P1: OTA/XYZ P2: ABC
JWST061-11 JWST061-Caers April 7, 2011 13:26 Printer Name: Yet to Come
194 CH 11 VALUE OF INFORMATION
running a well-test analysis, consulting an expert, running logging surveys, doing a reser-
voir modeling study, and so on. The intuitive reason for gathering information is straight-
forward: if the information can reduce uncertainty about future outcomes, decisions can
be made that have better chances for a good outcome. However, such information gather-
ing is often costly. Questions that arise include (a) is the expected uncertainty reduction
worth its cost, (b) if there are several potential sources of information, which one is most
valuable, and (c) which sequence of information sources is optimal. This type of question
is framed under “the value of information problem.” This question is not trivial to answer
because it is necessary to assess this value before any measurement is taken.
Decision analysis and value of information (VOI) have been widely applied to deci-
sions involving engineering designs and tests, such as assessing the risk of failure for
buildings in earthquakes, components of the space shuttle and offshore oil platforms. In
those fields, gathering information consists in doing more “tests” and if those tests are
useful, that is, they reveal design flaws (or lack thereof), then such information may be
valuable depending on the decision goal. It seems intuitive that it is necessary to come
up with some measure of “usefulness” of the test. Indeed, if the test conducted does not
at all inform the decision variable of interest, then there is no point in conducting it. The
“degree of usefulness” is termed the “reliability” of the test in the traditional value of
information literature. In engineering sciences, the statistics on the accuracy of the tests
or information sources that attempt to predict the performance of these designs or com-
ponents are available, as they are typically made repeatedly in controlled environments,
such as a laboratory or testing facility. As will be seen, these statistics are required to
complete a VOI calculation, as they provide a probabilistic relationship between the in-
formation message (the data) and the state variables of the decision (the specifications of
the engineering design or component).
Many challenges exist in applying this framework to spatial decisions pertaining to
an unknown Earth. Instead of predicting the performance of an engineer’s design under
different conditions, the desired prediction is the response of the unknown Earth – which
can be very complex and poorly understood and certainly cannot be put in a laboratory –
due to some external action and the spatial uncertainty. Geophysical surveys and remote
sensing surveys are some of the most commonly used sources of information in the Earth
Sciences. For example, VOI can be used to evaluate whether better 3D seismic coverage
is worth the extra survey costs to explore for untapped oil reserves.
11.2 The Value of Information Problem
11.2.1 Introduction
A catch-22 situation presents itself when determining VOI of any measurement in the
Earth Sciences. VOI is calculated before any information is collected. Note that data col-
lection can be done only once and is rarely repeatable. However, the VOI calculation
cannot be completed without a measure that describes how well the proposed measure-
ment resolves what one is trying to predict. This measure is known as the “data reliability