A Soil Conductivity Map Is Not an RF Performance Map

A colourful map can make a difficult RF subject feel intuitive. That is useful—until the picture quietly changes the quantity being discussed.

Consider a map built from modelled geological resistivity. Resistivity is converted to conductivity, the most conductive terrestrial category is declared the reference, and every poorer category is assigned a negative number in decibels. The legend then asks how many dB the operator is “leaving on the table”.

The accompanying caveat may insist that these numbers are only an illustrative index and not predicted antenna gain or loss. Unfortunately, the title, colour scale and “dB loss” language have already told the reader otherwise.

The issue is not that soil conductivity is irrelevant to radio. It can be extremely important. The issue is that a ratio of conductivities is not a ratio of radiated power, field strength, antenna efficiency or received signal.

The arithmetic is easy

Suppose the chosen reference conductivity is 50 mS/m. Define an index:

Iσ = 10 log10(σ / 50)

The resulting values are:

The calculation is internally consistent. The interpretation is not.

Conductivity is not a power quantity, and no RF law says that received power or field strength varies in direct proportion to soil conductivity. Choosing the factor 10 creates a private logarithmic scale. Choosing 20 would double every number. Nothing in the map itself tells us which multiplier corresponds to an observable radio result, because no propagation or antenna transfer function has been applied.

Calling the result an “index” can make the arithmetic honest, but it does not make “loss”, “ground quality” or “leaving dB on the table” honest descriptions of that index.

dB needs a defined ratio

A decibel does not identify a physical quantity by itself. The engineer must state what ratio is being expressed.

• For a power ratio: 10 log10(P2/P1)
• For a field or voltage ratio under the required equal-impedance conditions: 20 log10(E2/E1)
• For antenna efficiency: 10 log10(η2/η1)
• For a defined propagation comparison: the difference between two predicted or measured field strengths or transmission losses

Those quantities have a direct physical meaning. A “conductivity dB” number has only the meaning assigned to it: one ground model contains a different conductivity value from another. It does not say how much signal changed.

This is a category error: one electrical quantity is substituted for another because both can be written on a logarithmic scale.

There is no universal “better ground for RF”

The phrase sounds harmless, but “better” requires an objective. Better for what?

The same conductive ground can therefore be favourable for one defined mechanism, neutral for another and undesirable for a third. A Beverage receive antenna over seawater is a useful counterexample to any blanket rule that “more conductive” always means “better radio”.

The BGS resistivity model is not the mistake

The British Geological Survey resistivity dataset is a legitimate geological and engineering product. BGS describes it as a 1:50,000-scale spatial model intended for applications such as the earthing characteristics of the ground.

The model uses geological classifications and expected ranges of saturation, porosity, clay content and pore-fluid resistivity. BGS derives a statistical distribution for each geological classification and reports values such as the median and percentile range for the upper 3–5 m. Field resistivity soundings and airborne electromagnetic estimates are used for verification.

That is valuable information. It is not a field measurement in every garden, and a geological polygon is not a promise that the instantaneous RF surface impedance is uniform throughout that polygon.

Several distinctions matter:

• Modelled is not measured everywhere. A classified geological unit inherits a statistical resistivity distribution; it is not sampled continuously across the map.
• Map scale is not point accuracy. A 1:50,000 source scale does not resolve the moisture, fill, drainage, buried services or soil treatment at one antenna site.
• One representative value hides a range. Seasonal moisture, saturation, pore-water chemistry and temperature can move the real value.
• RF ground is complex. Propagation calculations require conductivity together with relative permittivity, and both can depend on frequency and the effective depth sampled by the field.
• Layering matters. Superficial material, bedrock and water table can contribute differently depending on frequency, geometry and penetration depth.

Converting resistivity to conductivity using σ = 1/ρ is algebraically correct when both describe the same material under the same conditions. It does not create the missing frequency dependence, permittivity, local variability or RF model.

What ITU-R P.368 actually predicts

ITU-R P.368-10 is a ground-wave propagation prediction method for frequencies from 10 kHz to 30 MHz. Its output is field strength or the corresponding basic transmission loss for defined conditions.

The recommendation does not accept a conductivity value and return “antenna performance”. Its homogeneous-Earth method requires frequency, distance, conductivity and relative permittivity. The reference source is a short vertical monopole on a perfectly conducting plane, normalised to a specified radiated power and field strength. The antennas are on or near the Earth’s surface, and the method predicts the vertical far-field component. It also states limits concerning ionospheric reflections, antenna height and other conditions of use.

For a path containing different ground sections, Annex 2 applies the Millington mixed-path method. Each section needs its own length, conductivity and permittivity. The order and composition of the path therefore matter; a national colour beneath the transmitter cannot represent the whole route.

