45408-AC8
Reflectivity from Noise
We have completed the first year of our research titled "Reflectivity from Noise".
There are four new developments sponsored in our ACS grant.
We utilize scattered energy to go beyond the law of square root (N) signal-to-noise enhancement, where N is the number of geophones that record seismic data. Our numerical experiments validate that the signal-to-noise ratio can be improved by square root (NM) where (M=T/To). Here T is the total recording time of a trace and To is the dominant period of the source wavelet. Two 120-channel field tests were conducted, one in a Utah steam tunnel and another in Arizona. The application for this new methodology is for a variety of uses, including detecting trapped miners in lost mines and for detecting the location of hydro-frac sources.
We also validated the super-resolution properties of multiple scattering in a seismic field experiment, perhaps the first time this has been done. I (Schuster) have included the results in my new book "Seismic Interferometry" which will be published sometime in 2008.
We have developed a new method to image VSP and SSP data using the interferometry concept. We use the VSP data to migrate below salt without needing to know the velocity model. This is an important improvement over the previous breakthrough of Calvert et al. (2004) who created virtual seismic sources from VSP data but were restricted to imaging data around the well. We have broken that restriction. An extended abstract discussing all these results is accepted to be presented in the 2008 SEG annual meeting committee.
We recently were given 17 minutes of 3D passive seismic data from the company "Microseismic" and 8 hours of 3D passive seismic data from the company “Aramco”. Our results from the first set of data (17 minutes) are very encouraging as we are able to compute virtual direct waves and surface waves. There is strong evidence that we have extracted reflections as well. The second set of data contains surface and downhole records as well as synthetic records; we will start processing this set of data in August 2008.
In this research, we were able to extract inexpensive RVSP (reverse vertical seismic survey) and SSP (surface seismic survey) surveys from passive data. The passive seismic data were collected at Tooele Army Depot (TEAD), Utah using a crane-driven hammer drill as a source and two overlapping lines of geophones along the ground to record seismic data. Despite the high level of ambient noise at this army base, the raw data traces show clear reflection arrivals even though the events overlap from different source excitations. The traces were autocorrelated and a gapped prediction error filter was applied to these data and the result was a RVSP shot gather with no overlapping events and a high signal-to-noise ratio. These RVSP records were then correlated and summed for different source positions in depth to give a virtual surface seismic profile data. Strong coherent surface waves prevented the formation of visible reflection arrivals so the next step is careful surface wave filtering. Nevertheless, the clear RVSP records obtained after deconvolution of 10-second records suggest the possibility of using hammer-drill sources to collect 3D RVSP data in environmentally delicate areas where vibroseis trucks are not allowed (e.g., towns or cities). If the surface waves are removed then it is likely that these RVSP data can be used to extract a virtual 3D SSP survey.
We have acknowledged the support of ACS in all publication that are mentioned in this report: "The American Chemical Society is acknowledged for their 2007-2009 grant ACS-45408-AC8."
