The following expanded abstract has been submitted to the SEG to present as a talk at the 2009 convention.  If accepted, you can hear the whole talk there.

 

GrAZ 3D for volume based fluid contact detection and anomaly assessment

Michael J. Padgett,* Quantum Earth Corporation

 


Summary

 

GrAZ 3D is a volume based method for fluid contact detection and anomaly assessment.  It is the extension to 3D volume analysis of the corresponding horizon based 2D GrAZ methodology.  In a 3D GrAZ calculation, a dot product is formed between a structural dip vector field and a gradient vector field of an attribute volume.  The result is a 3D volume of dot products after final filtering.  While the horizon based method is better constrained, the requirement that a horizon be interpreted in advance limits its applicability to anomaly assessment.   The 3D GrAZ methodology has broader applicability, but as an effective derivative of an attribute volume, the artifact load is high and extensive post filtering is necessary.  Given the existence of paleo fluid contacts, in addition to data artifacts, a GrAZ 3D anomaly is best viewed as a necessary but not a sufficient indicator in an exploration workflow.  The calculation can be done in a scanning mode useful for anomaly identification before detailed horizon interpretation. 

 

Introduction

 

Before beginning the process of detailed structural or amplitude interpretation, it is useful to have indicators that lead an interpreter to focus in certain areas.  GrAZ 3D is designed to glean first-pass indicators of hydrocarbon fluid contacts, and thus provide leads to regions deserving more intensive work.  GrAZ 3D is a straightforward extension of the horizon based equivalent but without the need of a previously interpreted horizon.

 

The problems of converting a 2D horizon process to a 3D process that can be run in a scanning mode are many.  The existence of an interpreted horizon greatly increases the legitimacy of subsequent calculations, having already had both geology and interpretive noise rejection incorporated into the horizon itself.  In proceeding to a 3D scanning implementation, a goal is to incorporate some geologic reasoning and interpretive noise rejection in the computation processes.

 

 

2D GrAZ calculation

 

2D GrAZ is a method of determining whether changes in a derived horizon attribute change abruptly in the up-dip direction at a constant time or depth and, hopefully, due to a fluid contact.  A typical horizon based attribute might be the maximum instantaneous amplitude near the picked event or simple maximum negative trough amplitude.    To make this attribute change determination, the gradients of both the interpreted Z values and of the extracted attributes are computed in 2D.  This produces 2 gradient vectors at each horizon grid location.  A dot product is formed using both vectors. The dot product can be put through a Horizon Binning process to make a graph like the one shown in Figure 1, or can be mapped directly, as in the example of Figure 2.  The goal of GrAZ 3D is to duplicate a map, like Figure 2, but without needing to interpret a horizon in advance.

 

 

Figure 1:  In an 2D GrAZ plot after Horizon Binning, the X-axis is time or depth and increases to the right.  The vertical axis is the binned dot product  of the gradient of the structure map with the gradient of the seismic attribute.  In the water-leg, isotropy is inferred by a very small dot-product values to the right, down structure..

 

 

 

 

Figure 2:  A 2D GrAZ dot product can be mapped.  Shown is a 2D GrAZ calculation with contours over-lain.  The location of the anomaly (roughly north-sourth) corresponds with an inferred fluid contact.  Down-dip is to the north-east.  Note that the anomaly appears to break at likely fault locations.



 

 

 

 

 

 

 

GrAZ 3D calculation: 3 steps

 

Following the basic concept of a 2D GrAZ calculation, there are 3 steps in computing a GrAZ 3D volume.  First, a volume of structural dip vectors must be computed.  Second, a dot product of the gradient of the attribute volume with corresponding structural dip vectors must be performed.  Third, since the artifact load of this sort of derivative process is high, a sequence of filters is normally applied before loading onto a workstation (or Segy clustering to horizons).  These steps are described in more detail in reference 3.

