Page 107 - Modern Spatiotemporal Geostatistics

88 Modern Spatiotemporal Geostatistics — Chapter 3
associated with soft data, but at the same time the risk of missing out on
what might be a very important and subtle source of knowledge is maximized.
As a matter of fact, the importance of uncertain (soft) data was recognized
centuries ago. Saint Thomas Aquinas, following Aristotle, argued (see, e.g.,
Schumacher, 1977) that

The uncertain knowledge that may be obtained of the highest
things is more desirable than the most certain knowledge obtained
of lesser things. Summa Theologica

In light of the above considerations, modern spatiotemporal geostatistics
may be viewed as a field of concepts and methods whose boundary conditions
are the available knowledge bases. Both knowledge bases Q (general) and 5
(specificatory) will be used in the BME theoretical construct leading to the
spatiotemporal map of the phenomenon. These two knowledge bases must
mesh in coherent interaction with the new information provided by the map
in order to provide us with an explanatory rationale for the phenomenon of
interest.
The modern geostatistics paradigm requires not only that the mathemati-
cal model or technique chosen be the best possible, but also that the processing
of the various forms of knowledge be achieved by means of logically plausible
rules and the updated knowledge be derived from coherent inferences. These
epistemic requirements are discussed in the following chapter.
By being able to incorporate physical knowledge (about structural con-
nectivity, laws, mechanistic models, etc.), BME may move one step ahead of
empirical (or statistical) methods. Indeed, unlike empirical mapping techniques
that describe an existing set of data and are only locally predictive (i.e., inter-
polation is possible only within the range of the available data), BME is able
to integrate physical knowledge (in the form of scientific laws, empirical rela-
tions, etc.) and, thus, it has explanatory and global prediction features (i.e.,
extrapolation is possible beyond the range of observations). This is important,
if geostatistics is to be considered a respectable scientific discipline.
In "Sources of physical knowledge" (Chapter 1, p. 20), strong emphasis
was laid on the argument that, as an applied scientific discipline, geostatistics
is intended to produce marketable products, capitalizing on the stores of ba-
sic knowledge that have accumulated thus far in a richly productive century.
Surely, the meaning of the term "basic knowledge" in the definition of applied
science above goes far beyond observational facts and includes several other (3
and .5 bases of physical knowledge. Hence, in the following chapters we will
be concerned with the development of a group of modern spatiotemporal geo-
statistics models which have the epistemic and technical capacity to account
for these knowledge bases in a rigorous and systematic fashion.