AVO Helps Seismic Imaging In Deepwater Environments

Chuck Skidmore, Richard O. Lindsay
Diamond Geoscience Research Corp.
Tulsa

Davis Ratcliff
Diamond Geoscience Research Corp.
Houston

Amplitude and frequency variations related to offset should be analyzed routinely during interpretation of seismic data acquired in deepwater environments.

Amplitude variation with offset (AVO) in three dimensions is the key exploration tool in deep waters of the Gulf of Mexico. But application of the tool requires special care.

Three-dimensional AVO helps the interpreter understand stratigraphy and the meaning of amplitude anomalies. Used in conjunction with well log data, it can help the interpreter distinguish amplitudes related to the presence of hydrocarbons from those that result from, for example, rock-property changes within a non-hydrocarbon-bearing layer, such as a shale, or residual gas (fizz water) in high-porosity sands.

Examples from gulf

Data from deepwater environments in the Gulf of Mexico show strong, interpretable relationships between amplitude values and offset (distances between seismic sources and receivers).

Fig. 1 [112,461 bytes] shows how a map view of the subsurface changes when data are segregated according to offset distance. All the data are from 3.4 sec.

Fig. 1a is an amplitude map generated from the full data volume; that is, it contains all offsets from 300 m to 6,200 m. Strong amplitudes are in yellow and red, while weaker reflections are mapped in blue and green.

Fig. 1b shows how the same mapped horizon looks on the near-offset stack (300-1,500 m). Fig. 1c is the same horizon on the far-offset stack (3,500-6,000 m).

Two strong amplitude anomalies exist on the map generated from the conventional full stack (Fig. 1a), one updip under structural closure and one downdip. The near-offset map of Fig. 1b shows only the downdip event as an amplitude anomaly (yellow and red).

On the far-offset map of Fig. 1c the updip amplitude is brighter than it was on the full stack and conforms to structure, while the downdip amplitude is dimmer with increasing offset.

What do these changes in 3D AVO mean to the geoscientist and production engineer? How do the AVO changes affect the exploration economics of a prospect or the field development of a discovery? Rock property analysis based on log data from nearby wells provides insight into the differences in the 3D AVO responses and their significance (Fig. 2 [119,248 bytes]).

Fig. 2a shows log data and normal-incidence (zero offset) synthetic seismic traces modeled from the well information. From left to right, the log data categories are gamma ray, resistivity, porosity, density, compressional velocity (Vp), shear wave velocity (Vs), and Vp/Vs curves. The normal-incidence synthetic appears on the far right.

Three sands are highlighted in yellow on the gamma ray curve. The middle sand (Sand B) is associated with hydrocarbons and shows up as a high amplitude on the synthetic traces. The synthetic traces allow the explorationist to link rock-property data from the well with seismic data collected in the field.

The logs provide the basis for a standard seismic interpretation step that combines measurements of density and velocity to show the earth section's potential to reflect acoustic energy. The result of this step is called a reflectivity series, which shows reflection potential as a function of depth.

The interpreter mathematically blends, or convolves, a model of the sound pulse, or wavelet, with the reflectivity series to produce a synthetic seismic trace. This procedure essentially predicts the seismic response of a geologic model made from well measurements.

Usually, construction of a synthetic trace assumes zero-offset energy behavior, which is what traces exhibit on a near-offset stack seismic section. Rock-property modeling that accounts for both compressional (P-wave) and shear-wave (S-wave) velocities allows the interpreter to estimate amplitudes of reflections on non-zero-offset energy, similar to what occurs in the field. The resulting traces show what, if any, AVO effects can be expected in the area.

Fig. 2b shows an AVO synthetic generated from the well data in Fig. 2a. Offset increases from left to right. The A and B sands are highlighted. The pay sand (Sand B) exhibits an increase in amplitude with offset in a pattern called a positive AVO. Sand A, the brine-filled sand, shows a weak reflection response on the near offset, which weakens with offset in a negative AVO pattern.

Fig. 3 [108,418 bytes] shows what occurs on the synthetics when gas is substituted for brine in sand A. In Fig. 3a, the normal-incidence synthetic has a high amplitude associated with Sand A, with hydrocarbons modeled in for brine. In Fig. 3b, Sand A modeled with hydrocarbons has a positive AVO response.