 

Structural dip vector volume

 

There are many good methods available for computing dip vectors using 3D seismic data.  Several are described by Marfurt in reference 2.  In a GrAZ 3D calculation, the dip vectors must function as the gradients of “horizons,” in analogy to the 2D GrAZ calculation.  This means that the dip vectors must indicate the overall dip in an “up-structure” direction, in a manner that is stable and geologically reasonable.  This author has found that the dip vectors taken from a standard calculation are too unstable to use for fluid contact detection without further processing.  In particular, variations in phase near fluid contacts can cause dip vector variations in the exact location where one needs stability prior to taking the dot products.  For this reason, from a normal dip vector calculation, “structural dip vectors” are extracted which indicate the overall structural dip, focused on locations that are likely to have stable values.  These structural dip vectors are used in computing the GrAZ 3D dot product.

 

 

Dot product and artifacts

 

It is straight-forward to compute a gradient of a 3D attribute volume and then compute a dot product with appropriately selected structural dip vectors.  This is effectively a directional derivative of the seismic attribute in an up-dip direction.  In the water leg of a reservoir unit, one assumes that changes will be isotropic and cause small, randomly-placed dot product values.  This allows the inference that large dot products at a consistent time (depth) are related to fluid contacts, if other causes can be discounted by filtering or later interpretation.

 

Final filtering and interpretation

 

As effectively a directional derivative of seismic data, there are many things that can give large values.  Faults, salt faces and rapid stratigraphic changes can all produce large derivative values.  In addition, paleo fluid contacts exist and can cause significant derivative values.  Paleo fluid contacts mean that “legitimate fluid contacts” may be detected at locations where un-economic wells would be drilled.  This means that the existence of a GrAZ 3D anomaly is not sufficient to assert hydrocarbons in an exploration prospect analysis.  Conversely, the lack of a GrAZ 3D anomaly, at a prospect whose petrophysics would infer one, is very negative for the prospect.

 

Normally, before loading to a workstation, a sequence of final filters is applied.  These are designed to reduce randomly large derivatives, possible faults, and other features that one would not normally consider to be associated with a legitimate fluid contact.  Figure 3 shows a time slice through a GrAZ 3D volume after a full sequence of filters.  The “white areas” in this time-slice are “0.0.” Interpretation is normally done on time slices, since on a vertical section the anomalies may appear as small “dots.”  In a time slice, the “dot’ may have extent and extend across a fault block.  On a workstation like SMT or Geoframe, the GrAZ 3D anomalies can be picked as faults (in time slice) and then incorporated into the rest of the interpretation.  In any project involving GrAZ 3D work, one finds that the calculations are relatively fast and serve as a starting point for more extensive work to follow.

 

 

Figure 3:  A GrAZ 3D anomaly associated with the fluid contact for the perf shown. The anomaly basically spans one fault block.  It can be picked as a fault on a workstation or extractged as a horizon using a clustering algorithm.that the anomaly appears to break at likely fault locations.


 

 

Conclusions

 

The problems of converting a 2D horizon process to a 3D process that can be run in a scanning mode are many.  The existence of an interpreted horizon greatly increases the legitimacy of any subsequent calculation, having already incorporated both geology and interpretive noise rejection.  In proceeding to a 3D scanning implementation, geology and interpretive noise rejection will come later, but the need to pre-interpret horizons is removed.

 

GrAZ 3D detects locations of fluid contacts when there is a change in a seismic attribute from down-dip to up-dip, emphasizing constancy in time or depth.  As a derivative based operation, there are many artifacts that must be removed before or during the interpretation process.  Some of the artifacts are due to paleo fluid contacts, these cause GrAZ 3D anomalies to be viewed as necessary but not sufficient, in an exploration setting.

 

In any project involving GrAZ 3D work, the calculations are relatively fast and serve as a starting point for more extensive work to follow.  This technique is not a “silver bullet,” but can logically augment a normal exploration, delineation and/or development work-flow in many settings around the world

 

Acknowledgements

 

The author would like to thank OpenSpirit Corporation for permission to use the data analyzed in Figure 2 and to thank Centerline Geophysics for permission to show the results in Figure 3.