1
Chapter 3
Seismic Velocity
Introduction
The objective of this chapter is to learn about the porosity-velocity relationship and
types of seismic velocities and how to compute them.
Seismic velocity of subsurface rocks is the most important parameter in seismic
exploration.
Seismic velocity is used in various stages of seismic data processing and
interpretation such as:
NMO correction and stacking
Interval-velocity estimation
Time-to-depth conversion
Migration
General interpretation purposes
Model of a sedimentary rock
The simplest model of a sedimentary rock consists of spheres (grains) arranged
(packed) in a specific pattern. The packing pattern controls the amount of pore space.
Porosity is the most controlling factor on seismic velocity.
Porosity () is defined as the ratio between the pore volume (Vp) and the total rock
volume (Vb):
= Vp/Vb. (1)
2
Porosity-density relation: Porosity is related to the rock (overall), matrix (solid), and
pore-fluid densities (r, m, and f) by the following volume-weighted average
equation:
r = f + (1 - m. (2)
Porosity-velocity relation: Porosity is related to the rock, matrix, and pore-fluid P-
wave velocities (r, m, and f) by the following time-average (Wyllie’s) equation:
1 (1 )
. (3)
r f
m
Figure.
Density-velocity relation: Gardner’s rule is an empirical equation that relates the
density and P-wave velocity in sedimentary rocks as follows:
r 0 . 31 r ,
1/ 4
(4)
where r is the rock (overall) density (gm/cm3) and r is the rock (overall) P-wave
velocity (m/s).
This figure shows plots of r and r as functions of .
The following statements are generally true about the overall P-wave velocity in
sedimentary rocks:
Velocity decreases with porosity.
Velocity increases from sandstone to limestone to dolomite.
In the same rock, P-wave velocity is smaller, when gas fills the pores, than when
oil or water fills them.
Velocity increases with age, depth, pressure, or cementation.
3
The following is a table of matrix densities and velocities of some rocks:
Rock m (km/s) m (gm/cm3)
Shale 3.4 2.52
Sandstone 5.5 2.65
Limestone 6.4 2.71
Dolomite 7.0 2.87
Anhydrite 6.1 2.96
Salt 4.6 2.16
The following is a table of densities and velocities of some fluids:
Fluid f (km/s) f (gm/cm3)
Water 1.5 1.00
Oil 1.3 0.80
Gas (at 1 atm, 25 ºC)* 0.33 0.0008
* Density and P-wave velocity in gas increase with pressure and temperature.
Seismic velocity types
We will study the following seismic velocity types:
Interval
Average
NMO
RMS
Stacking
4
Dix
The interval velocity (Vi) is the velocity in a single layer, which can be
determined from sonic logs or laboratory measurements on cores from the layer.
The average velocity ( V N ) to the Nth layer is defined in terms of the layers
properties as:
V T i 0i
i 1
VN N
T 0i
i 1
where N: the total number of layers, Vi: the interval velocity in the i-th layer, and
T0i = T0i – T0i-1; where T0i-1 and T0i are the zero-offset traveltimes to the top and
bottom of the i-th layer, respectively (T00 = 0).
The average velocity is the velocity that we get by dividing the depth (ZN)
over the zero-offset one-way traveltime (T0N/2) to the bottom of the Nth layer:
2Z N
VN
T0 N
The NMO velocity (VNMO) to the bottom of the Nth layer is the velocity found using
the approximate NMO correction formula (equation (5) of Chapter 2).
VNMO is found practically by searching for the velocity that will align the true T-X
curve horizontally using the approximate NMO correction formula.
This is usually done through the constant velocity stack (CVS) method during the
velocity analysis phase of the seismic data processing flow (Figure).
X
It can be found directly from TNMON(X) as V NMO
2 * T 0 N T NMO N ( X )
N
5
The root-mean-square velocity (VRMS) to the bottom of the Nth layer is defined, in
terms of layers properties, as:
V
2
i
T0 i
i 1
V RMS N
N
T0 i
i 1
It is defined, in terms of the true T-X curve, as the reciprocal of the square root of
the X2 coefficient we get by fitting a polynomial to the true T 2 - X 2
curve. That
is, fitting a polynomial of the form: T N2 C 0 N C 1 N X 2 C 2 N X 4 ... to the true
1
T
2
-X 2
curve, V RMS N .
C 1N
Note that VRMS is also the reciprocal of the square root of the slope of the tangent
1
V RMS N
dT
2 2
to the true T - X curve at X = 0: 2 2 .
N
/ dX X 0
The stacking velocity (VS) to the bottom of the Nth layer is defined as the velocity
found by fitting a hyperbola to the true T-X curve of the form:
2
X
T (X ) T
2 2
N 0N 2
VSN
Note that the stacking velocity is a special case of the RMS velocity (i.e., when
only the first 2 terms are used in the polynomial fit).
6
By fitting a hyperbola to the true nonhyperbolic T-X curve, we are lumping all the
layers above the Nth reflector into a single virtual layer and assigning this virtual
layer a velocity of VSN.
VS is determined practically by searching for the velocity that will produce the
best-fit hyperbola to the true T-X curve.
This is usually done through the velocity spectrum method during the velocity
analysis phase of the seismic data processing flow (Figure).
At small offsets (X/ZN < 1),VRMSN ≈ VSN ≈ VNMON .
Dix velocity (VN) of the Nth layer is the interval velocity calculated from the RMS
velocities to the top and bottom of the Nth layer (VRMSN-1 and VRMSN) using Dix’s
following formula:
2 2
V RMSN T 0 N V RMSN 1
T 0 N 1
VN
T0 N
Out of VRMSN, VSN, and VNMON, only VRMSN can be related directly to the interval
velocities of subsurface layers through Dix velocity formula.
However, if only small offsets are used, we can use VSN-1 and VSN or VNMON-1 and
VNMON in place of VRMSN-1 and VRMSN in Dix’s velocity formula (More details).
7
Methods of velocity determination
1. Time-distance methods: This includes:
(1) X2-T2 method:
If we assume that the time-distance curve is a hyperbola, then by plotting T2
versus X2, we get a straight line whose slope is 1/V2, where V is the stacking
velocity (VS) to the reflector.
We can also fit higher-order polynomials to the true T-X curve to get the RMS
velocity (Figure).
Exercise: Fit polynomials of increasing orders to the true T-X curves of the
second and third layers in this sheet and estimate VS or VRMS and compare
your results to the true model parameters.
This method is more suited for high-quality small (experimental or synthetic)
datasets because it requires picking of time at many offsets, which is time
consuming and inaccurate on real datasets.
(2) Best-fit methods: These are the most commonly used methods for determining
velocity in seismic exploration. They are carried out during the velocity analysis
phase of the seismic data processing flow. They include:
Velocity spectrum method: by fitting the best-fit hyperbola to the true T-X
curve and finding the corresponding stacking velocity to each reflector
(Figure).
Constant-velocity stack method: by finding the NMO velocity that produces
the best NMO-corrected and stacked section for each reflector (Figure).
8
2. Borehole methods: These include:
Check-shot survey (Shooting a well): A shot at the wellhead is fired and
receivers at specific depths in the well are activated. The time-depth data of the
direct arrivals are used to determine the velocities to each depth (Figure).
Vertical Seismic Profiling (VSP) survey: Several shots at known offsets near a
well are shot while receivers at specific depths in the well are activated. The
time-depth-offset data is used to determine velocities near the well (Figure).
Acoustic (Sonic) Logging: Vertical transit time in the borehole wall is recorded
continuously using closely-spaced receivers. The velocity is calculated from the
transit time and receiver spacing (Figure).