The rock-property analysis of this area indicates that pay sands are bright (amplitude-related) and have a positive AVO response, while brine sands have a weak normal-incidence response with a negative AVO signature. The implications for the mapped horizons in Figs. 1b and 1c are that:

• Hydrocarbons are present in the updip, structurally trapped amplitude since positive AVO is evident.
• The downdip amplitude is associated with a strong lithologic contrast since it is bright but has negative AVO. Nothing in the rock-property modeling has that correspondence.

The data in Fig. 1 are from a Gulf of Mexico area in 400 ft of water where the depositional environment compares closely with those in the much deeper waters that have become the focus of much industry attention.

Well control

Fig. 4 [194,664 bytes] shows similar results in an area with well control. Fig. 4a is a time slice of a full 3D data volume, in an area with two wells that encountered hydrocarbons.

The offset range in this example was broken into three segments according to angle of incidence, the angle of energy arrival at a reflecting horizon measured from a line perpendicular to the horizon. Angle of incidence increases with offset and decreases with reflection time.

Changes in offset in these examples create strong changes in apparent stratigraphy and amplitude. Well control establishes that the high amplitudes result from oil and gas-bearing sands. A strong build-up in amplitude is apparent as offset increases in Figs. 4b-4d. The shape of the sand feature becomes clearer and more definitive with the use of 3D AVO maps.

Fig. 5 [206,678 bytes] shows the amplitude change with offset in cross-section view for Line A (left-to-right) in Fig. 4. Fig. 6 [253,163 bytes] shows the same thing for Line B, which intersects Line A from top to bottom at the bottom well location shown on the plot.

On both lines, the amplitude increase with offset is obvious and is associated with hydrocarbons.

Improving detail

Fig. 7 [128,818 bytes] illustrates a technique that improves detail of AVO interpretation but that is not widely used. The technique interprets maps of horizons from data volumes of low, middle, and far offset ranges. It then subtracts amplitude values of one map from those of another.

This differs from the more normal practice of subtracting near-offset data from far-offset data, which can lead to erroneous maps if there is any residual moveout in the data.

(Moveout is the hyperbolic shape of a reflection in a common depth-point CDP gather that results from the increase in reflection time associated with increasing offset. A processing step called moveout correction flattens the reflection in time so that traces in the gather can be summed, or stacked, into a single, zero-offset trace.)

In this technique, maps are of like horizons, wherever they appear in time, interpreted in data volumes segregated by offset range. The focus thus is on amplitude change associated with offset, regardless of the time shift that may result from residual moveout.

In Fig. 7, middle-range amplitude values of the horizon of interest in Fig. 4 have been subtracted from far-range values of the same horizon.

This technique highlights areal extent, potential "sweet spots," and separation between sand bodies of a horizon 10,000-11,000 ft below the sea floor. The stratigraphic detail is striking in comparison with the full-stack time slice through the same feature in Fig. 4a. The deep blue trend running along the right of the top-to-bottom seismic line is a lithologic change of some sort, possibly a fault or permeability barrier, trapping hydrocarbons to the right.

Frequency-dependent analysis

A technique called 3D frequency-dependent analysis (FDA), used in conjunction with 3D AVO, can further improve stratigraphic and geologic interpretation. It essentially improves resolution of a specific amplitude on a stack section.

Like 3D AVO, 3D FDA compares the full data volume with data volumes partitioned, in this case according to frequency bandwidth rather than by offset range.

For example, the full data set may involve frequencies of 5-70 hz. A typical FDA application would divide the range into 20 bands for comparison with the full volume. Bandpass filters accomplish the division. The interpreter selects bands that provide the most information about stratigraphy, then compares them.

Fig. 8 [106,783 bytes] compares a full-bandwidth data volume (Fig. 8a) with a volume including a middle range of frequencies. A channel sand, with distinct sand bodies, appears in the middle frequency range that was only vaguely suggested in the full-bandwidth volume.

Contours on this plot are time values. Yellow and red indicate where the thickest portions of sand bodies are and suggest the possibility of 3D estimation of sand thickness as a target of technological development.

The geologic changes apparent with frequency change here result from tuning effects and the effects associated with blocky beds as opposed to gradational beds. Tuning effects are amplitude increases that occur when reflection energy overlaps because of the close spacing of horizons.

Dividing a volume of traces into different frequencies thus improves insight into geology, sand deposition, sand geometry, and sand thickness.

Three dimensional AVO applied with careful rock-property analysis provides important information about reservoir shape and content. Three dimensional FDA further delineates reservoir shape and can help identify the thickest portions of sand bodies. Used together, the tools are powerful guides to drilling in deepwater environments.

The Authors