Oil & Gas - Maximising Recovery Programme (formerly SHARP) IOR Views

Introduction

Around 40 wells within the Palaeocene West of Shetlands were positioned on an amplitude or AVO anomaly with only nine encountering significant hydrocarbons. Analysis shows that most of the wells that failed to find hydrocarbons were located on poorly understood amplitude anomalies (various lithologies including igneous), AVO artefacts and spurious DHIs (direct hydrocarbon indicators), which also include multiples. A large number of these failed wells were positioned on AVO or high amplitude features interpreted to coincide in many cases with the termination or up dip limit/pinch-out edge of a sandstone interval. Furthermore, AVO work carried out by the companies on these features implied that a hydrocarbon accumulation was present. For a number of failed cases, the cause of the AVO or high amplitude features was poorly understood.

Complications in AVO response due to overlying condensed sections or variations in rock property can significantly reduce or even destroy AVO responses. Hence AVO studies cannot be used as the only key measure of prospect risk, but they must be combined with other techniques.

Foinaven

Foinaven displays a classic and easily understood Class 3 AVO response, which, if seen within an exploration prospect, would be significant in reducing prospect risk. Foinaven is also an excellent example of a soft/negative acoustic response that increases with offset angle. However, in other cases in which hard shales overlay hard sands, the far offset is usually a negative response that actually dims with offset.

Figure 1: Foinaven 3D Seismic Line and Interpretation

Hydrocarbon-saturated sands generate strong seismic amplitude anomalies and help to define the extent of the pools.  Figure 1, from Foinaven, shows a combination structural/stratigraphic trap in T31 to T34 Vaila sandstones, whilst the amplitude anomalies generally conform to the structure.  The nature of the amplitude anomaly can be further demonstrated through AVO cross plotting.  Essentially this is a scatter plot of two AVO attributes, usually zero offset P-wave reflectivity (A) against the AVO gradient (B).  Figure 2 shows the T34 (left) and T32 (right) Vaila sandstones data plotted in this manner.  Both plots clearly illustrate Class 3 AVO response (both A and B negative), which indicates the presence of hydrocarbon.  The response on the left is due to a thin gas cap within T34 Vaila sandstones and on the right there is clear separation of shales from deeper T32 hydrocarbon bearing Vaila sandstones.

Figure 2: Foinaven AVO Cross Plots

Assynt

Figure 3 (left) shows a 3D seismic line for Assynt and the anomaly targeted by well 204/18-1 (2001).  On the right is a schematic Assynt-Foinaven geoseismic section.  Prior to well 204/18-1, Assynt had been interpreted as a direct fairway analogue of Foinaven, but compared to Foinaven, there is no evidence of true amplitude conformance with depth.

Figure 3: Assynt 3D Seismic Line and Schematic Assynt-Foinaven Geoseismic Section

Post-well AVO analysis (E Liu, personal communication) of the Assynt amplitude anomaly (Figure 4 – left) shows a fundamental difference to the operators AVO analysis, which pre-drill suggested the presence of hydrocarbons (Class 3 type AVO). Actually, on the near and mid offset stacks (375-2241 m) the “Assynt” amplitudes are quite strong, however, on the far offsets (2241-3174 m) the amplitudes are much weaker. The AVO and various attribute analysis conclusively shows no evidence of hydrocarbon presence. More significantly, the analysis of the Assynt amplitude anomaly represents a Class 1 type AVO. In a geologic/geomorphologic context, the strong amplitudes are mainly confined to channels that show evidence for incision into the underlying strata. They are deduced to be recording a significant contrast in rock properties between the high-porosity channel fill and the surrounding sediments, but without supporting geological evidence this should not have been interpreted as conclusive proof of hydrocarbon charge in the prospect.

Figure 4: Assynt (204/18-1) and 204/17-1 AVO Cross Plots

Approximately 9 km west of Assynt, a more recent well 204/17-1 (2003) was positioned on an AVO anomaly. The pre-drill AVO work carried out by the operator for this well suggested a Class 3 type response. However, the anomaly was caused by the interface of Volcaniclastic tuffs and underlying sandstones. Moreover, the gathers show no increase in amplitude with offset and the cross plots without doubt demonstrate a Class 1 type AVO anomaly (Figure 4 – right).

Conclusions

AVO methods can in certain cases add reliable constraints to quantitative reservoir characterisation, if underlying concepts and how to apply the technology is understood. Much of the AVO work was carried out on 3D seismic datasets that had angle offsets of up to 35 degrees, realistically reliable for AVO analysis to a sub-sea depth of approximately 2 km.  At least 75% of the drilled AVO anomalies were at depths greater than this.

For increased accuracy and confidence in AVO analysis at depths greater than 2 km, there is a need to acquire seismic data with offsets longer than 3 km. Despite the pitfalls, the use of AVO/DHI has been fairly widespread West of Shetland since the early 1990s. In parallel to carrying out AVO analysis there is a fundamental need to also understand the key geologic ingredients that represent a successful trap. By doing so this will inevitably contribute to the success of future exploration utilising AVO techniques.

Acknowledgments

The author would like to thank Enru Liu, Mark Chapman and Ian Andrews of BGS for their help in this study, and also BP, Chevron, Total and Veritas for supplying data.