Using 3-D Seismic Attributesin Reservoir CharacterizationUsing 3-D Seismic Attributesin Reservoir Characterization

Susan NissenGeophysical ConsultantMcLouth, KS

Susan NissenGeophysical ConsultantMcLouth, KS

Kansas Next Step 2007 Seminar: New Technology/Seismic/Seismic InKansas Next Step 2007 Seminar: New Technology/Seismic/Seismic InterpretationterpretationAugust 9, 2007August 9, 2007

Hays, KSHays, KS

OutlineOutline•• Brief overview of some reflection seismology basicsBrief overview of some reflection seismology basics

•• What are seismic attributes and what physical What are seismic attributes and what physical information can they provide?information can they provide?

•• Methods of interpreting attributes from 3Methods of interpreting attributes from 3--D seismic D seismic volumesvolumes

•• Reservoir Characterization ExamplesReservoir Characterization Examples–– Fault interpretation Fault interpretation –– PorosityPorosity–– Bed thickness estimationBed thickness estimation–– Fracture delineationFracture delineation

•• ConclusionsConclusions

The Seismic Reflection MethodThe Seismic Reflection Method

33--D seismic data volumeD seismic data volumeCMP gatherCMP gather

MoveoutMoveout, , stack, migratestack, migrate

CMP

Source

ρ1V1

ρ2V2

Receiver

Geophone (receiver)Geophone (receiver)Vibrator truck (source)Vibrator truck (source)

Figure Courtesy of Industrial Vehicles

Seismic Reflection InterpretationSeismic Reflection InterpretationUsually horizonUsually horizon--basedbased

HorizonHorizon -- the surface the surface separating two different separating two different rock layers; also, the rock layers; also, the reflection from this reflection from this surface.surface.

0.4

0.5

0.6

0.7

0.8

Trav

el ti

me

(sec

onds

)

ρ1V1

ρ2V2

ρ3V3

ρ4V4

Horizon 1Horizon 1

Dep

th

Horizon 2Horizon 2

Horizon 3Horizon 3

Seismic Applications in Petroleum ExplorationSeismic Applications in Petroleum Exploration

Structural analysisStructural analysis (1920s onward)(1920s onward)•• study of reflector geometrystudy of reflector geometry•• used to identify faults and used to identify faults and

locally high parts of formationslocally high parts of formations

deBruin et al. (2007)

Seismic sequence stratigraphySeismic sequence stratigraphy (1970s onward)(1970s onward)•• study of reflection sequences study of reflection sequences •• used to locate stratigraphic used to locate stratigraphic

traps and define the facies traps and define the facies framework of structural trapsframework of structural traps

Seismic attribute analysisSeismic attribute analysis (1970s onward)(1970s onward)•• study of seismic attributes study of seismic attributes •• provides information related to structure, provides information related to structure,

stratigraphy, and reservoir propertiesstratigraphy, and reservoir properties

What are Seismic Attributes?What are Seismic Attributes?

Any measurement Any measurement derivedderived from the from the seismic data is a seismic attribute.seismic data is a seismic attribute.

Seismic attributes typically provide Seismic attributes typically provide information relating to the information relating to the amplitudeamplitude, , shapeshape, and/or , and/or positionposition of the seismic of the seismic waveform.waveform.

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

General classes of attributesGeneral classes of attributes

11--D attributes D attributes -- operate on a single stacked operate on a single stacked seismic traceseismic trace

22--D and 3D and 3--D attributes D attributes -- calculated using calculated using information from adjacent tracesinformation from adjacent traces

Families of Seismic AttributesFamilies of Seismic AttributesTime Attributes Time Attributes (1930s)(1930s) –– related to the vertical position related to the vertical position of the waveform in the seismic section (e.g., of the waveform in the seismic section (e.g., horizon horizon timetime picks, picks, isochronsisochrons))

Envelope:Envelope: A(t) = [qA(t) = [q22(t)+r(t)+r22(t)](t)]1/21/2

Instantaneous Phase:Instantaneous Phase: θθ(t) = tan(t) = tan--11[q(t)/r(t)][q(t)/r(t)]Instantaneous Frequency:Instantaneous Frequency: ωω(t)=(t)=ddθθ(t)/(t)/dtdt

q(t)

A(t)

r(t)

θ(t)

Complex Trace Attributes Complex Trace Attributes (1970s)(1970s) –– The seismic data The seismic data is treated as an analytic trace, is treated as an analytic trace, which contains both real and which contains both real and imaginary parts.imaginary parts. Various Various amplitudeamplitude, , phasephase, and , and frequencyfrequency attributes can be attributes can be calculated. calculated.

