[CIG-SHORT] fault parameters on complex geometry

[Brad Aagaard]

Hongfeng,

There are several different ways to setup a spatial database to specify 
the spatial variation of a value for complex, nonplanar fault geometry:

(1) Use points on a horizontal surface with interpolation
(data-dim=2, query_type=linear)

   For this case I usually use z=0. This case works well when the 
spatial variation is given in terms of horizontal coordinates (for 
example lon/lat converted to x,y).

(2) Use points on a vertical surface with interpolation
(data-dim=2, query_type=linear)

   This case usually works well with a nearly planar dipping fault.

(3) Use points from your meshed fault surface without interpolation 
(data-dim=3, query_type=nearest)

   If you have very complex geometry or spatial variations, sometimes 
the best way to setup the spatial database is to use a script that grabs 
the points on the fault surface from your mesh, computes the values at 
the points, and then writes the spatial database. This results in big 
files, but the queries are usually relatively quick because only the 
nearest point in the spatial database needs to be found.

   Note that with query_type=nearest, the data-dim is not important, but 
you may want to specify the appropriate value to remind yourself whether 
the points lie on a plane, line, etc.

Regards,
Brad