Re: geostatistics and bedform characterisation

In article <[EMAIL PROTECTED]>, Thierry
Schmitt <[EMAIL PROTECTED]> wrote:

> I wonder if someone has got any idea or pointers about the use of
> geostatistic methods such as semi-variogram or auto-correlation on
> DTMs of bedforms (dunes or ripples).

If you do a literature search you might find some useful stuff on this. 
I would start by looking in Mathematical Geology.  I haven't done a
search, but I have some thoughts that might help you.

If you are thinking mostly about data analysis, variograms might help
you quantify some basic parameters like ripple size, anisotropy, and
orientation.  One thing you might need to watch out for (depending on
your needs) is the fact that standard variorgam analysis will not be
able characterize bedform cross section asymmetry.   A sawtooth profile
(a simple geometrical example not intended to represent any bedform...)
looks the same to a variogram whether the teeth lean to the right or
the left.

If you are thinking about modeling rather than analysis you can
probably come up with some creative combinations of gaussian simulation
with various variogram shapes, followed by transformations, to
stochastically simulate bedforms.  This is more geostatistics as an art
form, rather than geostatistics as an analysis method, but that is not
a bad thing.  I have done something like this to model rudist mounds
and the associated debris deposits with some success.  Because of the
variogram limitation mentioned above it will be more difficult to make
bedforms with cross section asymmetry, but not impossible.

-- 
Jim Jennings, Bureau of Economic Geology, The University of Texas at Austin
reply to jim dot jennings at beg dot utexas dot edu