Stone Studio

GPR and borehole radar are near-surface geophysical techniques that can provide high resolution images of the top few 10’s of meters of the earth. The radar data are acquired by sending electromagnetic waves through the earth and recording the timing and magnitude of energy that arrives at the receiver antennas. A radar image is actually an image of the dielectric properties of the subsurface, as it is the dielectric constant that controls the velocity and the path of electromagnetic waves

Laboratory studies – While radar methods give us an image of dielectric properties, what we want from the data is information about the geological material such as lithology, water content, porosity, permeability. There is also interest in using radar data for the direct detection of organic contaminants. Laboratory studies on rocks and soils allow us to study the relationship between the measured dielectric properties and the materials properties of interest.

Some related publications:
Li, C., P. Tercier, and R. Knight, The effect of adsorbed oil on the dielectric properties of sand and clay, Water Resources Research, 37, 1783-1793, 2001. PDF File
Chan, C. Y. and R.J. Knight, Laboratory measurements of electromagnetic wave velocity in layered sands, Water Resources Research, 37, 1099-1105, 2001. PDF File
Knoll, M.D., Knight, R.J. and Brown, E., Can accurate estimates of permeability be obtained from measurements of dielectric properties? Proceedings, Symposium for the Application of Geophysics to Environmental and Engineering Problems, Orlando, FL, (1995).
Knight, R.J. and Nur, A., The dielectric constant of sandstones, 50 kHz to 4MHz, Geophysics, 52, 644-654, (1987). PDF File

Theoretical/Numerical Studies – Understanding the link between the dielectric properties measured using radar methods and the hydrologic properties, such as water content, is essential if radar is to be used as a hydrologic characterization tool. Once we have an understanding of the relationship between dielectric properties and material properties at the lab-scale, we need to develop the relationships at the field-scale. Stephen Moysey has built a geostatistical framework for upscaling rock physics relationships from the lab to the field (for more information) . Related to this, we are also pursuing ways of integrating radar data with other geophysical and hydrologic data sources to provide hydrologic models that minimize parameter uncertainty.

Some related publications:
Chan, C.Y., and Knight, R., Determining water content and saturation from dielectric measurements in layered materials, Water Resources Research, 35, 85-93, (1999). PDF File

Field studies – GPR images of the subsurface contain much information about spatial heterogeneity – an important aspect of accurately modeling the properties of the subsurface. Stephen Moysey is exploring ways to use neural networks to classify regions in a GPR image so as to build a large-scale model of the subsurface (for more information). This is described in the following publication:
Moysey S., J. Caers, R.J. Knight, R.M. Allen-King, Stochastic estimation of facies using ground penetrating radar data, Stochastic Environmental Research and Risk Assessment, 17, 306–318, 2003.

Some related publications:
Moysey, S., R.J. Knight, and H.M. Jol, Texture based classification of ground-penetrating radar images, Geophysics, 71(6), K111-K118, 2006. PDF File
Moysey S., K. Singha, and R. Knight, A framework for inferring field-scale rock physics relationships through numerial simulation, Geophysical Research Letters, 32, DOI 10.1029/2004GRL022152, 2005.

Forward modeling allows us to predict the way in which variations in subsurface material properties (and the related variation in dielectric properties) is captured in a radar image. James Irving developed 2-D finite-difference time-domain (FDTD) forward modeling codes of GPR for MATLAB as a part of his PH.D. research. Forward modeling code (zip file)

Once we have this model, we have shown how geostatistical methods can be used to quantify the spatial variability seen in a GPR image. Current research is focused on the Hanford site in eastern Washington, where hydrogeologists need an accurate model of the subsurface to better understand the controls on contaminant transport.

Some related publications:
Rea, J., and Knight, R., Geostatistical analysis of ground penetrating radar data: A means of describing spatial variation in the subsurface, Water Resources Research, 34, 329-339, (1998). PDF File
Tercier, P., Knight, R., and Jol, H. , A comparison of the correlation structure in GPR images of deltaic and barrier spit depositional environments, Geophysics, 65, 1142-1153, 2000. PDF File

Additional research, addressed by James Irving as a part of his Ph.D.:

(i) efficient modeling of antenna transmission and reception for crosshole GPR (for more information, PDF File)

(ii) investigation of the effects of antenna length on crosshole GPR tomography (for more information, PDF File)

(iii) strategies for improving crosshole GPR tomography at close borehole spacings

(iv) GPR diffraction velocity analysis (with Paul Sava in SEP) (for more information)

[via Stanford Uni]