# Run multiple seismic attributes to visualize patterns

### SEG 2015 | Deborah Sacrey and Rocky Roden

*An excerpt:*

The main task for geologists and geophysicists is to identify and ascribe the geologic meaning to observable patterns in their seismic data. The isolation of such patterns and their use as possible identifiers of subsurface characteristics constitutes attribute analysis and can significantly impact reducing risk in hydrocarbon prospecting.

*Full Article Text...*

# Self-Organizing Maps and Interpretation

**What if you could...**

**Slide 2**

- Reduce interpretation cycle by advanced reconnaissance of your data?
- Reduce risk in drilling marginal/dry holes?
- Understand reservoir characteristics better?
- Employ an analysis to help sort through the mountains of attributes generated from your data?
- Maximize use of your seismic data during downturn in industry?

**Slide 3**

The main task for geologists and geophysicists is to identify and ascribe the geologic meaning to observable patterns in their seismic data.

The isolation of such patterns and their use as possible identifiers of subsurface characteristics constitutes attribute analysis and can significantly impact reducing risk in hydrocarbon prospecting.

Self-Organizing Maps (SOM) is a powerful cluster analysis and pattern recognition approach that helps interpreters identify patterns in their data that can relate to geological characteristics such as lithology, porosity, fluid content, facies, depositional environment, etc.

**Slide 4**

*Not your "Daddy's" Neural Analysis*

- Unsupervised neural analysis has been around for some time –but the technology has drastically changed because of increased computer power and the invention/creation of hundreds of new attributes from advanced processing of seismic data.
- This is NOT“black box”, but employs advanced understanding of various attributes and their contribution to finding solutions to specific problems in the seismic world. It can be “GIGO” if not used carefully.

**Slide 5**

*What is a Seismic Attribute?*

A measurable property of seismicdata, such asamplitude, dip, frequency, phase and polarity. Attributes can be measured at one instant in time or over a time window, and may be measured on a single trace, or on a set of traces or on a surface interpreted from seismic data.** Schlumberger Oilfield Dictionary**

Seismic attributes reveal features, relationships, and patterns in the seismic data that otherwise might not be noticed. **Chopra and Marfurt, 2007**

**Slide 6**

*Objectives for using Seismic Attributes*

- To take advantage of the seismic attribute analysis and today’s visualization technology,to minepertinent geologic informationfrom a huge amount of seismic data
- The ultimate goal is to enable the geoscientist to produce a more accurate interpretation and reduceexploration and developmentrisk.

**Slide 7**

## Most Common Seismic Attributes for Interpretation

- Instantaneous attributes
- (trace envelope, instantaneous phase, instantaneous frequency)
- Amplitude defining attributes
- (Average Energy, Sweetness, RMS)

- Coherency/Similarity
- AVO Attributes
- Inversion
- Spectral Decomposition
- Curvature

**Slide 8**

*Instantaneous Attributes*

- Reflection Strength (trace envelope, instantaneous amplitude)
- Lithological contrasts
- Bedding continuity
- Bed spacing
- Gross porosity
- DHIs

- Instantaneous Phase
- Bedding continuity
- Visualization of unconformities and faults

- Instantaneous Frequency
- Bed thickness
- Lithological contrasts
- Fluid content (frequency attenuation)

**Slide 9**

*Amplitude Accentuating Attributes*

These attributes help define how the amplitude stands out against surrounding reflectors and background events.

- DHI characteristics
- Stratigraphicvariations
- Porosity
- Lithologyvariations

- Average Energy
- Sweetness (frequency weighted envelope)
- RMS
- Relative Acoustic Impedance

**Slide 10**

*Coherency/Similarity*

Coherency, similarity, continuity, semblance and covariance are similar and relate to a measure of similarity between a number of adjacent seismic traces (multi-trace analysis). They convert data into a volume of discontinuity that reveals faults, fractures, and stratigraphicvariations.

