Seismic reflection studies of an artificial magma

Seismic reflection images of a solidified artificial magma body ~5 cubic meters in volume were compared to cross-sections produced independently from temperature measurements and core drilling the cooled, solidified rock. The data were obtained during in situ vitrification (ISV) of a simulated waste disposal trench at Oak Ridge National Laboratory. Four migration schemes were compared, prestack Kirchhoff time migration and three post-stack migrations: Kirchhoff, Stolt FK, and finite-difference. The images produced using prestack Kirchhoff migration were superior to images from the post-stack migrations, and detail both the top and bottom of the ISV rock. Images produced using post-stack Kirchhoff migration were contaminated by high-amplitude noise, while images from Stolt FK and finite- difference migration contained less noise but showed only the upper portion of the ISV rock.

Different field experiment designs were compared by processing subsets of the data. Almost all information concerning the geometry of the ISV rock was found in the backscattered (geophones on the same side of the ISV rock as the source) subset. Wide-angle reflection data (geophones on the opposite side of the ISV rock from the source) provided only limited information about the upper surface. Data from horizontal (radial) component geophones contained significantly stronger reflections from the ISV rock than vertical component data, although both geophone components recorded sufficient information to image the body. Combining partial images from radial and vertical data produced the most complete pictures. This work demonstrated that it is possible to produce satisfactory seismic reflection images of complexly shaped objects at relatively shallow depths when near-vertical incidence data cannot be obtained. The results from this study apply to imaging magma chambers and plutons at crustal scales, and to applied waste management problems at shallow depths.

A new procedure called seismic masking was applied to seismic reflection data recorded during the ISV experiment. The procedure masks selected volumes in the earth containing the known objects prior to stacking, and is based on the seismic response of hypothetical scattering points (masking points) located within the volumes. Portions of the data traces in each shot gather that coincide with arrival times from the masking points are muted. The procedure involves forward modeling the traveltimes for each masking point, and thus requires an accurate velocity model. The masking algorithm was tested by successfully attenuating reflections from the limestone in data acquired prior to melting. Images of the melt showed unwanted reflections from a subsidence zone above the melt, which interfered with reflections from the melt itself. Masking the subsidence zone significantly improved the seismic image for the upper portion of the melt.