Flay, Shaun Alexander (2006) Climatology of Queensland landfalling tropical cyclones: evaluating instrumental, historical and prehistorical records. PhD thesis, James Cook University.

	PDF (Thesis front) Download (206kB)
	PDF (Thesis whole) Download (2MB)

Abstract

Knowledge of the probability of occurrence of major tropical cyclone events forms an integral part of developing hazard mitigation strategies in regions prone to their impact. Queensland landfalling tropical cyclones are analysed in this study in order to provide a climatology of event frequency and magnitude. The adopted modelling approach differs from previous studies in the region in that it focuses specifically on the incorporation of historical and prehistorical information. The availability of such records offers a means to test whether the satellite-based instrumental record, which encompasses only the last few decades, provides a representative sample with which to characterise extremes of the process.

A methodology based on Bayesian statistical techniques is presented and applied to facilitate the incorporation of historical information. Through this approach, historical observations are specified as prior information for models of seasonal activity and intensity. When combined with reliable information from the instrumental record, subsequent inferences on the landfall climatology can be made with greater precision and confidence. Among the statistical models considered is a Poisson distribution for seasonal counts, a Generalised Linear Model that incorporates an index of ENSO as a predictor for seasonal activity, and a Generalised Pareto Distribution for tropical cyclone minimum central pressures.

The inclusion of historical information is shown to lead to increased certainty in parameter estimates for these models. Furthermore, the incorporation of historical information on storm intensities leads to predictions of the frequency of major landfall events that are higher than what would be expected from an analysis using only the instrumental record. A further outcome of implementing a Bayesian strategy is the development of predictive distributions showing the probability of specific levels being reached in future periods.

A trend analysis identified the presence of decadal to multi-decadal variability in seasonal storm numbers and in the strength of the relationship between ENSO and tropical cyclone activity. A statistically significant downward trend in landfalling storm intensities over the 20th century was also detected. For Coral Sea region tropical cyclones, there is also evidence of decadal variability in storm numbers. Interestingly, a marginally significant upward trend in peak intensities is found for the Coral Sea region over the period 1960/61-2004/05, which is in contrast to the general downward trend in landfall intensities. No evidence is found to suggest that ENSO has a direct effect on either regional or landfall storm intensities.

A simulation model was then derived from the Coral Sea region satellite record and subsequently applied to generate a series of landfall events. Comparison of the observed landfall record with this simulated series showed close agreement. A further comparison of observed and simulated records with prehistoric data, previously reconstructed from storm ridge sequences found throughout the Great Barrier Reef region, showed some discrepancy. In particular, estimates based on observed and simulated data were tending to underestimate the frequency of major events. Uncertainties inherent in reconstructing storm intensities from the geological record complicate the utility of this comparison, however, suggesting further work is needed to address the use of prehistoric records as an independent data source.

Item ID:	17525
Item Type:	Thesis (PhD)
Keywords:	tropical cyclones, landfall, risk prediction, forecasting, cyclone tracks, storm counts, storm intensities, ENSO, El Nino southern oscillation, seasonal activity, information sources, instrumental record, geological record, historical observations, Coral Sea, simulations, mathematical models
Date Deposited:	28 Nov 2011 23:23
FoR Codes:	04 EARTH SCIENCES > 0401 Atmospheric Sciences > 040105 Climatology (excl Climate Change Processes) @ 100%
SEO Codes:	96 ENVIRONMENT > 9602 Atmosphere and Weather > 960202 Atmospheric Processes and Dynamics @ 100%
Downloads:	Total: 481 Last 12 Months: 18
	More Statistics