The 2004 Sumatra-Andaman Earthquake and Indian Ocean Tsunami shattered the paradigm that guided our understanding of giant subduction zone earthquakes: that massive, magnitude 9+ earthquakes occur only in subduction zones experiencing rapid subduction of young oceanic lithosphere. Although this paradigm forms the basis of discussion of subduction zone earthquakes in earth sciences textbooks, the 2004 earthquake was the final blow in an accumulating body of evidence showing that it was simply an artefact of a sparse and biased dataset (Okal, 2008). This has led to the realization that the only factor known to limit the size of megathrust earthquakes is subduction zone length. This new appreciation of subduction zone earthquake potential has important implications for the southern Asia-Pacific region. This region is transected by many thousands of km of active subduction, including the Tonga-Kermadec, Sunda Arc, and the Makran Subduction zone along the northern margin of the Arabian Sea. Judging from length alone, all of these subduction zones are capable of hosting megathrust earthquakes of magnitude greater than 8.5, and most could host earthquakes as large as the 2004 Sumatra-Andaman earthquake (Mw=9.3). Such events are without historical precedent for many countries bordering the Indian and Pacific Oceans, many of which have large coastal populations immediately proximate to subduction zones. This talk will summarize the current state of knowledge, and lack thereof, of the tsunami hazard in the southern Asia-Pacific region. I will show that 'worst case' scenarios threaten many lives in large coastal communities, but that in most cases the uncertainty in these scenarios is close to 100%. Is the tsunami risk in SE Asia and the SW Pacific really this dire as the worst-case scenarios predict? The answer to this question relies on our ability to extend the record of tsunamis beyond the historical time frame using paleotsunami research.

The major tsunamis of the last few years have dramatically raised awareness of the possibility of potentially damaging tsunami reaching the shores of Australia and to the other countries in the region. Here we present three probabilistic hazard assessments for tsunami generated by megathrust earthquakes in the Indian, Pacific and southern Atlantic Oceans. One of the assessments was done for Australia, one covered the island nations in the Southwest Pacific and one was for all the countries surrounding the Indian Ocean Basin

Keynote presentation to cover * the background to tsunami modelling in Australia * what the modelling showed * why the modelling is important to emergency managers * the importance of partnerships * future challenges

The high risk of natural disasters in developing nations has considerable implications for international aid programs. Natural disasters can significantly compromise development progress and reduce the effectiveness of aid investments. In order to better understand the threat that natural disasters may pose to its development aid program, AusAID commissioned Geoscience Australia to conduct a broad natural hazard risk assessment of the Asia-Pacific region. The assessment included earthquake, volcanic eruption, tsunami, cyclone, flood, landslide and wildfire hazards, with particular attention given to countries the Australian Government considered to be of high priority to its development aid program. Geoscience Australia's preliminary natural hazard risk assessment of the region aimed to help AusAID identify countries and areas at high risk from one or more natural hazards. The frequency of a range of sudden-onset natural hazards was estimated and, allowing for data constraints, an evaluation was made of potential disaster impact. Extra emphasis was placed on relatively rare but high-impact events, such as the December 2004 tsunami, which might not be well documented in the historical record. While a detailed risk assessment was well beyond the scope of this study, it was recognized that some understanding of the potential impact of natural disasters could be achieved through the simple means of developing appropriate overlays of population and hazard. For example, given an estimate of the frequency and magnitude (VEI) at which volcanic eruptions in a certain region occur, the populations impacted could be roughly estimated by considering the average population close enough to a volcano to receive a significant impact from ash fall.

The report summarises earthquake and tsunami information worldwide in 1997 but with a focus on Australia for use by scientists, engineers and the public. Maps of the seismicity are presented on a state-by-state basis and isoseismal maps are included for the significant earthquakes.

As the tsunami moves across the open ocean it is almost undetectable on the ocean surface. In this example, the tsunami waves are only about half a metre high but have a wavelength of 200 kilometres. Travelling at speeds of up to eight or nine hundred kilometres an hour (the speed of a commercial passenger jet), it will take each wave about 15 minutes to pass a slow moving ship.

As the Australian plate slowly pushes under the Eurasian plate, massive stresses build up in the crust. These stresses also cause the Eurasian plate to be slowly forced upwards - part of the process that builds the mountains and volcanoes of Indonesia, as well as creating the many earthquakes felt in that region of the world each year. When the stresses get too great, the plates will suddenly slip causing massive movements in the seafloor. The part of the crust nearest to the fault zone rapidly moves upwards by a metre or so, lifting the entire body of water above it. A hundred kilometres away the opposite may happen: the seafloor drops and the ocean above it also falls. These two movements (the sudden rise and fall of the seafloor hundreds of kilometres apart), combine to cause a series of tsunami waves which move away from the line of the fault in both directions.

Tsunamis can be produced from volcanoes in a number of ways. During a volcanic eruption, hot fast moving bodies of gas and rock (known as pyroclastic flows) can travel into the ocean, pushing the water outwards and creating a tsunami. In other eruptions, the volcano may collapse inwards or produce large landslides, both of which can cause tsunamis. More than 90 volcanic tsunamis have been recorded worldwide in the last 250 years. The 1883 Krakatau eruption in Indonesia caused tens of thousands of deaths, including 77 about 800 kilometres away from the eruption. The effect of the tsunami was reported up to 10 kilometres inland and one large ship was raised 10m above sea level and carried 3 kilometres inland.

The tragic events of the 2004 Indian Ocean Tsunami highlighted the real threat posed by tsunamis to coastal communities worldwide. With subduction zones to the north and east of Australia, tsunamis pose a real threat to the Australian coast. Geoscience Australia has been developing methodologies for quantifying the severity of tsunami impacts to assist emergency management authorities in planning for this threat. Tsunami inundation modelling is computationally intensive and is often restricted to a small number of discrete communities. As a result, communities must be prioritised for this detailed modelling. One method for prioritisation is the Probabilistic Tsunami Hazard Assessment (PTHA) of Australia. In the PTHA, tsunamis were modelled from all likely earthquake sources across the deep ocean using a computationally faster linear solution and coarser model domain. Results are considered valid only to the 100 m depth contour, where we calculate return periods for tsunami wave height around Australia and generate a database of tsunami wave forms. However, tsunami waves are shaped considerably by the bathymetry between the 100 m depth contour and the coast, and tsunami behaviour near the coast is therefore highly non-linear, dependent on elevation, coastline shape, wave height, period and momentum. This non-linear reality of the near shore environment raises a number of questions. Is offshore wave height alone the best predictor of onshore tsunami hazard? Analytical solutions to the 1-D shallow water equations exist for predicting wave run-up on plane beaches. Can these be applied to offshore waves to measure inundation potential? Or are other metrics, such as wave energy, more appropriate? Comparisons with results from detailed inundation models will explore the utility of these measures for prioritising communitie

The aim of the present work is to determine to what extent event-specific tsunami amplitude forecasts from different numerical forecast systems differ, and therefore, how the related products from RTSPs might differ. This is done through comparing tsunami amplitudes for a number of different hypothetical tsunami events within the Indian Ocean, from a number of different tsunami scenario databases.