Tropical cyclones - Extreme Events: A Physical Reconstruction and Risk Assessment - page 104

more active and likely to occur, and similar periods when they are less likely
to occur? It is difficult to answer these questions definitively at this stage but
some research suggests that there may be periods when cyclogenesis is active
and at other times not. The 1900s in Queensland, Australia appear to have been
relatively quiescent compared to the early 1800s when many more very intense
cyclones struck the coast (Nott, 2003). Quasi-periodicities also appear to be evi-
dent in the US historical hurricane record. The frequency of all hurricanes in
thetropical north Atlantic region did not vary over the last century but the fre-
quency of intense storms (Saffir--Simpson categories 3--5) was variable and cyclic
at decadal scales (Landsea, 2000). The 1000 year record from southern China also
shows decadal episodes of enhanced cyclogenesis with two periods between AD
1660--1680 and AD 1850--1880 experiencing a much higher frequency of cyclones
occurring than at other times (Liu et al ., 2001).
Risk assessment of tropical cyclones using historical versus
prehistorical records
There is little doubt that prehistoric data do not record the magnitude
or frequency of cyclone events as accurately as modern instrumented data. How-
ever, by its very nature prehistoric data often preferentially record the most
extreme events, as subsequent lesser magnitude events are unable to remove
this information from the landscape. Hence, when extreme magnitude cyclones
have occurred in the past they can often be clearly differentiated from those in
thehistorical record, for the palaeocyclone surge deposits and erosional terraces
sit much higher above sea level, or farther inland, than the highest recorded his-
torical event. Likewise, while prehistoric chronologies derived using geological
dating techniques are less precise than historic chronologies, the uncertainty
margins of the technique (radiocarbon dating) allows a maximum possible age
(at the 95% probability level) to be determined. The beginning of the historical
record provides a minimum age if no events of these magnitudes are recorded
historically. Hence, while the prehistoric data have obvious shortcomings there
is no doubt that it is precise enough to reasonably determine when these events
occurred, and the extent to which they differ in magnitude to those recorded
historically. When prehistoric data register events larger than that seen in the
historic record it becomes a very valuable source of information because it shows
that the historical record does not include the entire range of cyclone events
likely in a given region. This has substantial implications for risk assessments.
Few comparisons have been made between the historic and prehistoric records
of cyclones for a region. Murnane et al .(2000) compared model predictions of
theprobabilities of occurrence of wind speeds during tropical cyclones, based on
the historical record, with the palaeocyclone record from the washover deposits