Taner et al. (1989)

Families of Seismic AttributesFamilies of Seismic AttributesWindow Attributes Window Attributes (1980s)(1980s) –– attributes which summarize attributes which summarize information from a vertical window of data.information from a vertical window of data.

Fourier Attributes Fourier Attributes (1990s)(1990s) –– frequency frequency domain attributes obtained through domain attributes obtained through Fourier analysis (e.g., Fourier analysis (e.g., spectral spectral decompositiondecomposition))

MultiMulti--trace Attributestrace Attributes (1990s)(1990s) -- attributes calculated attributes calculated using more than one input seismic trace, which provide using more than one input seismic trace, which provide quantitative information about lateral variations in the quantitative information about lateral variations in the seismic data (e.g., seismic data (e.g., coherencecoherence, , dip/azimuthdip/azimuth, , volumetric volumetric curvaturecurvature))

Definitions of selected attributes

Definitions of selected attributes

Envelope:Envelope: A(t) = [qA(t) = [q22(t)+r(t)+r22(t)](t)]1/21/2

Instantaneous Phase:Instantaneous Phase: θθ(t) = tan(t) = tan--11[q(t)/r(t)][q(t)/r(t)]Instantaneous Frequency:Instantaneous Frequency: ωω(t)=(t)=ddθθ(t)/dt(t)/dt

Instantaneous AttributesInstantaneous Attributes

q(t)

A(t)

r(t)

θ(t)

Taner et al. (1989)

Spectral DecompositionSpectral Decomposition

SingleSingletracetrace

All tracesAll traces

FF11

FF22

black = low amplitudeblack = low amplitudewhite = high amplitudewhite = high amplitude

analysisanalysiswindowwindow

3D Seismic Volume3D Seismic Volume

}Uses the Fourier transform to Uses the Fourier transform to calculate the amplitude calculate the amplitude spectrum of a short time spectrum of a short time window covering the zone of window covering the zone of interestinterest

The amplitude spectrum is The amplitude spectrum is tuned by the geologic units tuned by the geologic units within the analysis window, within the analysis window, so that units with different so that units with different rock properties and/or rock properties and/or thickness will exhibit different thickness will exhibit different amplitude responses.amplitude responses.

Seismic CoherenceSeismic Coherence

black = low coherenceblack = low coherencewhite = high coherencewhite = high coherence

fault = low coherencefault = low coherence

fault = low coherencefault = low coherence

3D Seismic Volume3D Seismic Volume

For each point in a 3D seismic volume, compare the waveformFor each point in a 3D seismic volume, compare the waveform

over a short vertical windowover a short vertical windowof adjacent traces (e.g., red trace compared to blue traces)of adjacent traces (e.g., red trace compared to blue traces)

Coherence CubeCoherence Cube

A measure of the traceA measure of the trace--toto--trace similarity of the seismic trace similarity of the seismic waveform within a small analysis windowwaveform within a small analysis window

Sigismondi and Soldo, 2003

May be computed at any azimuth about a point

Generally computed normal to tangent plane

Principal Curvatures (kmaxand kmin) can be combined to define other curvature attributes

Volumetric CurvatureVolumetric CurvatureCurvature describes how bent a surface is at a Curvature describes how bent a surface is at a particular point and is closely related to the second particular point and is closely related to the second derivative of the curve defining the surface.derivative of the curve defining the surface.

Volumetric curvature is computed for every point Volumetric curvature is computed for every point within a 3within a 3--D seismic volume.D seismic volume.

