What is an air reflection in GPR data?

December 22 2021

Sometimes, something known as air reflections will appear in your GPR data. What is it? And how can you differentiate between an air reflection and a reflection originating from actual buried objects?

Air reflections are created when a part of the GPR signal isn’t transmitted into the ground but through the air instead. This part of the wave will bounce off nearby trees, houses, fences, cars etc., and appear as hyperbolas or layers in your data set. Air reflections can mask real, but weaker, subsurface signals, or they can create anomalies that will mistakenly be interpreted as a real objects or layers in your data.

Air reflections created as layers are harder to see, while hyperbolas can often be distinguished due to its wider shape. This wider shape is a result of the velocity of electromagnetic waves in the air, which is a much faster media compared to ground. When doing a simple hyperbola fitting the difference is velocity is clearly seen. And, of course, you can use the nanoseconds to calculate the distance to the object. In the shown case, 60 ns, gives us 18 meters to the reflected tree.

Air reflections appearing as layers: Measurements along a road. The asphalt layer is marked in red. The next clear layer is actually a metallic fence running along the road.

Air reflections appearing as point objects: Hyperbola 1, with a velocity of 96 m/µs originates from an actual point object in the ground. Hyperbola 2, with a velocity of 300 m/µs (speed of light), is an air reflection and was created when the signal bounced off an object above ground.

Air reflections tend to be more common in areas with conductive soils whereas, more or less, all the energy from a GPR pulse will be transmitted into the ground in less conductive soils.

The number or air reflections appearing in your data can also depend on the antenna type you are using (shielded or an unshielded antenna). Most high-frequency (>100MHz), ground-coupled, GPR antennas are shielded, with the aim of focusing the transmitted wave down into the ground (which will minimize the number of air reflections in your data). However, as low frequency (<100MHz) antennas are larger in size, a shield may make the antenna too bulky and heavy. As a consequence, such antennas are more often unshielded, and the radar wave is transmitted in any direction (which will make air reflections a common part of your collected dataset).

To avoid misinterpretation of results, the following should be noted while the survey is carried out (and afterwards during the interpretation):

Note possible nearby sources of potential air reflections while doing your field survey. If you measure towards or away from a large structure you may be able to identify air reflections while gathering the data.
Be especially aware of large metal objects, debris, cable bundles or poles. These often form strong reflectors and will create strong reflections in your data.
Check the velocities of hyperbolas or slopes in the resulting radargram. (Note that hyperbola fitting doesn’t work if your radar measurements are trigged by time). If the hyperbola fitting tool produces velocities of around 300 m/µs, you have identified an air reflection.

With experience from different types of surveys and an awareness of air reflections, misinterpretations can quite easily be avoided. Even though the pattern and amplitude of the air reflections can vary considerably depending on the object, the position, the antenna polarization etc., it is often quite easy to identify them if you know what you are looking for.

What is an air reflection in GPR data?

More to read