Next: Velocity Analysis to Determine
Up: Basic Processing Steps
Previous: Velocity Filter Data to
The CMG traces in Figure 1.9
are misaligned with one another
so that a brute stack will produce cancellation of the signal.
To avoid this cancellation we flatten out the
reflections by using
equation 1.6
to apply
a normal moveout (NMO) time shift
t_{NMO}(x) to the data,
where
t_{NMO}(x) 
= 
t(x)t(0), 


= 

(1.10) 
and we have replaced the v_{RMS} by v_{NMO}.
Once flattened, the traces in a NMOcorrected
gather can be stacked together for constructive reinforcement
of the reflection events.
But how do we in practice determine the v_{NMO} values?
A systematic means for determining
Vnmo is described in the next section on velocity analysis.
The MATLAB script for NMO corrected data is given below:
%
% data(x,t) = CMG data
% datanmo(x,t) = CMG data with NMO correction
% t0 = 2way normal incidence time
% v(t0) = Stacking velocity as a function of 2way normal incidence time
%
for x=1:nx;
for t0=1:nt;
tx = sqrt(t0^2+(x*2/v(t0))^2);
datanmo(x,t0) = data(x,tx);
end;
end;
Application of a script like this to the LHS of
Figure 1.16 will
"flatten" the primary reflections to give the
NMO corrected traces shown on
the RHS.
Figure 1.16:
Mobil's
Gulf of Mexico CMG (LHS) before
and (RHS) after NMO correction.
Note, the absence of surface waves (why?) and the cleaner
appearance (no static problems) of
these marine records compared to the messy land data from Utah.
Marine data, typically, are cheaper to acquire and cleaner
than land data. The primary problem with marine data, however,
is usually the presence of seafloor multiples.

Next: Velocity Analysis to Determine
Up: Basic Processing Steps
Previous: Velocity Filter Data to
Gerard Schuster
19980729