Dielectric Constant in GPR: Permittivity and Radar Speed

What is the dielectric constant in GPR?

It is the relative permittivity of the ground, a number that runs from 1 for air to about 80 for water. It describes how strongly a material stores electrical energy and so slows electromagnetic waves. In GPR it fixes the wave velocity, v = 0.3 / √er m/ns, which converts recorded time into depth and controls how much energy reflects at each boundary.

Ground penetrating radar (GPR) is a geophysical survey technique that reads the subsurface from radar echoes, and permittivity decides how fast those echoes travel. Physicists use the term dielectric constant in a wider sense, and permittivity does vary with frequency. Across the usual GPR band, roughly 10 MHz to 3 GHz, the simple velocity rule above holds well enough for field work.

Typical values for common materials

Water content dominates the dielectric constant in GPR work: the same sand can read 4 when dry and 25 when saturated.

Ranges follow widely used published GPR tables and vary with moisture. The velocity we calibrate does not change the per-metre price; to work out your GPR survey cost, use the length of your scan lines.

Why does it matter on a real survey?

Because a wrong velocity scales every depth by the same error. Assume er 9 (0.10 m/ns) where the ground is er 16 (0.075 m/ns). An echo at 20 ns is then reported at 1.0 m when the pipe sits at 0.75 m. Our Two-Way Travel Time (GPR) guide shows that conversion step by step.

Permittivity also affects reach. Wet, high-permittivity soils are often conductive as well, so the signal dies sooner; see GPR depth of penetration for that side of the story. Seasonal changes in ground conditions across north India alter readings between the dry months and the monsoon. A velocity from one visit should not be reused blindly on the next.

A worked calibration at a valve chamber

A known depth turns the dielectric constant from a guess into a measurement. Suppose one of the water pipes entering a valve chamber is taped at 1.05 m below the cover and its echo on the adjacent scan arrives at 21 ns. Velocity is 2 × 1.05 / 21 = 0.10 m/ns, so er = (0.3 / 0.10)² = 9. Further along the street a cable echoes at 30 ns, which places it at 1.5 m. Had the software kept an assumed er of 6 (0.122 m/ns), the same cable would have been drawn at about 1.84 m. That is some 34 cm too deep, enough to put an HDD bore straight through it.

Common mistakes with permittivity values

Most errors come from treating one number as fixed for a whole job.

• Keeping the software's start-up value instead of calibrating on the site itself.
• Using one figure for a route that passes from dry verge to waterlogged median.
• Calibrating through a concrete slab and applying that figure to the soil below it.
• Carrying a dry-season value into monsoon fieldwork, or the reverse.

Each of these scales depth wrongly while the plan position stays correct, which is why a drawing can look right on the ground and still mislead the drill crew.

Where does the value appear in a Mezux report?

On every depth-bearing sheet, beside the depths it produced. Our drawings state the velocity in m/ns and the equivalent relative permittivity for each section of the route. They also give the method used to obtain it (hyperbola fit, known-depth target or both) and the date of the scan. If the value changes along the corridor, the change point is marked by chainage. Anyone re-using the data collected on that visit can then rescale our depths if the ground has changed. ASTM D6432-19, the general guide to the surface GPR method, and ASCE 38-22, which classes such depths as quality level B, both expect such a traceable record.

How do we measure it on site?

Mostly by fitting hyperbolas, and by checking against a target of known depth. We never take the dielectric constant in GPR software on trust; the usual checks are:

• curve-fitting several hyperbolas along each line and averaging them;
• scanning over a valve chamber, culvert or manhole whose depth can be taped;
• on larger projects, a common-midpoint test that moves the receiving antenna away from the transmitter.

The stronger the contrast between a buried object and the soil around it, the stronger the reflection. The reflection coefficient between two layers is (√e1 − √e2) / (√e1 + √e2), so dry sand over wet clay reflects about 43% of the amplitude, a bright, easy-to-follow line. Examples sit in the GPR Radargram Library.

For utility detection where depth accuracy counts, request a written estimate for your project.

Frequently asked questions

Is dielectric constant the same as relative permittivity?

In GPR work, yes. Relative permittivity is the precise name; dielectric constant is the older, common one.

What dielectric value should I assume for Delhi NCR soil?

None by default. Alluvial silts and sands change with moisture, so we measure velocity on each site before mapping underground utilities.

How does velocity show up on a radargram?

In the shape of every hyperbola: narrow for slow ground, wide for fast ground. See How to Read a GPR Radargram.

Does a bitumen road surface change the value you use for the soil?

Not directly. The road layers have their own permittivity, but most of the travel path to a utility lies in the soil below, so we calibrate on targets beneath the pavement.

Why can the value change after rain?

Rainwater fills pore space, and water has a permittivity near 80. Even a few per cent more moisture lowers the velocity noticeably.

Can permittivity tell a PVC pipe from a steel one?

Only indirectly. Radar sees metallic and non-metallic objects alike, but steel reflects almost all the energy, while an empty PVC pipe gives a weaker return signal of opposite polarity. Exposure is still needed to be certain.

Written by Saniya Parveen · Technically reviewed by Rahul Rai · Updated 8 November 2026