PositiveCurvature

ZeroCurvature

ZeroCurvature

DippingPlane

Anticline

Syncline

NegativeCurvature

Flat

Z

X

After Roberts, 2001

PositiveCurvature

ZeroCurvature

ZeroCurvature

DippingPlane

Anticline

Syncline

NegativeCurvature

Flat

Z

X

PositiveCurvature

ZeroCurvature

ZeroCurvature

DippingPlane

Anticline

Syncline

NegativeCurvature

Flat

Z

X

Z

X

After Roberts, 2001Curvature (k)=1/R

R

2-D 3-D

What physical information is provided by seismic attributes? What physical information is

provided by seismic attributes? EnvelopeEnvelope-- presence of gas (bright spots), presence of gas (bright spots), thinthin--bed tuning effects, lithology changesbed tuning effects, lithology changesPhase Phase –– lateral continuity of reflectors, lateral continuity of reflectors, bedding configurationsbedding configurationsFrequency Frequency –– bed thickness, presence of bed thickness, presence of hydrocarbons, fracture zoneshydrocarbons, fracture zonesSpectral Decomposition Spectral Decomposition –– bed thicknessbed thicknessCoherence, Volumetric Curvature Coherence, Volumetric Curvature –– faults, faults, fractures, lateral stratigraphic fractures, lateral stratigraphic discontinuitiesdiscontinuities

Methods of interpreting attributes from 3-D seismic volumes

Methods of interpreting attributes from 3-D seismic volumes

Identify spatial patterns/trends in attribute dataIdentify spatial patterns/trends in attribute data–– CrossCross--sectional view sectional view –– Map view (attributes extracted along horizon or from Map view (attributes extracted along horizon or from

zone of interest) zone of interest) –– 3D visualization3D visualization

Tie attributes to well control using statistical Tie attributes to well control using statistical methods (e.g., methods (e.g., crossplotscrossplots))Automatically analyze multiple attributes (with or Automatically analyze multiple attributes (with or without well control)without well control)–– GeostatisticsGeostatistics–– Principal component analysisPrincipal component analysis–– Cluster analysisCluster analysis–– Texture analysisTexture analysis

Reservoir CharacterizationExamples

Reservoir CharacterizationExamples

Fault Interpretation – Offshore TrinidadFault Interpretation – Offshore TrinidadSeismic Time SliceSeismic Time Slice Coherence SliceCoherence Slice

Complex Complex faulting faulting difficult to difficult to detect on detect on seismicseismic

Coherence Coherence shows shows lateral lateral continuity continuity of faultsof faults

Gersztenkorn et al., 1999

Limits of Porous Reservoir Limits of Porous Reservoir ----

Mississippian Dolomite ReservoirMississippian Dolomite ReservoirJudicaJudica FieldField

Ness and Gove Counties, KSNess and Gove Counties, KS

GR

Nt Phi Guard Res

20% O/W -1938

DO

LOM

ITE

LS

LS

Base Warsaw LS

OsagianSeries

MeramecianSeries

Spergen

Warsaw

Mis

siss

ippi

anSystem

Penn

.

?

Judicapay zone

Top Miss

After Dubois et al., 2003

Judica Field Stratigraphy

Judica 3-D seismic survey

Bhattacharya et al., 2004

ADry holes on structural Dry holes on structural

high due to low porosity high due to low porosity within reservoir intervalwithin reservoir interval

Base Warsaw LSTop Miss

0.80 s

0.85 s

A'A A'

Top Mississippian structure

Base Warsaw LS amplitude map

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

-18000 -16000 -14000 -12000 -10000 -8000 -6000 -4000

seismic amplitude - Base Warsaw LS horizon

phi-h

(por

osity

-ft)

Bhattacharya et al., 2004

Approximate porosityof reservoir interval: 5% 25%

Synthetic seismic section

Velocity model

Modeled variation in amplitude of Base Warsaw LS horizon due to increase in porosity of reservoir zone

-0.3

-0.25

-0.2

-0.15

Am

plitu

deB

ase

War

saw

LS

Top Mississippian Structure Map Base Warsaw LS Amplitude Map

Reservoir compartment Reservoir compartment mapped from 3mapped from 3--D seismic D seismic structure and amplitudesstructure and amplitudes

Bhattacharya et al., 2004

Judica 3-D attribute analysis results

Judica 3-D attribute analysis results

Seismic amplitude of the base of Warsaw LS Seismic amplitude of the base of Warsaw LS correlates with porositycorrelates with porosity--thickness of the thickness of the JudicaJudica pay zone, providing a method for pay zone, providing a method for discriminating between dry and productive discriminating between dry and productive wellswells

