Introduction - Extreme Events: A Physical Reconstruction and Risk Assessment - page 2

to interpret this natural variability. Nature, however, effectively records its own
extreme events, and often its not so extreme ones, providing us with a docu-
mented history in the form of natural records that can display the full range
of variability of most hazards that confront society. Long-term records, there-
fore, are the only real source for uncovering the true nature of the behaviour
of natural hazards over time.
Scientists who study the Quaternary -- the most recent period of geologi-
cal time (or approximately the last 2 million years) know that natural events
including hazards often occur in clusters or at regular to quasi-periodicities.
Over longer time intervals, the Quaternary record shows us that the periodic-
ity of events, including climatic changes, is governed by many factors, some of
which are external to the Earth. For example, climatic changes of various scales
occur at intervals from 100 000 years to 11 years based upon regular variations in
theorbit of the Earth around the Sun, along with the tilt of the Earth's axis and
precession of the equinoxes, to sunspot activity. There are many other regular
cycles of climate change that occur in between the 100 000 and 11 year cycles
that have been uncovered from a variety of natural records. The clear message
from Quaternary records is that many of the climatic changes that occur on
Earth do not occur randomly and in this sense are not independent of time. The
same could be expected of many extreme natural events.
While it is true that natural records do not document the event as accurately
as the instrumented record, they are nonetheless of sufficient precision to show
us the magnitude of the most extreme events and how often these events are
likely to occur. Even more importantly, natural records provide us with a very
effective means by which to test the assumptions of the stationarity of the mean
and variance, and randomness of occurrence of a natural hazard. In the absence
of these tests we cannot hope to realistically assess community vulnerability and
exposure, and therefore risk. Social scientists, involved in that part of the risk
evaluation process devoted to community and social parameters, also need to
be aware of the assumptions made by scientists and engineers in determining
thephysical nature of the hazard. This is frequently not the case and planners
are left with false impressions of which areas are safe for urban, industrial and
tourism developments.
We can only really attempt to reduce risk from natural hazards when we
factor the dependence on time of a hazard into our risk equations, or at least
test for it, and then see where the historical record fits in the sequence. Unfor-
tunately, this approach is rarely adopted. It is imperative that we examine each
of the forms of evidence, being the instrumented, historical and prehistoric
records when undertaking risk assessments. Many practitioners, however, are
unfamiliar with prehistoric records and are hesitant to incorporate them into