Reciprocal Refraction

When shooting a refraction profile across a river, it is advantageous to exchange the conventional shot and geophone locations for the survey. The river environment is typically very noisy due to the river current and its disturbing effects on the geophone. Placing the geophones on the river bank, and deploying an airgun source off an existing bridge can be very successful in developing a refraction profile. An example is shown in Figure 38. In theory, one could use just 2 geophones, one on the north bank, and one on the south bank. When sorted to common geophone gathers, this would provide reverse profiles for refraction analysis. However, deploying a grid of geophones costs very little more in effort, and opens up the opportunity for array forming to further increase the signal to noise. Selected files from this project may be downloaded from the BSU Database web page.
Go to https://cgiss.boisestate.edu/~pm/BSU/. Click on the check boxe for Horse Shoe Bend and then on submit. It will list the download archive. ID-103.zip contains the raw EGG SEG2 files, processed BSEGY and SEGY files with some scripts. Assuming you have the ID-103 archive and unpacked it in a convenient location, you will note that it has the following directory structure:

{treelist2}
ID-103
|--- bref
|--- seg
|     |---temp
|--- seg2
`--- sgy

treelist2

Figure 38: Reciprocal shooting for refraction surveys across rivers. Bridge foundation investigations benefit from placing the geophones on land, and the source suspended from the bridge in the river.

Packaged with BSU is the Octave program delaytmR.m. The “R” in the name indicates it is for use with reciprocal shooting. This program is nearly identical to the previously described program without the “R” in the name. There is one significant difference, and that is the requirement for an additional water depth file, wds.dat. This file consists of 2 columns, (water depth, shot station number).

The reciprocal nature of the problem is done entirely with BSU programs which resort to common geophone gathers. The roll of shot and receiver are reversed, so pretend that when bref is run, the question of number of shots is answered by the number of receiver gathers. The Octave processing is essentially blind to this reversal in roles, since it merely reads the bref generated matrices and treats the picks as though they were from common shot gathers, with the additional processing of a water layer.