Fermer

Main projects

DeeSse

The DeeSse software is our implementation of the direct sampling multiple-point statistics algorithm. DeeSse can be used to generate time series, or random fields using examples provided as training data sets. DeeSse is extremely flexible and versatile since it can handle a broad range of constraints. The improvements of DeeSse and its extension to new cases is an ongoing research effort that involves mainly Julien Straubhaar in the team, but also several PhD candidates (Przemyslaw Juda, Valentin Dall'Alba, Lucile Chauveau, etc.) who test some new ideas, or develop specific extensions or accelerations.

Geologically constrained inversion

This is one of the core topic of the group. Many different techniques are developped and tested in this framework. We explored for example ensemble method, markov chain monte carlo approaches or importance sampling. We are particularly interested in the links between multiple-point statistics and inversion techniques but explore as well what can be done with novel deep-learning methods or fast machine learning tools like decision trees. The main person involved in that research direction at the moment is Przemyslaw Juda.

Phenix

The Phenix project is funded by the SNF and is dedicated to the development of new methods for the characterization of alluvial aquifers. This project involves currently two PhD candidates (Lucile Chauveau and Alexis Neven), one senior researcher (Julien Straubhaar), a set of external collaborators in particular in the Hydrogeophysics group of the University of Aarhus (Denmark), and local companies and authorities in relation with a test site in the upper Aar Valley, close to Bern (Switzerland).

Roussillon

The Dem'Eaux project is a large research effort devoted to characterize the coastal aquifer of the Roussillon plain in the South of France. The aquifer is exploited and threatened by population growth and climate change. Within that project, Valentin Dall'Alba is conducting his PhD on the geological, geostatistical and hydrogeological modeling of the plain.

Flowcast

We are participating to Flowcast project lead by the Geological Survey of Austria and in collaboration with the Korean Institute of Geoscience.  The aim of this project is to investigate the feasibility of geophysical characterization of karstic features using a combination of electro-magnetic and electric methods. This project is lead by Manon Trottet.

World wide hydrogeological parameter database

You need an estimate of the porosity of a sandstone? or the specific storage a clayey sand? You need statistics about these parameters? It's often tedious to search for such values in books or papers. This is why we worked on developing an open and collaborative data base of world wide hydrogeological parameters. These parameters are classified by scale and rock types. You can consult the data at wwhypda.org.

Do not hesitate also to contribute, just fill a spreadsheet and upload it on the site to share your data with the world.

If you want to know more about the structure or detailed content, have a look at this paper.

Hydrogeologists Time Capsule

Have you ever wondered about the personalities of C.V. Theis, J. Bear, P. Whiterspoon or G. Dagan? All these famous hydrogeologists who changed the way we work. Since more than 10 years, we are collecting video interviews and publish them online. If you want to meet these masters, have a look at the Hydrogeologists Time Capsule web site. This is a project supported by the International Association of Hydrogeologists.

The world wide hydrogeological parameter data base: wwhypda.org.
 

github

The github of the group is available there : https://github.com/randlab