Tropical cyclones - Extreme Events: A Physical Reconstruction and Risk Assessment - page 99

last 3000--5000 years. A more detailed investigation of two of these sites (Curacoa
Island and Princess Charlotte Bay) by Hayne and Chappell (2001) confirmed the
conclusions of the earlier studies.
The frequency of cyclones throughout the late Holocene at these sites was
determined by radiocarbon dating fragments of coral from the core of the
coral rubble ridges. By dating most, if not all, of the ridges at a site the total
number of ridge building cyclones over time can be ascertained. Of course the
rate of ridge building is dependent upon the number of cyclones passing the
area with time and also the rate of coral reef replenishment or regeneration.
Hayne and Chappell (2001) concluded that the narrow, fringing coral reef, which
wasthe source of rubble for the ridges on Curacoa Island, was able to regen-
erate after 80 years. This time period is considerably less than the apparent
recurrence interval of ridge building (storm deposit) events of 280 years. The
other ridge sites along the length of the Great Barrier Reef showed that ridge
building events have occurred on average every 177--280 years (Nott and Hayne,
2001).
While the northeast Australian palaeorecord shows no change in the fre-
quency of tropical cyclones at a multi-century scale, the US record shows a
2400 year period of increased cyclogenesis during the late Holocene. The record
from Western Lake, Florida suggests that at least 12 high-intensity cyclones (cat-
egory 4 or 5) struck this immediate section of coast over the past 3400 14 Cyears,
with only one event during the past 1000 years. The period between 5000 and
3400 14 Cyearswasalsorelatively quiescent. The record at Shelby Lake, Alabama
shows five prominent sand layers were deposited during the past 3200 14 Cyears
and no sand layers were deposited between 4800 and 3200 14 Cyears. Both records
suggest a marked change in cyclone frequency over time with a distinct phase
of heightened activity between approximately 3400 and 1000 14 Cyears BP (Liu
and Fearn, 2000, 2002).
Both techniques (coral rubble ridges and barrier washover deposits) of recon-
structing long-term cyclone frequencies are limited in their ability to determine
themagnitude of the cyclones responsible. In the case of the coral rubble ridges
it was not known just which magnitude cyclones, or if all cyclones, affecting a
region had remained unchanged in their frequency of occurrence. The same is
also true for the barrier washover deposits because to date no assessments have
been made of the chronostratigraphy of the overwashed barrier dunes. Hence,
some of the sand layers may be from high-magnitude cyclones, others from less
intense cyclones after the barrier dunes are eroded. However, if the barrier dunes
have remained unchanged over the period of the palaeocyclone record, then the
magnitude of the cyclones responsible can be bracketed broadly.