30

4.6 Common Mid-Point Surveying

4.6.1 Common Mid-Point Gathers

There are two disadvantages to using only a shot gather for analysis:

(1) As we saw in Section 4.1, reflections tend to have a low amplitude (generally less
than 20% of the incident wave amplitude). This means that noise in the seismic data can
obscure reflections.

(2) Each reflection occurs at a different point on the interface. The analysis of shot
gathers assumes uniform horizontal layers. If there are significant lateral variations in
structure, this assumption is no longer valid and the analysis can result in errors.

These problems can be overcome by using shots from a number of different points and
multiple detectors. The source and detectors are moved in between shots.

From the complete data set, a subset of traces are chosen that have a common reflection
point. The reflection point is taken to be halfway in between the shot and the detector.
• this is valid for areas with horizontal layering
• if the reflection occurs at a dipping interface, this is an approximation, but does
not introduce large errors (see Kearey et al., fig 4.9).
 31

The collection of traces with the same reflection point is called a common mid-point
gather (CMP gather) or common depth point (CDP) gather.

As in shot gathers, the traces are arranged in order of distance from the shot.
• travel time curves for reflections from an interface will plot as parabolas
- the travel times can be analyzed in the same manner as a shot gather to
determine the velocity structure in the vicinity of the CMP

CMP Gather
0

0.1

0.2

0.3

0.4

Travel time (s) 0.5

0.6

0.7

0.8

0.9

1
-200 -100 0 100 200
Distance from shot (m)

The traces in CMP gathers have sampled the same point on the reflecting interface. They
can then be stacked (added together). Seismic traces contain both coherent energy (e.g.,
reflections) and random noise. When traces are added together, the coherent signals will
increase in amplitude, and the random noise will cancel. This increases the signal to
noise ratio (SNR), enhancing weak reflections.

Note that prior to stacking it is necessary to apply a time correction to account for normal
move-out.

Seismic Survey Procedures

The fold is the number of traces in a survey that share a common mid-point – this will be
the number of traces that are stacked together. If the sources and receivers are moved in
such a way that no traces share a common mid-point, the survey coverage is single-fold.
(A shot gather would be called single-fold.)
 32

To increase the fold of the data, seismic surveys follow this procedure:
• a seismic source is located at the front of the array of detectors
• the detector spacing is Δx, and there is a space of Δx is between the first detector
and source
• a shot is fired and the resulting seismic waves are recorded by the detectors
• the whole array, including the source, is moved forward a distance of nΔx, where
n is an integer
- on land, this involves moving the last geophone (or more) to the front of
the line and drilling a new hole for the source
- in a marine survey, this corresponds to the air gun and hydrophone
streamer moving forward with the ship
• if there are N geophones and the array moves forward a distance of nΔx between
shots, it can be shown that the number of rays that share the same common mid-
point is N/2n. This is the fold of the data. Fold is sometimes reported as a
percentage of coverage (100×N/2n %)

In the example below, there are 6 geophones and the array moves forward a distance of
Δx between shots. This gives 3-fold CMP coverage or 300% coverage.
• in most surveys, the fold is 24-60, but may be over 1000 for exceptional cases.
 33

4.6.2 Time Corrections – Statics

The goal is to stack the traces in the CMP gather, in order to increase the amplitude of
reflections. For this, the reflections must occur at the same time on all traces, so that they
will stack coherently. Thus, prior to stacking, it is necessary to apply two types of time
corrections. The corrections shift each trace vertically (in time).

The first correction is a static correction. Until now, we have assumed that the seismic
data has been collected on a flat surface above a layer with uniform velocity. As we
know, the Earth has significant topography and the near-surface geology can be highly
variable, due to weathering and other effects. Both factors will cause the travel time of
seismic waves to vary between traces. These delays are called statics because they cause
a shift in the arrival time of all the arrivals in a seismic trace.

Elevation statics: if a geophone is at a higher elevation than a reference level (datum),

the travel time will be slightly longer relative to a geophone at the reference level – all
arrivals will be delayed. Conversely, a geophone at a lower elevation will have slightly
earlier arrival times. The reference level is usually taken to be sea level

Question: What is the effect of a 20 m change in elevation on the arrival times of

reflections? Assume a velocity of 3000 m/s and no weathering layer.

