Stanford Rock Physics Laboratory - Tapan Mukerji
Basic Geophysical Concepts
14
Stanford Rock Physics Laboratory - Tapan Mukerji
Body wave velocities have form: velocity = modulus
density
K + (4 / 3)µ λ + 2µ
VP = =
ρ ρ P wave velocity
µ S wave velocity
VS =
ρ E wave velocity
E
VE =
ρ
where
ρ density
K bulk modulus = 1/compressibility
µ shear modulus
λ Lamé's coefficient
E Young's modulus
ν Poisson's ratio
M P-wave modulus = K + (4/3) µ
Moduli from velocities:
⎛ ⎛ 4⎞ ⎞
µ = ρV 2
K = ρ⎜ VP2 − ⎝ ⎠ VS2 ⎟
S
⎝ 3 ⎠
M = ρ VP E = ρVE
2 2
In terms of Poisson's ratio we can also write:
V P2 2 (1 − v ) VE2 (1+ v )(1−2v) VP2 −2VS2 VE2 −2VS2
2 = 2 = v= 2 2 =
VS (1 − 2v ) VP (1− v) 2(VP −VS ) 2VS2
Relating various velocities: 2
V E2 VP
4− 2 3 −4
V P2 VS VE2 VS2
= 2 =
V S2 V E2 VS VP2
3− 2 −1
VS VS2
15
Stanford Rock Physics Laboratory - Tapan Mukerji
We usually quantify Rock Physics relations in
terms of moduli and velocities, but in the field
we might look for travel time or Reflectivity
ρ1V1
ρ2V2
The reflection coefficient of a normally-incident P-
wave on a boundary is given by:
ρ2V2 −ρ1V1
R = ρ V +ρ V
2 2 1 1
where ρV is the acoustic impedance. Therefore,
anything that causes a large contrast in impedance
can cause a large reflection. Candidates include:
•Changes in lithology
•Changes in porosity
•Changes in saturation
•Diagenesis
16
Stanford Rock Physics Laboratory - Tapan Mukerji
AVO
Amplitude Variation with Offset
V P1, VS1, ρ1 φ1 Deepwater
Well Oil
#2 Sand
Reflected
θ1 S-wave
Incident Reflected
P-wave P-wave
Transmitted
P-wave
φ2 θ2
Transmitted
VP2, V S2, ρ 2 S-wave C N.4 D
Recorded CMP Gather Synthetic
In an isotropic medium, a wave that is incident on a
boundary will generally create two reflected waves (one
P and one S) and two transmitted waves. The total shear
traction acting on the boundary in medium 1 (due to the
summed effects of the incident an reflected waves) must
be equal to the total shear traction acting on the boundary in
medium 2 (due to the summed effects of the
transmitted waves). Also the displacement of a point in
medium 1 at the boundary must be equal to the displace-
ment of a point in medium 2 at the boundary.
17
Stanford Rock Physics Laboratory - Tapan Mukerji
AVO - Aki-Richards approximation:
P-wave reflectivity versus incident angle:
Intercept Gradient
⎡ 1 ∆VP VS2 ⎛ ∆ρ ∆VS ⎞ ⎤ 2
R(θ ) ≈ R0 + ⎢ −2 2⎜ +2 ⎟ ⎥ sin θ
⎢⎣ 2 VP VP ⎝ ρ VS ⎠⎥ ⎦
1 ∆VP
+
2 VP
[tan 2 θ − sin2 θ ]
1 ⎜⎛ ∆VP ∆ρ ⎞
R0 ≈ + ⎟
2 ⎝ VP ρ⎠
In principle, AVO gives us information about
Vp, Vs, and density. These are critical for
optimal Rock Physics interpretation. We’ll
see later the unique role of P- and S-wave
information for separating lithology,
pressure, and saturation.
18
Stanford Rock Physics Laboratory - Tapan Mukerji
Seismic Amplitudes
Many factors influence seismic amplitude:
• Source coupling
• Source radiation pattern
• Receiver response, coupling, and pattern
• Scattering and Intrinsic Attenuation
• Sperical divergence
• Focusing
• Anisotropy
• Statics, moveout, migration, decon, DMO
• Angle of Incidence
…
• Reflection coefficient
Source Rcvr
19
Stanford Rock Physics Laboratory - Tapan Mukerji
Intervals or Interfaces?
Crossplots or Wiggles?
Rock physics analysis is usually applied to intervals, where
we can find fairly universal relations of acoustic properties to
fluids, lithology, porosity, rock texture, etc.
Interval Vp vs. Phi
Interval Vp vs. Vs
In contrast, seismic wiggles depend on interval boundaries
and contrasts. This introduces countless variations in
geometry, wavelet, etc.
20
Stanford Rock Physics Laboratory - Tapan Mukerji
Convolutional Model
Impedance Normal Incidence
Reflectivity
vs. depth Seismic
Convolve
With
wavelet
Rock properties Derivatives of Smoothed image
in each small layer of derivative of
layer properties impedance
Normal incidence reflection seismograms can be
approximated with the convolutional model. Reflectivity
sequence is approximately the derivative of the
impedance:
1 d
R(t) ≈ ln(ρV )
2 dt
Seismic trace is “smoothed” with the wavelet:
S(t) ≈ w(t)∗ R(t)
Be careful of US vs. European polarity conventions!
21
Stanford Rock Physics Laboratory - Tapan Mukerji
Inversion
Two quantitative strategies to link interval
rock properties with seismic:
•Forward modeling
•Inversion
•We have had great success in applying
rock physics to interval properties.
•For the most part, applying RP directly to
the seismic wiggles, requires a modeling
or inversion step.
We often choose a model-based study,
calibrated to logs (when possible) to
•Diagnose formation properties
•Explore situations not seen in the wells
•Quantify signatures and sensitivities
22
Stanford Rock Physics Laboratory - Tapan Mukerji
The Rock Physics Bottleneck
At any point in the Earth, there are only 3
(possibly 4) acoustic properties: Vp, Vs,
density, (and Q).
No matter how many seismic
attributes we observe, inversions can
only give us three acoustic attributes
Others yield spatial or geometric information.
Reservoir
Seismic Acoustic Properties
Attributes
Properties
Traveltime Porosity
Vnmo Vp Saturation
Vp/Vs Vs Pressure
Ip,Is Density Lithology
Ro, G Q Pressure
AI, EI Stress
Q Temp.
anisotropy Etc.
etc
23
Stanford Rock Physics Laboratory - Tapan Mukerji
Problem of Resolution
Log-scale rock physics may be different
than seismic scale
24
Stanford Rock Physics Laboratory - Tapan Mukerji
Seismic properties (velocity, impedance,
Poisson Ratio, etc)
… depend on pore pressure and stress
Units of Stress:
1 bar = 106 dyne/cm2 = 14.50 psi
10 bar = 1 MPa = 106 N/m2
1 Pa = 1 N/m2 = 1.45 10-4 psi = 10-5 bar
1000 kPa = 10 bar = 1 MPa
Stress always has units of force/area
Mudweight to Pressure Gradient
1 psi/ft = 144 lb/ft3
= 19.24 lb/gal
= 22.5 kPa/m
1 lb/gal = 0.052 psi/ft
25