- Cross-correlation-Based Coherence
- Semblance-Based Coherence
- Variance-Based Coherence
- Eigen structure-Based Coherence
- Gradient Structure Tensor-Based Coherence
- Least-Squares-Based Coherence

**Slide 11**

*AVO Attributes*

AVO attribute volumes are computed from prestackdata (gathers) . They include combinations of near, mid, and far offset or angle stacks and depending on approximations of the Knott-Zoeppritzequations, various AVO components. Most of the AVO attributes are derived from intercept and gradient values or equivalents. They are employed to interpret pore fluid and/or lithology.

- Intercept (A)
- Gradient (B)
- Curvature (C)
- A*B
- A-C
- ½ (A+B)
- ½ (A-B)
- Far-Near
- (Far-Near)Far
- Poisson Reflectivity
- Fluid Factor
- Lambda-Mu-Rho

**Slide 12**

*Seismic Inversion*

- Inversion transforms seismic reflection data into rock and fluid properties.
- The objective of seismic inversion is to convert reflectivity data (interface properties) to layer properties.
- To determine elastic parameters, the reflectivity from AVO effects must be inverted.
- The most basic inversion calculates acoustic impedance (density X velocity) of layers from which predictions about lithologyand porosity can be made.
- The more advanced inversion methods attempt to discriminate specifically between lithology, porosity, and fluid effects.
- Inversion is most helpful in determining reservoir quality and extent in unconventional plays

**Slide 13**

*Spectral Decomposition*

Use of small or short windows for transforming and displaying frequency spectra (Sheriff, 2005 Encyclopedic Dictionary of Applied Geophysics). In other words, the conversion of seismic data into discrete frequencies or frequency bands.

- Layer thickness determinations
- Stratigraphicvariations
- DHI characteristics (e.g. shadow zones)

- Discrete Fourier Transform
- Fast Fourier Transform
- Short Time Fourier Transform
- Maximum Entropy Method
- Continuous Wavelet Transform
- Gabor
- Gabor-Morley
- Gaussian

- Spice
- Continuous Wavelet Packet-Like Transform
- Wigner-Ville Distribution
- Smoothed Wigner-Ville Distribution
- Matching Pursuit
- Exponential Pursuit

**Slide 14**

*Curvature*

Curvature is a measure of bends and breaks of seismic reflectors.

Another way to describe curvature for any point on a seismic reflecting interface is the rate of change of direction of a curve.

Curvature is especially useful when working in the unconventional resource environment –to see fracture systems and avoid larger faults in horizontal well bores.

- Fractures
- Folds
- Faults

**Slide 15**

*Seismic Attributes - different distributions*

Normal Distributions

- Amplitude
- Imaginary Part
- Dip Variance
- Relative Acoustic Impedance
- Instantaneous Frequency
- Envelope Time Derivative
- Envelope Second Derivative
- Curvature Maximum
- Curvature Minimum
- Curvature Most Positive
- Curvature Most Negative
- Curvature Dip Direction
- Curvature Strike Direction
- Curvature Mean
- Attenuation
- Similarity Variance

Lognormal or Exponential Distributions

- Average Energy
- Chaotic Reflection
- Curvature Curvedness
- Dip of Maximum Similarity
- Dominant Frequency
- Envelope Modulated Phase
- Instantaneous Dip
- Instantaneous Lateral Continuity
- Instantaneous Q
- Parallel Bedding Indicator
- Smoothed Dip of Maximum Similarity
- Sweetness
- Trace Envelope

Odd Distributions

- Curvature Gaussian
- Shale Indicator
- Zones of Unconformity
- Dip Azimuth
- Instantaneous Frequency Envelope Weighted?

Uniform Distributions

- Instantaneous Phase
- Normalized Amplitude

Left Skewed Distributions

- Similarity
- Smoothed Similarity

Bi or Tri-Modal Distributions

- Acceleration of Phase
- Event Continuity
- Paraphase
- Thin Bed Indicator

**Slide 16 - Translating from the 2D World to the 1D World in color mapping**

**Slide 17-23 - How Paradise "translates" color: Fasken Ranch, 10x10 topology, time slice at 1.44 seconds 2D example**

**Slide 24 - Case Histories**

**Slide 25**

*Unconventional Resource Plays*

The essential elements of unconventional resource plays encompass the following categories:

