Ground penetrating radar (GPR) can be an effective subsurface imaging tool in urban environments. However, interpretation of urban GPR data is complicated by 1) a lack of information regarding the electrical properties of the subsurface materials and 2) electromagnetic (EM) wave reflections from above-ground cultural features. Reflections from elongate, profile-parallel above-ground objects (overhead power lines, metallic fences) mimic subsurface planar reflections. Diffraction hyperbolae from localized objects (parked cars, utility poles) may be mistaken for subsurface point reflectors. In conductive materials, where signal attenuation limits radar range, these air wave events are especially prone to misinterpretation. Common midpoint (CMP) soundings provide a quick, on-site method for estimating depth of penetration. Further, because the slopes of air and ground wave events differ markedly on CMP plots, CMPs quickly differentiate between above-ground and subsurface reflections without lengthy post-acquisition data processing. CMP soundings are commonly conducted to determine signal velocity following completion of reflection surveys. This study suggests that CMP soundings should precede extensive reflection surveys in urban environments, particularly where the electrical properties of the subsurface materials are unknown. In addition, since CMP surveys require separate transmitting and receiving antennas, it is recommended that urban GPR investigations be conducted with equipment capable of being used in bistatic mode.

GPR was evaluated as a method for mapping a reclaimed stream valley in Philadelphia, PA where differential subsidence of the fill material has resulted in building distress and subsequent demolition of nearly 1000 "sinking homes". Initial reflection surveys suggested subhorizontal reflectors at a depth of 5 m. However, CMP soundings indicated that actual signal penetration was restricted to less than 1 m, and that the apparent subsurface reflections resulted from profile-parallel above-ground objects. The CMP soundings ruled out use of GPR for this particular application prior to execution of lengthy reflection surveys. The GPR data from Philadelphia are compared with those from a similar urban setting in Atlantic City, NJ, where CMP confirmed actual subsurface stratigraphy in low-loss materials.