This is a very common question. I get it all the time. Often people are concerned that they will not be able to identify archaeological features…AND THEY SHOULD BE!
In many archaeological contexts, people in the past used local materials to construct buildings or manufacture artifacts. If the remains of these items are buried in the same local soils that they were harvested from, then the contrast between the two (the archaeological material and the geology) will be minimal. Since it takes a change in GPR wave speed to create a reflection, and a change in material properties will cause a change in wave speed, if the materials are identical, then here may be no reflection of the GPR wave and thus no recording of the target.
One example of where this might be the case is if limestone was quarried to construct house floors in the past and now that limestone is buried in a karst soil. The archaeological material and the geological material might be so similar that it could be difficult to identify a reflection from the limestone house floor.
However, this is usually not the case for several reasons and it is often possible to identify archaeological targets even if the material is similar to the geology that it is imbedded in. First, preparation and use of artifacts and features generally alter the physical properties in some way and create at least minimal contrast between the archaeological target and the geology. Using the example from above, cut limestone blocks used to construct a floor will be denser than the karstic soil they are imbedded in. In other cases, where living floors were just soil, they become more compact over time creating variation from the surrounding soil.
A second reason why artifacts could potentially be identified using GPR is that an artifact might be filled with something other than the material it is made of. For example, a storage vessel made out of clay that is also embedded in clay might be filled with air or organic material that has a significant contrast with the clay itself. This principle can be applied to other situations as well. In utility locating for example, clay sewer pipes that are embedded in clay soil will likely have air or sewage inside the pipe. This would create a reflection event when the wave transitions to the air or sewage allowing the locator to detect he pipe.
Finally, archaeological targets can create variations in current conditions that are detectable with GPR. A ditch that is filled with the same soil that was originally excavated out will often be less compact and contain more air filled pores. This contrast could allow for its detection. Or, if water builds up at the bottom of a ditch, or grave shaft, or some other feature, that saturated soil will have a much higher dielectric value that causes a change in wave speed. When this happens, a reflection event will occur.
So how can you tell if it is archaeological or just geological variation…DO YOU HOMEWORK! Learn as much as you can about your site and your potential targets to give yourself the best chance at accurate interpretation.