- Reservoir Geology: thickness, lateral extent, stratigraphy, mineralogy, porosity, permeability
- Geochemistry: TOC, maturity (Ro-heat), kerogen% (richness)
- Geomechanics: acoustic impedance inversion, Young’s modulus, Poisson’s ratio (Vp/Vs), pressures
- Faults, Fractures, and Stress Regimes: coherency (similarity), curvature, fault volumes, velocity anisotropy (azimuthaldistribution), stress maps

**Slide 26**

*Fasken Ranch "Wolfberry" Analysis*

Client gave us the following data covering about 60 square miles

- Bandwidth Extension –14-100 Hz amplitude volume
- Brittleness
- Bulk_Modulus
- Density
- Lames_Modulus
- Poisson’s Ratio
- Rigidity
- Inversion
- Shear_Impedance
- VP
- VpVs
- Vs
- Young’s Modulus
- Culture data, Production data, logs on 14 wells, four horizons

**Slide 27**

*Zone of interest in BE Amplitude volume with RAI overlay*

**Slide 28-33**

Neural Analysis run 120 ms above Lower Wolfcamp horizon to 30 ms below, using 100 neurons (10x10 topology)

Wiggle overlay is Relative Acoustic Impedance showing relative “hard” and “soft” spots. Classification is showing

lithology changes.

**Slide 34**

*SOM Details for Buda*

Attributes from Client:

- Brittleness Coefficient
- Final Density
- LambdaRho
- MuRho
- P_Impedance
- S_Impedance
- Poisson’s Ratio
- Poisson’s Brittleness
- Young’s Brittleness

- Average Energy
- Dip Variance
- Final Density
- PSTM (Amplitude volume)
- Poissons_Brittleness
- Relative Acoustic Impedance
- S_Impedance
- Sweetness
- Instantaneous Frequency
- Envelope

Run with 8 x 8 topology, 80 Epochs, Time: 1.2 –1.6 seconds

**Slide 35**

Isolation of specific neurons which represent "Sweet Spot" in Buda

**Slide 36 - Buda Objectives**

Well with excellent shows, lost circulation, had to set liner, finally completed partial section flowing naturally at 300 BOPD, after 6 months was still producing 125 BOPD naturally

**Slide 37**

*Granularity in your data - Bone Springs - Delaware Basin*

Using a 4x4 topology (16 neurons) and a smaller time window, allows one to direct neuron patterns to focus areas and look at a more “regional view”

**Slide 38**

Using a 6x6 topology (36 neurons) and a smaller time window, shows more detail and complexity in the patterns. There is a clear distinction between the “geology/lithology” in the northeastern portion of the survey than in the western/southern portion.

**Slide 39**

Using an 8x8 topology (64 neurons) and a smaller time window, shows more detail and complexity in the patterns. There is still a clear distinction between the “geology/lithology” in the northeastern portion of the survey than in the western/southern portion, but there is another level of detail not seen in the previous two SOMs.

**Slide 40-42 - Southern Oklahoma - Lap-out play **

**Slide 43 - Conventional Type 2 AVO Yegua - thin pay**

**Slide 44 - Inline through well - Amplitude data**

**Slide 45 - Conventional PSTM Amplitude Map**

**Slide 46 - SOM Classification - Base of Yegua Pay**

**Slide 47 - SOM Classification Inline**

**Slide 48 - 2nd Successful Yegua well updip**

**Slide 49 - Volume rendering visualization of specific neurons**

**Slide 50-52 - Middle Wilcox - Wharton County**

**Slide 53**

*Summary and Conclusions*

- Principal Component analysis will sort through the attributes and show the ones which contribute most to the interpretation and analysis
- Unsupervised neural network analysis (SOM) employing specific sets of attributes can be used to reduce risk and identify solutions to problems within the seismic data.
- The more information used (wells, production, etc.) the better the solution can be tuned with targeted attribute selections.
- Color translations can be made from the 2D world to the 1D world
- Neural analysis can be done on 2D data or 3D data
- It is important to understand the functionality of the attributes one chooses for the neural analysis in order to understand the results
- Machine learning is a “step-change” in the way geoscientists handleinterpretation of their data.