Weathering statics: the upper few metres to 10’s of metres of the subsurface is
generally a zone of low seismic velocity due to fractures and joints associated with
weathering. Even though this layer may be thin, the velocities are so low that the seismic
waves experience a noticeable time delay.
CMP Gather with Statics
Question: The bedrock has a velocity of 0

3000 m/s and there is a 5 m thick weathered

layer at the surface with a velocity of 1000 0.1

m/s. What is the effect on the arrival time of

the seismic waves? 0.2

0.3

0.4
Travel time (s)

0.5
The plot on the right shows a CMP gather
that includes the effects of variations in 0.6
elevation and weathering between traces.
Note the vertical offset between adjacent 0.7

traces. All arrivals on each trace experience

the same offset. The maximum shift is ±10 0.8

ms.
0.9

1
-500 0 500
Distance from shot (m)
 34

Removing statics (static correction)

The effects of elevation and weathering are corrected by applying a time shift to the
entire seismic trace at each geophone.

• Elevation statics are corrected by making an accurate measurement of the

elevation. Using the velocity of the material below the weathered layer, the traces
are corrected for the time taken to travel the vertical distance between the
shot/detector elevation and the reference level.

• Weathering statics are corrected by making measurements of the thickness and

velocity structure of the weathered layer at each shot/detector location.
Techniques include refraction static analysis and uphole surveys (see Kearey et
al., pg 58-59).

• Often a residual static analysis is performed to correct for any remaining static
shifts between traces – compares traces and aligns them to produce the most
continuous seismic event. (see Kearey et al., Fig 4.16)

Marine surveys are much simpler – there is no weathered layer and the hydrophones and
source are in water, where the velocity is well-known. Marine traces are usually
corrected to sea level, using the depth of the hydrophones and airguns.

The following figure is from Kearey et al. (2002). It illustrates the effects of statics on
the arrival times of reflections and how the static corrections result in the alignment of
the reflections.
 35

4.6.3 Time Corrections – Velocity Analysis

The second time correction accounts for normal moveout of the reflections. Each trace is
shifted in time to become the equivalent of a zero-offset trace.

Consider a CMP gather for a single horizontal reflector. A velocity analysis is carried out
to determine what velocity should be used for the NMO correction. This is a trail and
error procedure:
• choose a velocity. This is called the stacking velocity (Vst)
• calculate the normal moveout as a function of offset (Section 4.4)
• subtract the normal moveout from each trace
• examine the results: is the reflection flat? after stacking, is the reflection a sharp,
high amplitude peak?
• vary the velocity to improve the results
Original CMP Gather NMO Corrected CMP Gather Stacked CMP
0 0 0

NMO correction with a

0.1 0.1 0.1
velocity of 2100 m/s:
0.2 0.2 0.2
After correction, the
reflection is still curved 0.3 0.3 0.3
Travel time (s)

downward. This means

0.4 0.4 0.4
that the NMO has not been
completely corrected. 0.5 0.5 0.5

Also, on the stacked trace, 0.6 0.6 0.6

the reflection shows up as a 0.7 0.7 0.7

broad peak.
0.8 0.8 0.8
-200 0 200 -200 0 200 0 0.5 1
Distance from shot (m) Distance from shot (m) Amplitude

Original CMP Gather NMO Corrected CMP Gather Stacked CMP

NMO correction with a 0 0 0

velocity of 1500 m/s: 0.1 0.1 0.1

Now the correction results 0.2 0.2 0.2

in an upward curvature of
the reflection. The NMO 0.3 0.3 0.3
Travel time (s)

has been over-corrected. 0.4 0.4 0.4

Again, the reflection is a 0.5 0.5 0.5

broad peak on the stacked
trace. 0.6 0.6 0.6

0.7 0.7 0.7

0.8 0.8 0.8

-200 0 200 -200 0 200 0 0.5 1
Distance from shot (m) Distance from shot (m) Amplitude
 36

Original CMP Gather NMO Corrected CMP Gather Stacked CMP

NMO correction with a 0 0 0

velocity of 1800 m/s:

0.1 0.1 0.1

The reflection is 0.2 0.2 0.2

relatively flat on the CMP
gather and is a sharp peak 0.3 0.3 0.3

Travel time (s)

on the stacked trace.
0.4 0.4 0.4

0.5 0.5 0.5

0.6 0.6 0.6

0.7 0.7 0.7

0.8 0.8 0.8

-200 0 200 -200 0 200 0 0.5 1
Distance from shot (m) Distance from shot (m) Amplitude

The same procedure can be used for multiple reflections. In this case, a different stacking
velocity is used to correct the NMO of each reflector. Because the correction varies as a
function of time along the trace, the NMO correction is also called a dynamic
correction.

