News from M2RC, Multivariate Recoverable Resource Consortium - December 2007

Most of the partners met together in Paris in December 2007 to review the final outcomes of the M2RC research and development. It turns out that the consortium has successfully reached its objectives. Given this success, GEOVARIANCES is proposing a 3rd year of research-development.

Indeed:
- the Multivariate Uniform Conditioning application has been achieved as well as the Direct Block Simulations,
- all the existing multivariate algorithms in ISATIS have been reviewed and improved: Bivariate Grade-Tonnage Curves, Gaussian Anamorphosis Modeling, Histogram Modeling, Support Correction, Variogram Regularization.

The M2RC-related new functionalities will be available to the M2RC partners in the next version ISATIS 8 which will be released in in the next few months (Q1/2008).

(PNG)
M2RC Multi Uniform Conditioning & Direct Bloc Simulations

Using these new algorithms, our geostatistical experts have run case studies using datasets provided by each M2RC member to validate the new applications.

These case studies have particularly been beneficial to discuss the limits of the different geostatistical models. What is the mean of my deposit? What is the link between this mean and my domaining? Is the stationarity a real issue when running simulations ? What can I do with my simulations in the day to day work ? All these discussions are of primary importance to improve the programming of ISATIS and its different workflows.

A third year of research/development for M2RC

Given the success of these two first years, GEOVARIANCES is offering to extend the consortium for a third year. The following research axes are proposed:
- Zero Effect: Treatment of the Zero Effect in the anamorphosis modeling and the gaussian transformation.
- Categorical and Continuous Variables: Combination of the two types of information.
- Simulations Post-processing Tools: What can we do with simulations?
- Domaining: Impact of domaining in geostatistics.
- New Anamorphosis and Zero Effect: Implementation of a new modeling and use of Gibbs Sampler Techniques.
- Conditional Expectation Tools: Grade tonnage curves with cutoffs on the main variable.
- 2007 Consortium Tools Improvements: Taking into account borderline case studies.

Our aim is to be as efficient as during the first two years in addressing the real questions of the practitioners and giving them some realistic solutions.

Three of the members are willing to continue. They agree to focus the 3rd year developments on the estimation/simulation of categorical and continuous variables taking into account the domaining issues as well as on the implementation of the new histogram modeling with zero effect treatment.

Appointment is made for the 18th of June 2008 in Paris to get the 2006/2007 return of experience from the partners and review the 3rd year current developments.

Note that a publication on the consortium work will be made during GEOSTATS 2008 congress in Santiago de Chile by Jacques Deraisme and Pedro Carrasco (from Codelco).

If you wish to become a partner of M2RC or for more information, please contact Julien Tan or call him at +33 1 6074 9102

More info on M2RC...