GeoComputation - GeoComputation - page 2

the visualisation of hitherto unseen scientific processes, and it offers a basis for the simulation of
complex systems which are too difficult for economical study by any other route. Computation per-
mits the investigator to test theory by simulation, to create new theory by experimentation, to obtain
a view of the previously invisible, to explore the previously unexplorable and to model the previ-
ously unmodellable. There is clearly considerable potential here that will be released in the new
millennium as computer speeds increase and a computational paradigm becomes a more common
paradigm for doing science in many more areas of scientific interest. It is probably as unavoidable as
it is inevitable but with the greatest developments having to wait for both faster computers and new
generations of computationally minded scientists.
So in science, there is now a strong and growing trend favouring a computational paradigm.
Indeed, many scientific experiments and investigations that were once performed in a laboratory, a
wind tunnel, or in the ield are now being increasingly augmented or replaced by purely computa-
tional alternatives. A common feature of computational science is that there appears to be an under-
lying implicit belief that the quality of the science depends in some way on the speed of the fastest
available computers. As computers have become faster, computational science has emerged as a
powerful and increasingly indispensable method of analysing a variety of problems in research, pro-
cess development and manufacturing. It is now being widely advocated and increasingly accepted
as a third methodology in engineering and scientific research that fills a gap between physical
experiments and analytical approaches. Computer simulations now provide both qualitative and
quantitative insights into many phenomena that are too complex to be dealt with by analytical
methods and which are too expensive or dangerous to study by physical experiments. For example,
the prohibition of atmospheric and underground nuclear weapons testing has stimulated the need to
be able to simulate nuclear explosions by numerical means. In the United States, this military need
has resulted in the Accelerated Strategic Computing Initiative (ASCI) which will eventually serve
many more civilian applications than purely military ones. Indeed it has already spawned the first
teraflop computers. In 1998, these were about 30 times faster than previous machines, for example,
the Cray T3E 1200 at Manchester, United Kingdom. So it is likely that the early years of the twenty-
first century will see increasing availability of teraflop supercomputers and hardware that will be
useful in many other areas including GC. As high-performance computing (HPC) becomes faster,
it stimulates entirely new areas of application which were previously computationally infeasible and
generally unthinkable.
The emergence of computational science is not a particularly new phenomenon although it is one
which has gathered speed throughout the 1990s. The availability of high-performance computers,
high-performance graphic workstations and high-speed networks, coupled with major advances in
algorithms and software, has brought about a silent revolution in the way many scientific and engi-
neering investigations are performed. In 1998, most of the UK research councils had for the first time
a 6-year programme of committed baseline investment in HPC following the inauguration of the
SCAR service late in 1998. See Birkin and Malleson (2014) in this topic for an update on e-Research.
Nevertheless, it should be readily apparent that there are similar attractions for a computational
style of approach in geography and the social sciences. GC can be regarded, therefore, as the appli-
cation of a computational science paradigm to study a wide range of problems in geographical and
earth systems (the geo ) contexts. Note that the geo includes human as well as physical systems. This
extension of the computational paradigm is such an obvious development that it may be a surprise
to discover that the word GeoComputation which seems so generically applicable was only recently
invented. Such is the power of language that a word is almost instantly able to describe whole areas
of research that have existed for two decades or more before the term was invented. A similar claim
could be made for GIS which, when it entered common usage aptly and almost instantly, described
many pre-existing areas of research and provided a central focus for their subsequent development
and dissemination.
The word GeoComputation first appeared in the author's spellchecker dictionary after cof-
fee time discussions relating to a computational geography conference being planned for 1996 in