Original CMP Gather NMO Corrected CMP Gather Stacked CMP

0 0 0
NMO Corrections:
0.00-0.45 s: Vst=1800 m/s 0.1 0.1 0.1

0.45-0.70 s: Vst=2500 m/s 0.2 0.2 0.2

0.70-1.00 s: Vst=3400 m/s 0.3 0.3 0.3
Travel time (s)

0.4 0.4 0.4

The stacking velocity of 0.5 0.5 0.5

each reflector is close to 0.6 0.6 0.6
the RMS velocity of that
0.7 0.7 0.7
reflector and the Dix
equation can be used to 0.8 0.8 0.8

determine actual 0.9 0.9 0.9

velocities. 1 1 1
-200 0 200 -200 0 200 0 0.5 1
Distance from shot (m) Distance from shot (m) Amplitude

Usually a computer program is used for the analysis. Traces are divided into small time
windows and the NMO correction is calculated for a range of velocities. The results are
plotted in a velocity spectrum, which shows the semblance (a measure of the alignment
of the NMO-corrected reflections) as a function of velocity and time. From the plot, the
appropriate velocity for each time period can be chosen. (See Kearey et al., Figure 4.18.)
 37

4.6.4 CMP Stacking

After correcting for statics and NMO, the reflections should be aligned across the CMP
gather. The traces are then stacked to produce a single trace.

As seen in the figures above, the direct wave and ground roll are not aligned on the CMP
traces. Therefore, when the traces are stacked, these waves are nearly removed.

For noisy data, stacking will increase the signal to noise ratio, as random noise will not
stack coherently.

Original CMP Gather NMO Corrected CMP Gather Stacked CMP

0 0 0

0.1 0.1 0.1

0.2 0.2 0.2

0.3 0.3 0.3

Travel time (s)

0.4 0.4 0.4

0.5 0.5 0.5

0.6 0.6 0.6

0.7 0.7 0.7

0.8 0.8 0.8

0.9 0.9 0.9

1 1 1
-200 0 200 -200 0 200 0 0.5 1
Distance from shot (m) Distance from shot (m) Amplitude

For each CMP gather, the same procedure is followed to obtain a single stacked zero-
offset trace. Each stacked CMP trace is plotted at the location of the common mid-point.
This results in a seismic section for the study area.
 38

4.6.5 Resolution in seismic surveys

When designing seismic surveys and interpreting seismic reflection data, it is important
to recognize the resolution of the method. This will govern the size of the smallest
features that can be resolved in the data.

Vertical resolution
The vertical resolution determines the minimum spacing of reflecting interfaces that can
be observed in seismic reflection data. Remember that the seismic source generates a
seismic pulse with a given length. This means that reflections on a seismic trace must be
separated by a minimum length of time in order to be clearly identified.

In the following examples, assume that there is a layer with a velocity of 2000 m/s. A
sinusoidal seismic pulse is generated with a wavelength of 200 m. (frequency = 10 Hz)

What is the smallest layer thickness that can be observed by this seismic wave?

To answer this, examine reflections for different layer thicknesses:

Velocity profile Reflection from top Reflection from bottom Top + Bottom
0 0 0 0

200 m thick layer 50 0.05 0.05 0.05

100 0.1 0.1 0.1

When the layer is very

thick, it is possible to
150 0.15 0.15 0.15

observe two distinct 200 0.2 0.2 0.2

reflections from the top

Depth (m)

Time (s)

250 0.25 0.25 0.25

and the bottom of the 300 0.3 0.3 0.3

layer.
350 0.35 0.35 0.35

400 0.4 0.4 0.4

450 0.45 0.45 0.45

500 0.5 0.5 0.5

1500 2000 2500 -0.05 0 0.05 -0.04 -0.02 0 0.02 0.04 -0.04 -0.02 0 0.02 0.04
Velocity (m/s) Amplitude Amplitude Amplitude

Velocity profile Reflection from top Reflection from bottom Top + Bottom
0 0 0 0

20 m thick layer
50 0.05 0.05 0.05

100 0.1 0.1 0.1

When the layer is very 150 0.15 0.15 0.15

thin, the two reflections

overlap and they can not
200 0.2 0.2 0.2
Depth (m )

be distinguished.
Tim e (s )

250 0.25 0.25 0.25

300 0.3 0.3 0.3

350 0.35 0.35 0.35

400 0.4 0.4 0.4

450 0.45 0.45 0.45

500 0.5 0.5 0.5

1500 2000 2500 -0.05 0 0.05 -0.04 -0.02 0 0.02 0.04 -0.1 -0.05 0 0.05 0.1
Velocity (m/s) Amplitude Amplitude Amplitude
 39

Velocity profile Reflection from top Reflection from bottom Top + Bottom
0 0 0 0

50 0.05 0.05 0.05

50 m thick layer 100 0.1 0.1 0.1

150 0.15 0.15 0.15

At this point, the crest of

the second wave 200 0.2 0.2 0.2

coincides with the trough

Depth (m)

Time (s)
250 0.25 0.25 0.25

of the first wave.

