out) is faster for deep soundings because you move two probes instead of four, and the fixed potential pair gives a smoother curve. It does not give a stronger signal — at the same array length it reads a smaller voltage than Wenner, which is why the potential spacing is widened as the array grows. Both feed the same layered-soil model, because they are the same measurement at different geometry.

Why take more than one traverse?

Ground varies laterally as well as with depth. Traverses in different directions (and offset locations) reveal that variation, so the soil model reflects the whole site rather than one lucky or unlucky line.