Long Paddock


Please let us know what you think of the Long Paddock website. Take 2 Minute Survey!

Stage 3 Report, July 2004

Report cover

Report cover

Surge Plus Tide Statistics for Selected Open Coast Locations along the Queensland East Coast

Executive summary

The frequency of storm surge plus tide during tropical cyclones was determined for 50 open coast locations along the east coast of Queensland, Australia (Figure 1). The goal was to produce return period curves for storm surge plus astronomical tide for return periods between 10 and 1000 years.

A series of sophisticated models was employed. First a tropical cyclone track and pressure model produced a synthetic dataset consisting of the time series of position and pressure for almost 10,000 storms. This represents 3000 years of data for the western Coral Sea. Numerical models with three nested grids with increasing spatial resolution for storm surge were established for the study areas. The storm surge model was validated in Phase 1 of this overall study (Harper et al., 2001).

It was computationally impossible to model all 10,000 storms to the needed resolution; therefore, a system was developed to determine which storms would contribute to return periods above 10 years. To do this all storms were modelled on the less computationally expensive coarse grid. The most severe results on these coarse simulations were used to cull the number of storms to be modelled on the finer resolution grids from 10,000 to about 500 for each of the 20 fine resolution C-grids.

An astronomical tidal signal was created using tidal analyses. Each storm surge time series from a single storm was linearly added to 500 separate tidal time series. The tide series were randomly chosen (with a weighting to reflect the monthly change in cyclone frequency) from a long tidal record. The maximum storm surge plus tide water level during each storm-tide event was determined and these values were ranked by magnitude and return period curves were created.

Establishing a datum and a tidal range at the project output points caused considerable difficulty. Most of the output points were not at established tidal measurement stations. Hence the datum at a location was often transferred from the nearest tide station and these official values (that were provided to us) did not always provide the best tidal information at that location. Considerable time and energy was expended to check and recheck values of MSL, AHD and HAT. Some official values were updated during the project in consultation with Maritime Safety Queensland. In the end, the decision was reached to model storm surge relative to the official MSL and use the official transfer from MSL to AHD.

The effect of greenhouse-induced climate change was investigated. Three separate scenarios were tested. These were (A) combined effect of an increase in maximum intensity (MPI) by 10% and a poleward shift in tracks of 1.3°. (B) increase in frequency of tropical cyclones of 10%. (C) mean sea level rise of 300 mm. In general the mean sea level rise is the most important effect especially at lower return periods. The 10% increase in tropical cyclone frequency is insignificant. The combined increase in intensity and poleward shift in tracks becomes increasingly significant with larger return periods. Both the magnitude and probability of greenhouse-induced mean sea level rise are more certain than greenhouse-induced changes in tropical cyclone frequency, central pressure, or track.

For all project reporting locations, the occurrence of a tropical cyclone was defined as any that occurred in the western Coral Sea regardless of its distance from the location. This has the property of merging the return period curve for tropical cyclone-induced storm tide into the return period curve for astronomical tide at the lower end of the curves. A more selective definition of tropical cyclone occurrence would have caused the return period curve of storm tide to decrease rapidly (sometimes below HAT) at the lower end of the curve. The adopted definition was used to avoid any misinterpretation of the frequency of water levels at return periods that may be dominated by non-cyclonic events.

There are several influences that affect water levels that are not considered in this study.

The largest storm surges in the synthetic ensemble, although severe, are not the largest possible. The probable maximum water level at a given location would be caused by a tropical cyclone and tide with the following characteristics. (1) very severe central pressure,
(2) large radius to maximum winds, (3) landfall point at a distance equal to the radius of maximum winds to the north, (4) forward speed of the eye equal to the short wave speed offshore and the shallow water wave speed over the shelf, and (5) most importantly for the Queensland region with its large tidal ranges, an astronomical tide level, at the time of maximum surge that is close to HAT. The combination of these characteristics would be very rare, but not impossible.

If an estimate of the probable maximum level were calculated by adding the largest storm surge from the 3000 years of simulations to the HAT level, then this would result in a total water level relative to AHD of about 5.3 m at Cairns, 6.8 m at Townsville and 6.8 m at Mackay Town Beach. These are approximately 2.4, 3.5 and 2.5 m, respectively above the 1000 year water levels. It must be emphasised that these probable maximums are not accurate estimates. As discussed in Section 5, the maximum Townsville surge appears to be an outlier, and this highlights the random and infrequent nature of very severe tropical cyclones.

A caution is necessary on the possibility of water levels much higher than the 1000 year levels that are presented in this report. The occurrence of the probable maximum water level could have devastating consequences for a nearby community. Although the probability of occurrence is very rare, a calculation of the risk (probability times consequences) is an important component of both disaster and longer term land use planning.

Download full report

* Requires Adobe Reader

Last updated 22 June 2009

About us