02.12.2014 Views

Annual Report 2000 - WIT

Annual Report 2000 - WIT

Annual Report 2000 - WIT

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

û<br />

ò é<br />

from ã äñ<br />

to ã î§ñ Þµó î=<br />

and ò<br />

ã<br />

4<br />

Ý<br />

2<br />

7:<br />

6ã<br />

:<br />

äñ<br />

to ã<br />

2<br />

7<br />

:<br />

:<br />

7<br />

þ<br />

7<br />

798 7<br />

Ý<br />

û<br />

3<br />

5<br />

3<br />

äñ<br />

to ã<br />

þ<br />

206<br />

äñ<br />

< ó î=<br />

: ; ò é ò ñ ò :<br />

6 6 6<br />

î§ñ Þ—ó î= ã î§ñ ã<br />

ã î§ñô ó î=<br />

from ã<br />

î§ñô ó î=<br />

are required.<br />

î§ñ<br />

,<br />

Figure 1: Determination of the coefficients ú²ñ¦= and ã ==<br />

: traveltimes ò ñ<br />

from ã<br />

The move-out velocity -/.10 å?> ýA@<br />

ò ñ ã<br />

û<br />

is a good approximation for the RMS-velocity<br />

of the Dix formula. Therefore our technique can be considered to be an extension of<br />

the well known , -method to arbitrary 3-D heterogeneous media.<br />

Ü—ÝöÞâßàÝ<br />

IMPLEMENTATION<br />

Equations (1) and (4) state that for small variations of óµã ä<br />

and ó¿ã î<br />

traveltimes can<br />

be interpolated with high accuracy if the according coefficient sets are known. This<br />

means that we can not only interpolate in between receivers but also in between<br />

sources. (Ursin, 1982) has presented examples for coefficients determined from<br />

ray tracing. We can, however, use (1) and (4) not only for interpolation but also<br />

for the determination of the coefficients if traveltimes for certain source-receiver<br />

combinations are given. Since we aim for migration such data are available. We will<br />

now demonstrate how to obtain the coefficients from traveltimes sampled on a coarse<br />

grid. Subsequent interpolation onto the required fine grid can then be carried out.<br />

In the following we will refer to cartesian grids for both source and receiver positions.<br />

Sources are located -B in the -plane with B , and corresponding to the indices<br />

ø•ñ<br />

and 5 ú²ñ<br />

as<br />

4 5 ã<br />

1,2 and 3 of the previous section. To determine the slowness ã vectors<br />

well as the second order derivative ã matrices ã , ã and , we need tables containing<br />

traveltimes from the source to each subsurface point for eight neighbouring sources of<br />

the source under consideration ã äñ<br />

at (for a 2D model the number of additional sources<br />

reduces to two). These additional sources are placed on the -B -grid with a distance<br />

to the central source that coincides with the coarse grid spacing 5 5<br />

and óCB . Please<br />

ó<br />

note that the method is not restricted to cubical grids. The components of ã<br />

, ã ø•ñ<br />

carrying indices and B only are determined directly from the traveltime tables, as<br />

ú€ñ<br />

and 5 . ã<br />

are all elements of ã<br />

and<br />

To give an example: to compute ú€ñ¦= and ã ==<br />

we need only the traveltimes ò ñ<br />

, ò é<br />

and

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!