Jo Eidsvik, IMF/NTNU 15.02.2016

Abstract: The course covers multidisciplinary concepts required for conducting value of information analysis in the Earth sciences, with a particular focus on petroleum. The value of information is computed before purchasing data, and can be useful for checking if data acquisition or processing is worth its price, or for comparing various experiments. Participants will gain an understanding for the integration of spatial statistical modeling and decision analysis for evaluating information gathering schemes. Examples demonstrate value of information analysis in petroleum exploration and development, where the decision maker could make better decisions by purchasing information (at a price) via exploration wells, seismic data or electromagnetic data.

Material: The course material is based on "Eidsvik, Mukerji and Bhattacharjya (2015). Value of Information in the Earth Sciences, Cambridge University Press".

Background: Some background in probability and statistics and mathematical modeling is recommended. The course is of interest to scientists who use statistics or decision analysis techniques and perform quantitative analysis of geophysical data.

Henning Omre, IMF/NTNU 15.02.2016

Abstract: The course covers methodology for inverting seismic AVO data into relevant reservoir variables like lithologies, porosity and fluids. Observations from one well is used as training data to calibrate the inversion model. The inversion is cast in a Bayesian inversion framework with prior models capturing general reservoir experience and with likelihood models representing the various types of reservoir specific observations. The resulting solution to the inversion will be a posterior model for the reservoir variables of interest. Both simulations which reflects the heterogeneity of the variables and the best predictions with prediction intervals can be provided based on this posterior model. Participants will gain understanding in probability and statistics relevant for Bayesian inversion of seismic data. Moreover, specific models for inversion of seismic AVO data will be discussed in some detail - and many examples/case studies will be presented.

Material: The course material consists of course notes and copies of publications, example: "Rimstad, Avseth and Omre (2012). Hierarchical Bayesian lithology/fluid prediction: A North Sea case study, Geophysics, 77, B69-B85".

Background: Some background in probability/statistics and geophysics is recommended.

Henning Omre, IMF/NTNU 15.02.2016

Abstract: The course covers methodology for assimilation of observations of spatio-temporal phenomena represented by numerical models. More specifically, reservoir description conditioning of production data from wells and/or lime-lapse seismic data. The acquisition is phrased as a spatio-temporal Bayesian inversion problem including likelihood and prior models. The solution is the posterior model for the reservoir variables on which simulation of realizations and predictions with prediction intervals can be based. Assessing this posterior model is usually complicated and algorithms like Ensemble Kalman or Particle filters will be presented. Participants will gain understanding in probability and statistics relevant for data assimilation in spatio-temporal models. Moreover, specific models for production data and time-lapse seismic data will be discussed. Several examples/case studies will be presented.

Material: The course material consists of course notes and copies of publications, example: "Sætrom, Hove, Skjervheim and Vabø (2012). Improved uncertainty quantification in the ensemble Kalman filter using statistical model-selection techniques, SPE Journal, 17, 152-162".

Background: Some background in probability/statistics and mathematical modelling is recommended.