Here “storm tide” refers to the combined effect of a meteorological storm surge, produced by a combination of high winds and low barometric pressure during storm events, and the astronomic tide. Lower amplitude deviations from mean sea level (in addition to tide), such as those produced by EL Niño Southern-Oscillation (ENSO) – may also contribute to flooding during storm events. In Australia, the tide component may be significantly larger than the surge component, and it is therefore important their cooccurrence probability is modelled.
The effect of waves, which may further elevate coastal water levels by wave setup, is not considered here. Wave setup can be regarded as a short-duration and localised contribution to storm tide flooding. While important for coastal erosion, wave effects alone will rarely lead to sustained and widespread coastal flooding.
The Coastal Oceanography Group at the University of Western Australia (UWA), in conjunction with the Bushfire and Natural Hazards CRC (BNHCRC) produced a 59-year sea level hindcast (1958 – 2016) for the entire Australian region6. The outputs provide continuous hourly sea level records at approximately 1-km resolution around the Australian coast. The three-dimensional finite element hydrodynamic model SCHISM was forced by TPXO tides and JRA-55 atmospheric reanalysis (wind and air pressure) to produce the hindcast.
Extreme value analysis was applied to the sea level data to predict Average Recurrence Intervals (ARI) at approximate 1-km spacing around the entire Australian coastline. An Annual Maxima approach and Generalised Extreme Value (GEV) distribution were used. These data are available via the sealevelx web tool.
The Japanese JRA-55 reanalysis provided wind and sea level pressure fields at 0.5-degree resolution at 3-hour intervals. Previous work suggests the JRA-55 model is robust in capturing storms transitioning from the tropics to extra-tropics7 and it also assimilates tropical cyclone track data to ensure that simulated tropical cyclones follow accurate trajectories.
Tropical cyclone intensities in reanalysis models are universally underestimated2 and as a result the long-term model runs undertaken by UWA-BNHCRC for Australia provide low estimates of the extreme sea levels at longer ARIs in tropical regions.
To account for this, the data available via the sealelvex web tool merges extreme sea level exceedence probabilities derived in previous work focused on tropical cyclones8 with the SCHISM model results used for the rest of Australia. The Haigh et. al model (which Risk Frontiers contributed to) was forced with wind and pressure fields from an earlier version of CyclAUS that synthetically extended the tropical cyclone record to 10,000 years. The stochastic model provided extreme sea levels for tropical cyclones whilst the SCHISM 59-year simulations included tidal and longer term (seasonal, interannual, ENSO) variability.
This means the combined storm tide ARI estimates provided in the sealevelx project include both tropical and extra-tropical cyclone storm influences, in addition to tide and lower amplitude sea level effects.
Interestingly, the only areas where the synthetic tropical cyclone ARI storm tide heights from Haigh et al. exceed those of the SCHISM model (using a 0.5-degree reanalysis that underestimates TC intensity) were for ARIs > 100 years in northern Western Australia and Northern Territory. Speculatively, this may be because:
- tide (i.e. the non-cyclone component) is a large contributor to storm tide around Australia
- storm surge magnitude is not only sensitive to TC intensity, but other track parameters such as forward speed and obliquity to coast which may be well captured in the JRA-55 reanalysis, in addition to local bathymetry
Whatever the reason, it suggests a storm tide hindcast derived from reanalysis does not underestimate extreme water levels at long ARIs for tropical cyclone-affected coasts outside of the Northwest Shelf region.
Storm tide heights (in m, relative to Mean Sea Level, which approximates to Australian Height Datum, AHD), calculated by UWA-BNHCRC and Haigh et al., were extracted at 45 locations around the coast of mainland Australia and Tasmania from the sealevelx web tool for the 5, 10, 20, 50, 100, 200, 500 and 1,000-year ARIs.
- Very High (5) – 100-year ARI water depth > 2.0 m
- High (4) – 100-year ARI water depth < 2.0 m
- Medium (3) – 100-year ARI water depth < 1.0 m
- Low (2) – 100-year ARI water depth < 0.3 m
- Negligible (1) – 100-year ARI water depth < 0.01 m
- Not Analysed (0) – not within the study area
Areas with risk ratings of 2 or higher may experience some damage from storm tide in a 1-in-100 year wind event, unless dwellings and other buildings are engineered to cope with such winds. This assumption is based on a floor height of 0.3m. Slab on ground construction (< 0.3m) is likely to experience damage. “Queenslander” type construction with a significantly elevated living area and no enclosure underneath are less likely to suffer water damage except in the Very High (5) category.
Proximity to shorelines refers to the shortest distance between an address and shorelines of any coastal waters (e.g., rivers, lakes, lagoons and estuaries) that are directly connecting to open ocean. For addresses located within 5km of the coast, a spatial resolution of 90m was used in the analysis, and for addresses located beyond 5km they were assumed to be outside the storm surge zone.