A combination of seismic structure and A combination of seismic structure and amplitude analysis allows us to better amplitude analysis allows us to better delineate reservoir compartment boundariesdelineate reservoir compartment boundaries

"Thin bed" thickness "Thin bed" thickness estimationestimation

----Upper CretaceousUpper Cretaceous “D” Sand“D” Sand

Sooner Unit, ColoradoSooner Unit, Colorado

“D” Sand Reservoir“D” Sand Reservoir

Upper CretaceousUpper CretaceousIncised valley fill Incised valley fill Fluvial and estuarine sedimentsFluvial and estuarine sediments

Cannon, 1998

"D" sand thickness from wells"D" sand thickness from wells

ObjectiveObjectiveDetermine “D” sand thickness between well control Determine “D” sand thickness between well control points.points.

10 ms10 ms

Sooner 3-D seismic survey“D” sand isochron map

Sooner 3-D seismic survey“D” sand isochron map

7 ms7 ms

Unfortunately….Unfortunately….Over most of the 3Over most of the 3--D survey D survey area, the “D” sand is below area, the “D” sand is below seismic resolution (a “thin seismic resolution (a “thin bed”).bed”).Below seismic resolution, Below seismic resolution, reflections from the top and reflections from the top and bottom of the sand maintain a bottom of the sand maintain a constant temporal separation, constant temporal separation, which is unrelated to the true which is unrelated to the true sand thickness. Amplitude, sand thickness. Amplitude, however, decreases with however, decreases with decreasing bed thickness.decreasing bed thickness.

Therefore….Therefore….EnvelopeEnvelope and and spectral spectral decompositiondecomposition, both related to , both related to amplitude, are likely to be amplitude, are likely to be better potential predictors of better potential predictors of “D” sand thickness“D” sand thickness

Two-way true thickness (ms)

Two-

way

app

aren

t thi

ckne

ss

AMPL

ITUD

E

THICKNESS

b/2TR

Max

imum

abs

olut

e am

plitu

deof

com

posi

te w

avel

et

2.0

1.0

0b/2

For our model wavelet (Ormsby 12/16-80/100):b/2 = tuning thickness = 7.9 ms (~ 53 ft)T = temporal resolution = 7.2 ms (~49 ft)R

25201510500

25

5

10

15

20

Maximum thickness of “D” Sand

1200012000

00

EnvelopeEnvelope-- extracted along top “D” sand horizonextracted along top “D” sand horizon

““D” sand thickness D” sand thickness contours from wellscontours from wells

Discrete Fourier component thin bed tuning analysis

Discrete Fourier component thin bed tuning analysis

After Partyka, 2001

ampl

it ude

10 Hz

20 Hz

30 Hz

40 Hz

temporal thickness (ms)0 5 10 15 20 25

Maximum thickness of "D" sand

envelope

Spectral DecompositionSpectral Decomposition-- 50 ms window centered on “D” sand50 ms window centered on “D” sand

60 Hz60 Hz50 Hz50 Hz 70 Hz70 Hz

20 Hz20 Hz 30 Hz30 Hz 40 Hz40 Hz

Am

plitu

deA

mpl

itude

00

0.80.8

0.80.8

00

Spectral Decomposition – 29 HzSpectral Decomposition – 29 Hz-- 50 ms window centered on “D” sand50 ms window centered on “D” sand

““D” sand thickness D” sand thickness contours from wellscontours from wells

Crossplots of attribute versus “D” sand thicknessCrossplots of attribute versus “D” sand thickness

Spectral Decomposition – 29 HzSpectral Decomposition – 29 Hz

IsochronIsochron

EnvelopeEnvelope

Fracture DelineationFracture Delineation----

Mississippian ReservoirMississippian ReservoirDickman FieldDickman Field

Ness County, KansasNess County, Kansas

Dickman Mississippian ReservoirDickman Mississippian ReservoirSubjacent to regional preSubjacent to regional pre--Pennsylvanian Pennsylvanian unconformity and karst unconformity and karst surface surface Composed of multiComposed of multi--layered layered shallow shelf carbonatesshallow shelf carbonatesProduction strongly Production strongly influenced by solutioninfluenced by solution--enhanced natural fracturesenhanced natural fracturesSupported by strong Supported by strong bottom water drivebottom water driveHigh waterHigh water--cut production cut production (>94%)(>94%)Seismic depth map of top Mississippian Seismic depth map of top Mississippian

