First Break - Volume 12, Issue 8, 1994

Falling water tables and deteriorating water quality are the major shallow hydrogeological problems currently facing groundwater management in the Netherlands; the former phenomenon is seriously affecting agriculture and nature conservation, where as shallow contaminated groundwater is a hazard for water supply companies and is also an environmental threat. Predictive scenarios are used when planning appropriate management strategies to deal with these problems. These scenarios use models of groundwater flow. Detailed imaging and accurate characterization of the shallow subsurface (down to about 50 m) are essential for the reliability of these models and consequently for the success of groundwater management.

During conventional processing of surface seismic reflection data, both statistical deconvolution and common midpoint (CMP) stacking with normal (NMO) corrections contribute to the suppression of multiple activity. Statistical deconvolution is limited to short operators of a few hundred milliseconds, and therefore collapses only short-path multiples. The NMO stack attenuates low-velocity multiples by exploiting their greater moveout with respect to the primary reflections. Neither process addresses the removal of long-path, high-velocity multiples. Thus, the two modes of multiple suppression are not fully complementary; a subset of the multiple activity falls beyond the reach of bath, and survives to contaminate the final image.