Geostatistical Reservoir Characterization:

Principles and Hands-on Practice

A Short Course by Clayton V. Deutsch

2001

Objectives:
The principles of geostatistical reservoir modeling or characterization will be covered in
lectures and hands-on exercises. Each participant will have an appreciation for the
variety of geostatistical techniques and tools available to address problems of
heterogeneity modeling and uncertainty quantification. The participants will be able to
apply key techniques and software with real reservoir data. The limitations and
assumptions of different methods will be revealed.
Upon completion of this course, each participant should be able to (1) describe situations
where the application of geostatistical techniques could assist with reservoir modeling,
(2) conduct an exploratory data analysis study for input to geostatistical modelling, (3)
quantify the connectivity of the variables using geostatistical tools such as the variogram,
(4) construct heterogeneous models, and (5) describe the limitations of the resulting
numerical models and the geostatistical model-building techniques.
Who Should Attend:
This course is primarily intended for engineers, geoscientists and computer scientists with
an interest in learning how to build geostatistical reservoir models. Students who have
taken a similar course will benefit from the new course material, new software, and case
study undertaken in the hands-on portion of the course.
Instructor:
Clayton V. Deutsch is a Professor in the Department of Civil & Environmental
Engineering Engineering at the University of Alberta. He teaches and conducts research
into geostatistics. Prior to joining the University of Alberta, Dr. Deutsch was an
Associate Professor (Research) in the Department of Petroleum Engineering at Stanford
University, a Research Specialist with Exxon Production Research Company, and an
Engineer with Placer Dome Inc.
 Course Content

This course will cover essential theory and practice of geostatistical modeling for reservoir
characterization; advanced material is covered at an introductory level. Course notes will be
provided. The second edition of the GSLIB book (Deutsch & Journel, 1998) and Dr. Deutsch’s
new book entitled Geostatistical Reservoir Modelling are available. The GSLIB software and the
winGSLIB interface will be used for hands-on exercises.

Following are the general subjects of each day. There is more material in the course notes and we
will adapt our schedule and pace to the participants. Real reservoir data will be used for the
hands on exercises:

Day One Overview of Geostatistics for Heterogeneity / Uncertainty

Introduction to Course Software
Elementary Statistical Analysis / Data Cleaning

Declustering for Representative Statistics

Coordinate Transformation / Geometric Modeling

Day Two Variogram Calculation and Interpretation

Variogram Modeling

Kriging for Mapping Applications

Cross Validation of Kriging / Estimation Methods

Day Three Sequential Indicator Simulation (SIS) for Facies Modeling

Truncated (pluri)Gaussian Facies Modeling

Surface-Based Facies Modeling / Image Cleaning

Object Based Modeling

Day Four Sequential Gaussian Simulation

SGS with Colocated Cokriging

Postprocessing Multiple Realizations

Connectivity and Ranking of Geologic Models

Additional Days Special topics and hands-on practice