Using Laboratory Measurements to Model Frequency Dependent Effects on GPR
Michael H. Powers and U.S. Geological Survey
Event name: 5th International Congress of the Brazilian Geophysical Society
Session: Use of Ground Penetrating Radar - GPR
Publication date: 07 November 1997
Info: Extended abstract, PDF ( 2.46Mb )
Using a frequency dependent, 2.5D, ground penetrating radar (GPR) modeling program and laboratory measurements of conductivity and complex, frequency dependent dielectric permittivity and magnetic permeability, the GPR response to a simple buried pipe model was generated for two sands in dry and wet conditions. Three different center frequencies, 500 MHz, 200 MHz, and 50 MHz were used on models that were appropriately scaled in size to represent entirely equivalent resolution problems. If the assumption of a stationary wavelet were accurate for these media, the responses to the different size models would be exactly the same for each different, but corresponding, input frequency. Instead, the different frequency models show changes in phase and amplitude, according to the frequency dependent propagation through the media. In addition to showing the effects of frequency dependence on the GPR response, this study also shows the great variability in electrical properties of different sands in dry versus wet conditions. Although these two sand samples were both collected in the Lechuguilla Desert near Yuma, Arizona, less than a kilometer apart, one becomes much more conductive when wet than the other. The combination of the above effects can thoroughly confuse a GPR practitioner.