300 0.3 0.3 0.3

350 0.35 0.35 0.35

400 0.4 0.4 0.4

450 0.45 0.45 0.45

500 0.5 0.5 0.5

1500 2000 2500 -0.05 0 0.05 -0.04 -0.02 0 0.02 0.04 -0.04 -0.02 0 0.02 0.04
Velocity (m/s) Amplitude Amplitude Amplitude

Velocity profile Reflection from top Reflection from bottom Top + Bottom
0 0 0 0

50 0.05 0.05 0.05

55 m thick layer
100 0.1 0.1 0.1

Now it is possible to see 150 0.15 0.15 0.15

two reflections.
200 0.2 0.2 0.2
Depth (m)

Time (s)

250 0.25 0.25 0.25

300 0.3 0.3 0.3

350 0.35 0.35 0.35

400 0.4 0.4 0.4

450 0.45 0.45 0.45

500 0.5 0.5 0.5

1500 2000 2500 -0.05 0 0.05 -0.04 -0.02 0 0.02 0.04 -0.04 -0.02 0 0.02 0.04
Velocity (m/s) Amplitude Amplitude Amplitude

The critical point where two reflections can be identified occurs when the waves are
separated by a distance of λ/2.

For a layer with a thickness of t, the reflection from the bottom boundary travels a
distance of 2t more than the reflection from the top.

Therefore the maximum resolution is: 2t = λ/2 Æ t = λ/4

Only layers that are thicker than λ/4 will be observable in seismic reflection data.
 40

Horizontal resolution
Horizontal resolution is determined by: 1) properties of the seismic waves, and 2) spacing
of the detectors. The factor that gives the larger value will determine the resolution of the
data.

Fresnel zones
In much of our analysis, we consider seismic reflections to involve only one ray that
samples a single point on the interface. Consider a source and detector located above a
horizontal interface. Ray theory predicts that the detector will record the ray that has
travelled vertically from the source to the interface and back up (normal incidence).

However, it is important to remember that the seismic source generates an infinite

number of rays that travel outward from the source as a spherical wavefront. Upon
striking a reflecting interface, the wavefront is decomposed into point sources that will
each generate spherical waves (Huyghen’s principle). This means that there could be
rays recorded at the detector that come from reflection angles that are not expected based
on ray theory.

These waves will be recorded by the detector, but because they have travelled a greater
distance than the normally-incident wave, they will arrive later. Following the same
ideas used above, the recorded waves will interfere constructively if the total distance
that they have travelled is within λ/2 of the normal (zero-offset) reflection - the waves
will appear as one reflection.

The part of the interface from which this energy is reflected is called the Fresnel zone. It
can be shown is a circular disc with a width, w = 2zλ .
 41

This width represents the horizontal length of the smallest feature that will be observable
in the data recorded at the surface. This equation indicates that the horizontal resolution
will decrease for deeper reflectors (larger z).

From the above discussion, short wavelengths will yield the highest vertical and
horizontal resolution. However, attenuation causes the preferential reduction in
amplitude of high frequency (short wavelength) energy. This means that only the
shallowest regions can be studied with short wavelengths. As waves travel to greater
depths, the dominant wavelength of the seismic waves will increase, and therefore the
vertical and horizontal resolution will decrease.

Detector spacing
As shown in the figure below, for a horizontal reflector, the reflections are spaced at a
distance that is half of the detector spacing. In designing a seismic survey, it is important
to make sure that this spacing is less than the Fresnel zone, so that the survey layout
does not limit the resolution of the data. A rule of thumb is that the detector spacing
should be no more than one quarter of the Fresnel zone width.

Questions
Seismic waves with a dominant frequency of 50 Hz travel through sediments with a
velocity of 3000 m/s.
• What is the smallest layer thickness that will be detected?
• If the deepest reflector is at a depth of 2000 m, what is the size of the smallest
horizontal feature that will be detected?
• What is the largest geophone spacing that should be used in the survey?

Now think about statics:

• What is the maximum static shift that will permit coherent stacking?
• Assume that all the static shifts are due to elevation variations. How accurately
do you need to know the elevation?