Deriving robust return periods for tropical cyclone inundations from sediments

• Published in: Geophysical research letters Vol. 40; no. 2; pp. 370 - 373
• Main Authors: Nott, Jonathan F., Jagger, Thomas H.
• Format: Journal Article
• Language: English
• Published: Washington, DC Blackwell Publishing Ltd 28.01.2013; American Geophysical Union; John Wiley & Sons, Inc
• Subjects: Archives & records; Bayesian analysis; Bayesian theory; Beach ridges; Beaches; Catastrophes; Coastal flooding; Cyclones; Distribution; Earth sciences; Earth, ocean, space; Exact sciences and technology; Extreme value distribution; Extreme values; Flooding; Floods; Global climate; Holocene; Hurricanes; inundation return intervals; Mathematical models; Meteorology; Overwash; Probability theory; Records; Ridges; Robustness; Sediments; Statistical analysis; Statistical models; storm surge; Storm surges; Storm tides; Storms; Synthetic data; Tropical cyclones; Australia; bayesian analysis; beach ridges; Australasia; global; upper Quaternary; coastal zone; Severe weather; Tropical cyclone; plains; Storm surge; Cenozoic; Natural hazards; uncertainties; Quaternary; probability; inundations; climate; pressure; loading; sand; Return period; extreme value; Deterministic model; Statistical model; buildings; standard samples; Phanerozoic; Holocene; storms
• ISSN: 0094-8276; 1944-8007
• DOI: 10.1029/2012GL054455

Return periods for tropical cyclone (TC) marine inundations are usually derived from synthetic data sets generated from deterministic models or by extrapolating short historical records. Such approaches contain considerable uncertainties because it is difficult to test their veracity until a sufficiently long period has elapsed. These approaches also often only consider storm surges or storm tides and not the total inundation, which includes waves, set‐up and run‐up, likely to flood a coastal property. An alternative approach is to examine sedimentary records of actual events that occurred throughout the late Holocene. Sedimentary beach ridges are unique amongst the different types of storm inundation sedimentary records because they record variations in the height of total marine inundations rather than a censoring level as occurs with overwash deposits. The limitation in using beach ridges to derive return periods for inundations has been the lack of a robust statistical model that accurately describes the distribution of these events over the past several millennia. Such a model is presented here using a Generalized Extreme Value distribution and Bayesian analysis of a sand beach ridge plain record of extreme TC‐generated marine inundations from northeast Australia. Using this approach, the return period of the marine inundation generated by severe TC Yasi is determined. This return period differs considerably from estimates determined using a probability‐based approach, which extrapolates from a short historical record. With global climate changing and the magnitude of marine inundations expected to increase, there is mounting pressure to develop national standards for marine flood loadings on coastal buildings. Deriving accurate return periods of these events will be critical to this endeavor, and this approach will be applicable at numerous localities globally where storm deposited beach ridges occur. Key Points Return intervals of storm inundation from beach ridges Return intervals based on records of actual events not synthetic data sets More realistic and accurate return intervals for coastal planning