A table labelled as a P.368 prediction must therefore provide enough information to reproduce every number. Each result needs both a frequency and a distance, not one or the other. It also needs the conductivity and permittivity of every path section, the homogeneous or mixed-path assumption, the reference source and the precise output being compared.

The recommendation itself warns that actual ground characteristics and reception conditions vary between locations. A propagation model is a defined engineering estimate, not a certificate that a coloured patch on a geological map costs every station the same number of decibels.

Local antenna ground and propagation-path ground are different problems

Another common mistake is to merge the soil beneath an antenna with the ground along the propagation path.

Near a ground-mounted vertical, strong reactive fields can penetrate the soil and dissipate power. The radial system, return-current geometry, feedline common-mode current and antenna current distribution all influence that local loss. This is an antenna-system problem.

Ground-wave attenuation concerns the field travelling along the Earth’s surface from transmitter to receiver. Conductivity and permittivity must be considered along that route. This is a propagation-path problem.

For HF skywave, the ground beneath the transmitting antenna may shape the elevation pattern, while the distant ionosphere and any ground-reflection region control other parts of the path. This is neither the same as local soil loss nor the same as surface-wave attenuation.

One conductivity colour cannot answer all three questions.

How to make a defensible RF ground map

A scientifically useful RF map is possible, but the output must be defined before the colours are chosen.

1. State the question. For example: “What is the predicted vertical ground-wave field strength at 1 MHz and 100 km for a specified reference source?”
2. Choose the propagation or antenna model. Use P.368 for a suitable surface-wave case, or a validated full-wave antenna model for local antenna efficiency and pattern.
3. Supply both electrical ground constants. Conductivity alone is insufficient; include relative permittivity and document their frequency and seasonal basis.
4. Describe the path. For mixed terrain or land/sea routes, use the section lengths and ground constants required by the selected method.
5. Keep antenna and path effects separate. Do not present surface-wave path attenuation as antenna gain or soil loss beneath the feedpoint.
6. Name the output precisely. Use “predicted field-strength difference”, “basic transmission-loss difference”, “radiation-efficiency difference” or another measurable quantity—not generic “RF quality”.
7. Publish the assumptions and uncertainty. Include model version, reference conditions, resolution, limitations and enough detail for another engineer to reproduce the result.
8. Validate where possible. Compare predictions with controlled field-strength, current or power measurements rather than treating colour agreement as proof.

The result will no longer be one universal map for every amateur-radio question. That is not a weakness. It is the point: different RF questions have different answers.

What an honest conductivity map can say

A geological conductivity map can still be useful without pretending to predict station performance. It can show:

• where the underlying model expects geological resistivity to be higher or lower
• where local RF ground measurements may be especially worthwhile
• which ground parameters might serve as initial values for a defined propagation study
• where a mixed-path calculation may need additional segmentation
• how wide the modelled geological variation is across a region

A suitable title would be “Modelled Near-Surface Geological Conductivity”. A suitable legend would show conductivity in mS/m, perhaps with the BGS percentile range. The explanatory text could state that the data are inputs to—not outputs from—an RF model.

What it should not say is “good RF ground”, “dB loss” or “how much you are leaving on the table” unless a defined RF calculation actually produces that quantity.

Takeaways you can trust

• Soil conductivity can materially affect antennas and propagation, but the mechanism must be defined.
• 10 log10(σ/σref) is a chosen conductivity index, not an RF loss prediction.
• A caveat does not undo a legend that labels a non-RF ratio as “dB loss”.
• Higher conductivity is often favourable for a specified surface-wave path, but it is not universally “better for RF”.
• BGS resistivity data are useful geological inputs; they are not measurements of instantaneous RF performance at every location.
• Conductivity, permittivity, frequency, distance, geometry and path composition must be combined in a suitable model.
• ITU-R P.368 predicts defined ground-wave field strength, not generic antenna gain, efficiency or station performance.
• Only predicted or measured RF quantities should be compared and labelled as RF differences in dB.

In Summary

The map is not wrong because it uses geological data. It is wrong when it converts those data into an apparently quantitative RF verdict without calculating an RF quantity.

A logarithmic conductivity scale can show that two model values are far apart. It cannot tell an operator how many decibels of signal, efficiency or performance have been lost. That bridge requires physics: a defined antenna or propagation model, complete inputs and a measurable output.

Precision in the legend cannot replace definition in the model.

Primary references

• British Geological Survey: BGS Resistivity
• BGS Electrical Resistivity Model of Great Britain: what the dataset shows
• ITU-R P.368: Ground-wave propagation prediction method for frequencies between 10 kHz and 30 MHz
• ITU-R P.527: Electrical characteristics of the surface of the Earth

Want more technical RF content? Subscribe for new deep dives and lab notes.

Have a question or field observation? Contact RF.Guru.