Repercussions of available long offset, random noise and impedance contrast on AVO analysis

Sergio Romahn and Kris Innanen

ABSTRACT

The amplitude variation with offset (AVO) or angle of incidence (AVA) is sensible to several factors that may affect the feasibility of doing this kind of analysis. This work evaluates how the available long offset, level of random noise and impedance contrast affect the estimation of the AVO parameters (intercept, gradient and curvature). An oil sand, with a class-III AVO anomaly, constitutes the geological framework. Fluid replacement modelling was used in order to extent the analysis to three different impedance contrast scenarios: gas, oil and water filling the pore space. Firstly, we analyzed the AVO response with no noise and offsets related to angle of incidence up to 45 degrees. This case was taken as reference to measure the error and standard deviation in the AVO-parameter estimation when decreasing the available offset and varying the level of noise. The results, for the gas case, show that the intercept is practically not impacted by the reduction of offset and is slightly affected by the level of noise. Errors around 3% were obtained with angle traces up to 15 degrees and S/N equals 2. The gradient is strongly affected by the maximum available offset and signal to noise ratio. Errors smaller than 20% allow estimating meaningful values of the gradient. 20% of error can be obtained with several combinations of maximum angle and level of noise, for example: angle traces up to 45 degrees and low S/N of 3, or angle traces up to 22 degrees and high S/N of 15. The most affected parameter is the curvature. We would need angle traces greater than 36 degrees and high S/ N of 15 to produce errors smaller than 20%. If the S/N≤5 we would need angles traces of 45 degrees to abstain the same error. When comparing the gas, oil and water scenarios, we observe that the error increases as the impedance contrast decreases. These results may be relevant when designing a seismic survey or for time-lapse seismic purposes.

View full article as PDF (3.70 Mb)