# Book Review—Numerical Earth Models, Jean-Laurent Mallet

## New EAGE publication provides math behind, and insights to, commercial earth
modeling solutions.

Numerical Earth Models* (NEM)
is a good blend of discussion and math, and, although we didn’t understand much
of the latter, the subtext turns out to be bang on in terms of the actualité
of commercial earth modeling. Indeed, Prof. Mallet, recently retired
from his position at the Nancy, France, School of Mines and father of GoCad,
has performed a remarkable feat in slipping in a compelling piece of marketing
for Paradigm’s new ‘Skua’ technology under the guise of an EAGE ‘Education Tour
Series’ publication.

Mallet takes it as a given that
earth modeling is the basis of E&P decision making. He dismisses the ‘classical,’
i.e. engineering CAD approach as being inadequate for modeling geological complexity
and presents his ‘discrete smooth interpolation’ approach as used in GoCad.
NEM includes some pretty good illustrations of complex geologies to back up
the mathematical exposé. The math is
tough, but accompanied by copious and relatively clear explanations for the
mathematically challenged.

Mallet moves on from topological
models to geostatistics, offering many examples of stochastically generated
channels and geobodies. A chapter on seismic interpretation is a little idiosyncratic
as it appears to dismiss conventional Fourier approaches to propose ‘trigonometric
polynomial’ modeling of the seismic trace.

But the payload of NEM is in
combining all of the above in the shared earth model and its successors, blending
seismically derived attributes with the geometrical model. Mallet makes a good
case for the inadequacy of ‘conventional’ pillar grid approaches of Schlumberger’s
Petrel and many other tools including previous incarnations of GoCad.

Mallet’s answer is the ‘unified
earth model,’ an approach that removes structural influences before performing
cellularization and attribute modeling. A final chapter extends all of the above
to upscaling and fluid flow modeling.

NEM is an extraordinary publication
that conveys Mallet’s enthusiasm for modeling the earth, seismics, fluid flow
and just about everything else. Is it ‘commercial?’ Definitely, and the EAGE
is to be congratulated for producing a great blend of product-oriented discussion
and science.

* EAGE
2008/ISBN 978-90-73781-63-4 and www.oilit.com/links/0807_9

This article originally appeared in Oil IT Journal 2008 Issue # 7.

*For more information or to comment on this topic email here*.