AAPG
EXPLORER
EXPLORER
Collecting optimal-quality seismic data across some onshore prospects can be a challenge for numerous reasons:
Because of these factors, a source that produces good-quality data across a sandy-soil grassland may produce poor data across an area of hard rock outcrops, and a receiver dimension that cancels ground-roll noise at prospect A may fail to do so at prospect B.
Before a seismic data-acquisition effort is launched across a prospect, seismic test data need to be acquired to determine:
The effort expended in acquiring this basic planning information is commonly referred to as wave testing or noise testing.
Figure 1 – Source-receiver geometry employed when recording seismic wave-test data. The distance “X” from source point SP1 to source point SPN should equal or exceed the depth to the deepest target that needs to be imaged. Distance “L” between successive source stations is the same as the length of the stationary geophone spread. The separation “d” between successive geophones stations should be small (two to three meters), and the receiver-patch length L needs to be reasonably long (200 to 400 meters).
Seismic wave-test data are usually acquired using a “walkaway,” or moving source geometry, such as illustrated on figure 1.
When recording the data, the receiver spread remains stationary and source stations are moved away from the receiver patch at distance increments L that equal the receiver spread length. The maximum source-to-receiver offset distance X should equal or exceed the depth to the deepest target that needs to be imaged with the surface-recorded seismic data.
Each receiver station within the receiver patch should be occupied by a single geophone – not by a string of geophones that extend over an appreciable distance – which prevents the cancellation of short-wavelength energy modes that need to be analyzed.
There should be no skipped source stations along the line of profile. If a source station is inaccessible for any reason, that source point should be moved closer to the receiver spread to prevent gaps in the offset profile.
Successive source stations should then be moved forward by the same distance to preserve a uniform spacing L between the remaining source-station positions.
Figure 2 – Wave-test data recorded for the purpose of analyzing signal frequency
bandwidth, identifying coherent and ambient noise modes and quantifying
signal-to-noise character of seismic data across a prospect area. In this
example, closely spaced test-data traces have been summed to represent data
acquired with a geophone array length of 110 feet (33.5 meters). The apparent
velocities of several modes embedded in the test data are labeled in units of
feet/second.
Site selection is important when recording wave-test data because the data sample only a small portion of a prospect.
A wave-test site should be representative of the entire prospect. If surface conditions change over a prospect, more than one wave-test site should be considered.
Wave-test data acquired across one prospect area circa 1975, using a recording geometry such as described on figure 1, are illustrated on figure 2.
These data illustrate several coherent noise modes. The dominant noises are a refraction event that has an apparent velocity of 9,500 feet/second (2,900 meters/second) and a band of surface waves with apparent velocities ranging from 2,750 to 4,200 ft/s (825 to 1,280 m/s). Several reflection events can be seen between 1.5 and 2.5 s at large offset distances beyond the surface-wave noise cone.
Because wave-test data are acquired using a small trace spacing d, any arbitrary-length inline receiver array can be simulated by summing an appropriate number of adjacent traces.
Figure 3 – Wave-test data acquired with d = 20 feet (six meters) after a 12-Hz low-cut
filter is applied and a running sum of 11 adjacent traces is calculated to
simulate data that would be acquired with an inline receiver array spanning 220
feet (67 meters). Hard evidence now has been produced that defines the receiver
patch dimension (220 feet) that should be used at each receiver station to
optimize data quality. 
For example, the data on figure 3 show the effect of adding 11 consecutive traces of wave-test data acquired with a geophone spacing d = 20 feet (figure 1) to simulate data that would be acquired with a string of geophones spanning 220 feet (67 meters) at each offset station. A 12-Hz low-cut filter has been applied to these data to aid in reducing low-frequency noise.
The frequency and wavelength content of the data inside the indicated Transform window is illustrated on figure 4 and confirms that the data are dominated by high-velocity reflection signals.
Running sums can be made using different numbers of test-data traces to simulate how receiver groups spanning any desired distance affect the frequency content and signal-to-noise ratio of prospect data. With this knowledge, seismic contractors can deploy receiver groups that have dimensions that will produce optimal quality data when production data recording is done.
Similarly, different sources can be deployed at each source station SP1 to SPN (figure 1) to compare data quality produced by vibrators versus impactors or by small explosive charges deployed at shallow depth versus large explosive charges placed at deep depths.
By comparing data quality generated by each source option, a contractor will know the best source to use across a specific prospect.
The concept described here should really be called horizontal wave testing, meaning geophones are deployed horizontally across the Earth surface when acquiring the test data.
Next month, we will consider the concept of vertical wave testing, where geophones are deployed vertically in a deep well as test data are acquired.
The Geophysical Corner is a regular column in the EXPLORER, edited by Bob A. Hardage, senior research scientist at the Bureau of Economic Geology, the University of Texas at Austin.
This month’s column is the first of a two-part series dealing with seismic wave tests – horizontal wave testing.