Effects of available long offset and random noise on simultaneous-AVO inversion

Sergio J. Romahn, Kristopher A. Innanen

The purpose of this work is to understand how the maximum offset available and the level of noise impact together the performance of simultaneous-AVO inversion. A low 30-m-thick gas sand reservoir, with a Class-III AVO anomaly, constitutes the geological framework. The evaluation is focused on the ability of separating the gas reservoir from the background (shale layers and brine sands) in a P- vs S-impedance plot. We took the original logs from the well as reference to measure the root mean square error (RMSE) of the inverted P- and S-impedance logs. The combined error in a P- vs S-impedance crossplot is represented by the multiplication of the inverted P- and S-impedance errors. We plotted the error as a function of noise and maximum angle in order to observe the combined effect of these two factors. Based on the reservoir discrimination, we classified the combined effect of maximum offset available and noise in four categories: best, acceptable, risky and non-acceptable results. The best results are produced with S/N greater than 7 and maximum angles between 40 and 45 degrees. Acceptable results arise for a S/N between 3 and 7 and maximum angles greater than 30 degrees; and with maximum angles between 30 and 40 with S/N greater than 7. Risky results are produced with S/N lower than 3 and angles smaller than 30 degrees. The results are not acceptable if we have maximum angles of incidence smaller than 25 degrees. We also observed that angles greater than 45 degrees may deteriorate the inversion result.