ABSTRACT

With increasing enphasis on shallow, high-resolution
geophysical techniques for environmental and engineering
applications, it has become important to implement and
evaluate tools for quantitative interpretation of
ground-penetrating radar (GPR) data. Both ray and Fourier 
algorithms are viable for numberical simulation of 2-D
monostatic GPR data, but they have different characteristics.
The ray algorithm uses geometrically complicated layers,
where within each the dielectric permittivity and attenuation
are constant. The algorithm produces accurate amplitudes for
reflection but does not include wave effects such as diffractions
from layer trunctations. The Fourier algorithm uses a grided
parameterizarion in which reflections are constructed by
superposition of diffractions in a background of constant
dielectric permittivity and constant attenuation. This technique
includes all wave effects, but it does not contain the antenna 
directivities. Both algorithms are able to simulate the main 
features in two representative, but very different, field data
sets: one from a quaternary fluvial/aeolian environment and 
one from a Cretaceous marine carbonate environment.