Datapages, Inc.Print this page

Gonzalez, Ezequiel F.1, Tapan Mukerji1, Gary Mavko1
(1) Stanford University, Stanford, CA

ABSTRACT: A Practical Inversion Methodology for Multi Offset P-to-S Elastic Impedance (PSEI): Application Using Well Logs and Synthetic Seismic to Discriminate Fizz Water from Commercial Gas

We present a practical methodology for applying P-to-S converted waves “elastic impedance” (PSEI) using logs and synthetic seismograms. The computationally economical procedure inverts for PSEI from PS seismic data using off-the-shelf commercial software designed for normal-incidence stratigraphic inversion of PP data. The key is to generate pseudo-Vs and pseudo-density logs in pseudo-depth units.
Differentiating fizz water (low gas saturation) from higher gas concentrations is difficult with only PP data. The abrupt reduction in Vp with the first few percent of gas dominates the PP seismic response. Usually only the presence of gas but not the saturation can be detected. In contrast, we find that PSEI monotonically decreases with gas concentration, responding to density changes. Consequently, PSEI can potentially differentiate better high vs. low gas saturations.
The fine-grained, poorly consolidated sandstones of the study were deposited in shallow to shelf marine environments. Although these sandstones have both commercial and non-commercial gas with similar PP response, the available logs only sample fully water-saturated zones. To validate the proposed methodology, PSEI was inverted from synthetic multi-offset PS traces computed from Vp-Vs-density logs simulated for different water-gas saturations. Using PSEI for two incidence angles, we obtained statistical classification errors of less than 10% in distinguishing “fizz water” (defined a-priori as 0.1<Sg<0.2) from “gas” (Sg>0.5). This represents an average 20% uncertainty reduction compared with using only PP attributes (-, -µ,  Ip-EI). One advantage of using two PSEI attributes, instead of a combination of PP-PSEI is that the PP-PS data time matching is avoided.


AAPG Search and Discovery Article #90026©2004 AAPG Annual Meeting, Dallas, Texas, April 18-21, 2004.