Geophysical diffraction tomography (GDT) was first suggested in 1984 as a quantitative method for high-resolution subsurface imaging. In this early form, the imaged quantity is the so called ''object function'' which is linearly related to the square of the complex refractive index. More recently, two algorithms have been proposed for separately imaging the spatial variations of the more fundamental mechanical properties of density and compressibility based on the GDT approach, acoustic measurements, and the exploitation of multiple frequencies. It was concluded in these works that these algorithms would not be sufficiently robust for the inherently noisy field applications. The postulated problems result from the existence of the density gradient appearing in the governing equation.

In this study, these variable density algorithms were applied to acoustic data acquired at a site known to contain the buried skeletal remains of a dinosaur. For one of the algorithms tested, images of object function and normalized density and compressibility of a buried bone were reconstructed that exhibited image sharpness consistent with the frequencies employed. In addition, the images of this bone, that was subsequently excavated, properly render its size and location and yielded qualitatively correct density and compressibility. ¿ American Geophysical Union 1993