Summary of 1996 Midyear UTAM Reoport

This report summarizes the 1996 midyear research efforts of the UTAM consortium.

Research accomplishments for 1996 include:

• Migration tests on 3D field data from Alberta, Canada demonstrate that quasi-Monte Carlo migration (QM) is no more than twice as efficient as regular grid migration (RM). This result of Chen is inconsistent with Zhou's previous work which showed that QM applied to the W. Texas data was 2-4 times more efficient than RM. Chen conjectures that QM was not very efficient for the Alberta data because of the restricted source-receiver configuration. No more than 100-200 receivers were active for any one shot, and so prevented an accurate simulation of quasi-Monte Carlo traces. Compare this to the W. Texas data where thousands of geophones were active for any shot.
• We describe a 3-D seismic experiment (about 100,000 traces) we performed over the Oquirrh Fault in May-June, 1996. The objective is to assess the effectiveness of the 3-D QM method and 3-D tomography. Processing steps are discussed and we hope to have first results by Feb., 1997.
• Responding to a request from a sponsor, we derived the resolution limits for VSP migrated sections. The theoretical vertical resolution limits agree well with the numerically computed limits for a range of practical VSP geometries. In contrast, the theoretical horizontal resolution limits did not agree well with the numerically computed limits. We are currently exploring reasons for this mismatch. Go to our homepage "http://utam.gg.utah.edu" and click on "Resolution Calculator" to compute VSP, CDP and crosswell resolution limits.
• The Greens functions for the equations of zero-offset and non-zero offset migration are derived under the far-field approximation. By specifying a homogeneous velocity model with irregular layers, these Greens functions can be used to quickly compute the migrated section without having to compute the forward-modeled data. We can now rapidly evaluate how different source-receiver geometries affect the migrated section. Future work should explore the extension of this Green's function to inhomogeneous media, summed prestack migrated sections, discrete arrays, waveform inversion and remigration.
• Adaptive grid finite-difference schemes are developed for 2.5-D and 3-D viscoelastic wave propagation. The instability in the 3D scheme reported at the 1995 UTAM meeting has been eliminated. For some models the 3D adaptive scheme is more than 4 times faster than the uniform grid finite-difference scheme. The axisymmetric scheme appears to work well without the presence of a borehole. If a borehole is included then the response outside the borehole and the tube wave along the borehole appear to be properly computed. However, there appears to be some numerical dispersion generated near the source. We are looking for a cure to this artifact.
• A new tube wave elimination scheme is presented, and some preliminary results are presented. The key idea closely follows that in the "Self Imaging Oil Fields" paper presented at the 1995 annual UTAM meeting. This Fall we hope to perform our first "Self Imaging Oil Field" experiment.
• A Conjugate Gradient (CG) migration method is developed to eliminate coherent noise from data. Theory and synthetic results are presented. We hope to have field data validation tests completed by Feb., 1997.

The 1996 annual meeting will provide a more complete analysis of the topics given in this year's mid-year report. We also expect to see some new research topics initiated prior to the annual meeting. Some new news:

• We recently accepted a new PhD candidate in Sept. of 1996: Hongchuan Sun.
• Fanlin Meng worked for Phillips this summer, Tamas Nemeth consulted for Amoco for several days, Dave Morey worked for Exxon this summer, Zhaojun Liu got married in China and so did Yue Wang. I hope all students benefited greatly by their summer experiences.
• I would like to point out our new homepage "http://utam.gg.utah.edu" where the midyear report (click on "UTAM Reports") is on-line with the username "pumpkin" and the password "pie". You might also check out our new electronic books (click on "Electronic Books") or the resolution calculator (click on "Resolution Calculator").
• I am also requesting from the UTAM members case history contributions for the GG522 or GG527 electronic books. If you are willing to write a case history (e.g., field data case history using different migration or predictive decon techniques) and have it included in the electronic book I would greatly appreciate it. Your name and company will be highly acknowledged. If you have any questions or comments please call (801) 581-4373 or e-mail (schuster@mines.utah.edu) me.
• By Sept. 29, 1996 we will have a Beta version of a "ProMax" modeling algorithm in the "Codes" section of our homepage. We welcome feedback on improving the ProMax interface. By the date of our annual meeting I expect that many of our 2-D forward modeling codes will be hooked up to Promax and given to the 1996 sponsors. We will soon purchase GOCAD and I hope to interface GOCAD with our forward modeling codes sometime in 1997. I will send everyone an e-mail reminder after the new ProMax code is in place.
• As a reminder the annual UTAM meeting will be held Feb. 13-14, 1997 at the University Park Hotel. Detailed information will be mailed to you later in the year. I look forward to everyone being at the meeting and hope I can get feedback on research directions, improving the ProMax modules, and help with the electronic books.

Jerry Schuster (Schuster@mines.utah.edu)