Seismic Attribute Categories:

Examples:

Comments:

P-Wave

PP reflection Data

All SAs unless specified

S-Wave

Psv, Mode Converted

Can give Bi-refringence

Pre-Stack

AVA (Amplitude vs. Angle) AVO (Amp. vs. Offset), Attenuation (Q), Azimuth, stacking Velocities, frequency, scattering, dispersion

More data intensive; requires more seismic knowledge & analysis; Can yield: direct Hydrocarbon Indicators (DHIs), geo-pressure, anisotropy, pore throat information

Post-Stack

Horizon & Window attributes, Spectral decomposition

Smaller 3-D Data Volumes, quicker, easier, more common & convenient

Single Trace

Freq., Phase, Amplitude

Early Attributes, easier to generate and model

Multi Trace

Coherence, Dip, Continuity,

Useful for Lithology, geo-body, fracture

2-D

Pre-stack, Post-Stack

Limited to single trace or 2-D multi-trace

3-D

Geometric Volumes, Sequence Stratigraphy, Coherency, Azimuth

The biggest growth area, better well ties, Used w/ visualization & statistical software

Time Lapse (4-D)

Re-shooting 3-D, & Analyzing differences vs. time

Useful for identifying fluid changes and for Reservoir Simulation matching

Mathematical

Most Attributes: Dip, "Fluid Factor"

Math function applied to data

Extracted

Horizon or Window Attributes

Dependent on the horizon interpretation accuracy & calculated attribute volume

Wavelet

StratiMagic®, RSI®

Wavelet shape can infer facies, in Map or Volume outputs

SA Analysis methods:

Examples:

Comments:

Supervised

Interpreter Input, pattern recognition, Emerge®

Used more when well data is available, can be used with a probabilistic method.

Unsupervised

Unbiased Mathematics, PCA

Often an interesting check of assumptions

Inversion

Pre. vs. Post, Model vs. Not

Many varieties, choose "fit for purpose"

Statistics

Typically SAs are correlated with Well Properties of interest

Can be very useful IF used correctly, seismic to well ties, modeling, & scaling are critical

Neural Networks

See Figure 2

One useful method of combining data

Principle Component Analysis (PCA)

Mathematically grouping data by the eignenvalues of the covariance matrix

Coherent energy~PC1= largest eigenvalue of the Covariance Matrix. Often difficult to understand the physical significance. Can show discontinuities

Kohonen, Self Organizing Map (SOM)

RSI’s Lithann® (fig.2), Stratimagic®, SeisClass®, dGB’s Detect®

Another popular and useful method of combining data