Environmental Engineering Reference
In-Depth Information
environmental science research of late, it is very difficult
to get away from hidden complexity because indicating
the whole complexity would disable any capacity for
understanding by others than the model architect.
bottom-up approaches with top-down approaches, at
least while computer power is limited.
Nevertheless, appropriate levels of aggregation of differ-
ent processes need to be defined, not least because they
represent the progress in our level of understanding.
•
As part of this question, parameterization may best be
understood as an emergent characteristic of a system at
different hierarchical levels.
27.1.10 What ismodellingfor?
•
We have elaborated in detail above how modelling is an
important component of explanation. It is a means of
evaluating our lack of understanding (Haraldsson and
Sverdrup; Mulligan and Wainwright, Chapter 11). But as
noted by Kirkby (1992), models can serve a whole range
of different purposes. Amongst other things, modelling
can be used for integration of information from dif-
ferent disciplines (Mulligan), interpolation (Wright and
Hargreaves), knowledge storage and retrieval (Engelen),
communication (Haraldsson and Sverdrup) and learning
(Engelen) and as a digital laboratory. It is a virtual Swiss
army knife to complement the one we take into the field!
Techniques must be developed to deal with errors at all
stages of the modelling (including the data-collection)
process. We know that error propagation is important,
but lack the means of overcoming it.
•
There is a need to encourage reuse of good modelling
concepts whilst ensuring that outdated concepts are not
incentivized to remain in circulation when they have
'passed their sell by date'. Tools need to be developed
that open up modelling to a wider audience through
the removal of the technical overheads which currently
exist in the development and application of models. This
will help break down the barriers between 'modellers'
and 'fieldworkers' and ensure a wider understanding
of the role and purpose of modelling in environmental
science. Generic modelling languages and graphical
model-building tools are the first step in this direction.
•
27.1.11 Moral andethical questions
Perhaps as a subset of the last question, modelling can
be important for addressing questions of environmental
change where direct experimentation would be morally
unacceptable. As well as moral grounds, a number of the
chapter authors suggest we can also address the practi-
cally unfeasible. Since environmental issues incorporate
economic, social, health, welfare, cultural, natural and
geographical factors, there will always be important moral
questions at stake and it is thus vital that we adopt an
appropriate ethical stance towards the results we produce
and the interpretations we allow to be placed upon them.
Models are a potentially important research output
because they can summarize research findings in a
much more interactive (and sometimes more useful)
way than scientific publications. They will never be
mainstream scientific publications until ways are found
to peer review and quality control them. Indeed it is
worrying that many models are so complex that their
construction is rarely transparent in the page-limited
publications that use them. This lack of transparency
is not acceptable for laboratory or field methodologies
and should not be acceptable for modelling experiments
either. The WWW increasingly provides the means
to distribute even complex model descriptions to the
audience of academic journals.
•
27.2 Research directions
There is a tendency to assume that all modellers follow the
same methodology. What should be apparent from the
different contributions in this topic is that there are many
different approaches, and a lot of healthy debate about
the ways forward. We provide here a brief overview of
developments that may be productive in this light. This list
should be read in conjunction with the text of Chapter 1:
27.3 Technological directions
Since the publication of the first edition of this topic
in 2004 computing power has continued to develop at
an exponential rate, as was first described by Gordon E.
Moore (1965). Moore co-founded the semiconductor
chip designers and manufacturer Intel with Robert Noyce,
in 1968. The effective doubling of computer power (mem-
ory, processing speed) every two years (so-called Moore's
Law) - along with similar increases in available magnetic
There is a need to define frameworks for the robust
analysis of environmental systems that deal with their
complex, open character.
•
Complex systems analysis may provide one way forward
in this respect but there is a clear need to integrate
•
Search WWH ::
Custom Search