A number of GUI prompts require responses from the user. These are
described as follows:
- Choose a G matrix file. Select G0003.mod, or perhaps your own edited
version. Click OK
- Choose a D vector file. Select D0003.mod, or your own version. Click
OK.
- Choose an E vector file. Select E0003.mod, or your own version. Click
OK.
- Number of Shots? Enter 2 and click OK.
- Number of Channels? Enter 48 and click OK.
- Enter smoothness constraint. Enter a value of 0.1. What this does
is increase the least squares error slightly, but adds a built in
change to the objective function (the spatial derivative of the delay
time solution). A value of zero here will produce the smallest error
in the solution, but may map picking noise into false structure).
- The text window will show the shot delay times. They should agree
with those specified in the constraint equations.
- The graphic window will have the refractor velocity in the title,
and it should agree with the constraint applied in our problem. The
plotted curve is a graph of the delay time solution, geophones.
- Enter overburden velocity. Here, type in a value of 664 as discussed
above.
- The graphic will show a plot of the ground topography and the refractor
structure. This is one end-limit solution that resolves the delay
times completely as variations in the refractor structure. This is
the preferred solution for this geologic setting.
- Enter a constant refractor depth. This is an alternative solution
which resolves the delay times as variations in the overburden velocity
(perhaps due to variations in water content). If you use the default,
10 m, you will see that the overburden velocity would have to vary
from 1265 to 300 m/s to explain the data. This end-limit solution
is not preferred in this geologic setting.
- Number of constraint equations? Enter 7 here, corresponding to the
number of rows at the bottom of the G0003 file that we edited. This
helps avoid a completely trashed plot where constraints become overprinted
with the time plot. With more than one shot record, there may be some
over plot of the computed times in this last graphic. You can fix
that in Xfig.