Spectral Decomposition AVO attributes for identifying potential hydrocarbon-related frequency anomalies

Abstract

Low-frequency seismic anomalies have long been a subject of interest to geoscientists involved in hydrocarbon exploration since such ‘gas-shadows’ can be a direct hydrocarbon indicator (DHI). Published studies have demonstrated evidence for them potentially resulting from increased seismic attenuation and velocity dispersion, as a result of hydrocarbon saturation. The topic gained wide interest during the 2000s with well-cited publications by Castagna et al. (2003), Ebrom (2004), Chapman et al. (2005, 2006) and Odebeatu et al. (2006), to name a few. The consensus of these studies was that hydrocarbon related frequency effects are predicted to be detectable on stacked seismic data. Furthermore, it has been suggested that low frequencies tend to show the highest sensitivity to fluid changes. (e.g. Korneev et al., 2004). The effect has been shown not only from seismic data; studies involving laboratory tests and borehole data provide similar conclusions. Overall there is general agreement that the effect exists. However the physical cause remains inconclusive. Several studies have also shown evidence that hydrocarbon reservoirs have an amplitude-versus-offset (AVO) frequency dependence (e.g. Chapman et al. (2005, 2006), Odebeatu et al. (2006), Liu et al. (2006), Ren et al. (2007), Zhang et al. (2007), Chen et al. (2008), Wu et al. (2014)). In the modelled case of gas-saturated sands, low frequencies have tended to show the greatest change in amplitude with offset (Figure 1a). The idea of using AVO and Spectral Decomposition (SD) to understand frequency anomalies and potential links to reservoir fluid content has been approached by several of these authors. The published work largely used model-based techniques to predict AVO effects for different frequencies (Figure 1b), and then applied this information to aid interpretation of anomalies observed on iso-frequencies sections. The results convincingly suggest that there are differences in the spectral content of seismic data which could be exploited for improved hydrocarbon identification. The results of these pioneering publications is the basis from which the workflow presented in this paper was conceived. Here these ideas are approached from an interpretation perspective and applied in a practical manner using commercial interpretation software.