The Seismic reflection method

What is a seismic section?

The seismic reflection method works by bouncing sound waves off boundaries between different types of rock (Figure 1). The reflections recorded are plotted as dark lines on a seismic section. A seismic section resembles a geological cross-section, but it still needs to be interpreted. One major difference between a geological cross-section and a seismic section is that the vertical axis is in time, rather than depth. In the earth's crust, seismic waves travel typically at about 6000 m/s so that 1 second of two-way travel time corresponds to about 3 km of depth. All the seismic sections presented in this atlas are plotted at 1:1 (no vertical exaggeration) assuming an average crustal velocity of 6000 m/s. Another difference is that the reflections are plotted halfway between the source and the receiver. These are referred to as unmigrated data. The process that moves the reflections in their correct spatial position is referred to as migration, and the resulting seismic section is referred to as a migrated section. Interpreters like to use both, and both unmigrated and migrated data are presented in this atlas.

Why the seismic reflection method?

The science of LITHOPROBE is spearheaded by the seismic reflection method because it is the geophysical technique which produces the best images of the subsurface. These data resolve mappable features such as faults, folds and lithologic boundaries measured in the 10's of meters, and image them laterally for 100's of kilometers and to depths of 50 km or more (Varsek, 1992). Seismic reflection profiling is the principal method by which the petroleum industry explores for hydrocarbon-trapping structures in sedimentary basins. Its extension to deep crustal studies began in the 1960s, and since the late 1970s reflection technology has become the principal procedure for detailed studies of the deep crust.

Seismic data acquisition

The method works by bouncing sound waves off boundaries between different types of rock (Figure 1). As opposed to earthquake seismology, where the location and time of the source is an unknown that needs to be solved for, seismic reflection profiling uses a controlled source to generate seismic waves. On land, LITHOPROBE has been using large truck-mounted vibrators as a source (the "Vibroseis" method), and occasionally dynamite is used. At sea, large arrays of airguns, which rapidly eject compressed air, are deployed. The reflected signals are recorded by geophones, or hydrophones at sea, which resemble ordinary microphones.

Figure 1. Seismic data acquisition.

During a seismic survey, a cable with receivers attached to it at regular intervals is laid out along a road or towed behind a ship. The source moves along the seismic line and generates seismic waves at regular intervals such that points in the subsurface, such as point P in Figure 1, are sampled more than once by rays impinging on that point at different angles. As a shot goes off, signals are recorded from each geophone along the cable for a certain amount of time, producing a series of seismic traces. The seismic traces for each shot (called a shot gather) are saved on magnetic tape in the recording truck.

Seismic data processing

Digital data processing is applied to raw seismic data to produce a seismic section (Figure 2). The following is an example of typical processing sequence.

Figure 2. Seismic data processing.

The data are read from tape and the shotrecords (i.e. all traces recorded for a given shot) are displayed (1). Bad seismic traces, due to noise or a short circuit in the recording equipment, are edited out (2). The traces are then reordered (3) so that each gather of traces belongs to a common reflection point, such as point P in Figure 1. Non-reflected arrivals, such as surface waves and direct arrivals, are removed by digital filtering and/or muting (zeroing of the data) (4). A correction is made for the time the reflected ray spends travelling laterally, so that the reflected arrivals now line up (5). These traces are then added to produce a single output trace (6). This process, referred to as stacking, cancels out random noise and reinforces the reflected signals. The waveform is then shrunk by frequency filtering or deconvolution to improve the resolution (7). Steps (4) to (7) are repeated for each common reflection point, and the resulting seismic traces are displayed as a seismic section (8) which is then interpreted (9).