The Inverse Modeling and Geostatistics Project
Research People Papers Software Teaching

Uncertainty Quantification in Geoscience

This project is a purely theoretical project funded from 1. February 2013 by DONG Energy and Production and Lloyds Register (DKK 2.000.000 or $369.000 for the first year).

Summary

The rules of probabilistic data analysis are, in general, well understood. However, application of these methods to concrete problems in geoscience leads to fundamental difficulties. Firstly, the physical/statistical methods for obtaining the required input probabilities are poorly understood, and secondly, it is virtually unknown how to validate probabilistic results. As a result, systemic errors may go undiscovered and unchecked.

This project will clarify what can be inferred from probabilistic results by explicating what they represent and how these representative claims may be substantiated epistemically. The project will establish and describe scientifically defendable ways of quantifying and tracking uncertainty to provide a consistent, integrated approach to probabilistic subsurface characterization from selected geological, petrophysical and seismic data, incorporating rock physics, geostatistics and probabilistic inversion and including validation of probabilistic methodologies.

(Collaborators: DONG E&P, Denmark; Lloyds Register Sweden/Denmark; Prof. dr. Jef Caers, Energy Resources Engineering Department, Stanford University, USA.)