(pre(pre--Penn unconformity surface)Penn unconformity surface)

0.5 mileC.I. = 10 ft

Shale-filled fractures intersected by horizontal well Ness County, KS

Shale-filled fractures intersected by horizontal well Ness County, KS

Karst-controlled10-100 ft intervalProvide a barrier to fluid flow

KarstKarst--controlledcontrolled1010--100 ft interval100 ft intervalProvide a barrier to Provide a barrier to fluid flowfluid flow

Carr et al., 2000

CoherenceCoherence

Seismic Attributes for Delineating Faults and Fractures

Seismic Attributes for Delineating Faults and Fractures

Volumetric CurvatureVolumetric Curvature

Calculated directly Calculated directly from seismic volumefrom seismic volumeExtracted along Extracted along interpreted horizoninterpreted horizon

Horizon CurvatureHorizon Curvature

Calculated from Calculated from interpreted horizoninterpreted horizon

0.5 mile

Volumetric Curvature – Gilmore City HorizonVolumetric Curvature – Gilmore City HorizonFrequency-Azimuth Rose Diagram

Length-Azimuth Rose Diagram

fault

NENE--trending lineamenttrending lineament

Interpreted shale- and debris-filled solution-enlarged fracture coincides

with NE-trending curvature lineament

Interpreted shale- and debris-filled solution-enlarged fracture coincides

with NE-trending curvature lineament

Thickness of karst zone in well versus distance to nearest NW and NE lineaments

Thickness of karst zone in well versus distance to nearest NW and NE lineaments

Thicker karst zonecloser to lineamentsNo relationship

0.5 mile

Oil production versus distance to nearest NW and NE lineaments

Oil production versus distance to nearest NW and NE lineaments

Increased oil production farther from lineaments

No relationship

A B

C

Water production versus distance to nearest NW and NE lineaments

Water production versus distance to nearest NW and NE lineaments

0.5 mile

No relationship

Increased water productionnear lineaments

A B

C

Dickman 3-D attribute analysis results

Dickman 3-D attribute analysis results

NENE--trending curvature lineaments appear to trending curvature lineaments appear to be barriers to fluid flow, and may represent be barriers to fluid flow, and may represent shaleshale--filled fractures.filled fractures.NWNW--trending curvature lineaments appear to trending curvature lineaments appear to represent open fractures, which serve as represent open fractures, which serve as conduits into the underlying aquifer.conduits into the underlying aquifer.Understanding the orientations of open and Understanding the orientations of open and filled fractures is an important prefilled fractures is an important pre--requisite requisite for effective reservoir management.for effective reservoir management.

General conclusions about attributesGeneral conclusions about attributesAttributes reveal information which is not readily Attributes reveal information which is not readily apparent in the raw seismic dataapparent in the raw seismic dataDozens of seismic attributes can be calculated, Dozens of seismic attributes can be calculated, some of which are more useful than otherssome of which are more useful than othersAttributes may be interpreted singly or using Attributes may be interpreted singly or using multimulti--attribute analysis toolsattribute analysis toolsDifferent attributes reflect different physical Different attributes reflect different physical properties of the underlying rock systemproperties of the underlying rock systemAttributes can aid in improving our Attributes can aid in improving our understanding of the reservoirunderstanding of the reservoirThe specific attributes to use in a reservoir The specific attributes to use in a reservoir characterization study will vary, depending on the characterization study will vary, depending on the type of reservoir and the problem being type of reservoir and the problem being addressedaddressed

AcknowledgmentsAcknowledgmentsTim Carr, Marty Dubois, and Tim Carr, Marty Dubois, and SaibalSaibal Bhattacharya, Bhattacharya, Kansas Geological SurveyKansas Geological SurveyKurt Kurt MarfurtMarfurt and Chuck Blumentritt, and Chuck Blumentritt, University of University of HoustonHoustonMull Drilling Company, Inc.Mull Drilling Company, Inc.Grand Mesa Operating Company Grand Mesa Operating Company Seismic MicroSeismic Micro--Technology, Inc. Technology, Inc. IHS, Inc.IHS, Inc.U. S. Department of EnergyU. S. Department of Energy