Introduction

The autocorrelation conjecture of Claerbout (1979) is a long standing idea that motivating passive seismic imaging that has yet to either bear fruit or be put to bed. Claerbout et al. (1988) begins a suite of articles in these consortium reports by defining the scope of the 1989 SEP passive seismic data acquisition effort that is then subjected to increasing interrogation by many former students. The results of this effort and it's many conterminous developments are sprinkled within the SEP reports until the thesis of Cole (1995) summed up much of the effort. Articles by James Rickett and co-authors beginning with Rickett and Claerbout (1996) then resurface the topic upon receipt of a solar seismic data set, that not only met the requirements of the passive experiment, but returned excellent results.

Taking a lead from takes a lead from Claerbout (1968) where it is shown that the autocorrelation function of the transmission series of the earth is directly related to a reflection seismogram, Schuster et al. (1997) develops the mathematics behind prestack migration of autocorrelograms. Schuster and Rickett (2000) then go on the generalize that work outlining the imaging conditions for both reflectivity and source location of correlation datasets. Anstey (1964) gives a very thorough treatment of the fundamentals of correlation techniques in general.

Pre-dating all of the above is a patent application submitted by Weller (1969) describing this exact process of collection and correlation of passive seismic receiver stations. Weller (1969) describes a 2-dimensional acquisition in the Gulf-coast region as returning convincing sequence boundary reflections from recordings on the order of eight hours long.

Recognizing that this imaging methodology will involve very long time records, the work of Kostov (1990) is applicable in handling spurious and coherent noise patterns endemic to recording stations with long residence periods. This thesis treats the closely related topic of seismic imaging utilizing a turning drill-bit as a source (also treated in Cole (1995)). Specific treatment of the handling and shaping of data in a pre-processing step as well as the multichannel and areal nature of his experiments are beneficial to the development of this effort. This same topic has recently been advanced through the work of Yu and Schuster (2001) that explains migration of crosscorrelogram data collected during seismic profiling while drilling experiments as an application of the previously mentioned Schuster and Rickett (2000) derivations.

While the ultimate goal is to acquire another passive seismic data set, this paper will begin to describe investigation of this subject through the manipulation of existing data sets from the seismology community. Utilizing the capabilities of SEP3D to handle irregular datasets, I will attempt to apply the imaging methodology as described below on the 1998 Santa Clara Valley Seismic Experiment (additionally, a 1997 USGS experiment at the Kilauea volcano, Hawaii Almendros et al. (2001) exists that features several areal arrays of seismometers).