Introduction

On December 26, 2004, a massive magnitude 9.0 earthquake hit just off the coast of north Sumatra, triggering a tsunami that sent waves up to 100 feet high into coastal areas throughout the Indian Ocean. More than a quarter million people in Indonesia and 13 other countries perished, in large part because they had little to no warning the tsunami was coming.

A tsunami, sometimes called a tidal wave, is a massive displacement of water usually caused by an undersea earthquake. The tsunami waves often travel across large water bodies wreaking tremendous damage on coastal areas when they make landfall.

An earthquake strong enough to potentially trigger a tsunami affecting Australia occurs on average about once a week. Within minutes of an earthquake striking, seismic data related to the quake is loaded into a Tsunami Decision Support Tool (DST) at the Joint Australian Tsunami Warning Centre (JATWC) hosted by the Australian Bureau of Meteorology (The Bureau) and Geoscience Australia (GA). The DST was developed in the IDL® programming language to assess whether the event poses a tsunami danger to Australia or the Indian Ocean region. If a tsunami is likely, the DST predicts the threat level, pinpoints a geographic target, and generates appropriate warnings.

The Bureau created the DST in the wake of the 2004 Indian Ocean tsunami, now considered the deadliest natural disaster in modern history. The Bureau selected the IDL programming language due to its advanced visualization and complex analytics capabilities. The NV5 Geospatial Custom Solutions team supported the original development and then completed a major upgrade to the DST in 2018.

“The DST is an end-to-end decision supporting system that starts with data collection to run the threat assessment models and then [it] creates and issues alerts,” said Yuelong Miao, the Bureau's National Manager for Tsunami Warning Services. “The beauty of IDL is that it’s not only processing data; it allows you to create a visualization interface.”

“IDL is a commercial off the shelf (COTS) product from NV5, that has hundreds of built-in formulas and algorithms written specifically for scientific data processing analysis and display,” said Bryan Justice, NV5 Program Manager.

Building the Tsunami Tool

As Australia’s national weather, climate and water agency, the Bureau is tasked with assisting its citizens in dealing with a variety of natural phenomena, including tsunamis. The Bureau and GA JATWC centers are staffed 24/7 as part of an international network of organizations monitoring earthquake and related events in the Indian Ocean and Pacific Rim region.

Of the dozens of data sets and observations streaming into the JATWC from its partner agencies, one of the most important is seismic measurements. Hundreds of seismographs in Australia and throughout the globe measure subsurface disturbances, determine their depths, and pinpoint their epicenters. The JATWC considers any undersea earthquake measuring greater than or equal to magnitude 6.5 to be a tsunami threat and inputs the data into the DST for further analysis.

The DST uses IDL to process the seismic readings and location data and matches these with pre-programmed tsunami wave model scenarios. In building the tool, the Bureau defined multiple tsunami threat levels based on expected wave impacts at the coastline. The threat assessment is the worst-case scenario for a given earthquake magnitude and the forecast threat levels are calibrated using historical tsunami impacts to Australia.

“We use the worst-case scenario rather than a percentage probability [that a tsunami will occur], just to be safe,” said Miao.

He explained that it makes more sense to alert the public to the most dangerous situation presented by a potential tsunami – even if one never comes – rather than lull them into a false sense of security.

Threat Scenarios Programmed into the Tool

Threat levels are split into ‘Marine’ and ‘Land’ categories. The Marine threat predictions relate to tsunamis that never inundate land beyond the immediate shoreline. But these may create currents, riptides and waves dangerous to swimmers and to small boats in offshore areas. The higher-level land threat, however, is a prediction of significant land inundation.

In addition to determining threat levels, the DST employs geospatial computer models created in IDL and built into the tool to calculate the geographic areas most at risk from the tsunami. The JATWC has divided the entire Australian and Indian Ocean coastlines into many warning zones of varying length. The models typically identify multiple zones in a broad area that are under tsunami threat at any given time.

Miao stresses that timeliness is a necessity for a tsunami warning. Residents of coastal areas may have less than 30 minutes to prepare for the waves that might come ashore once an earthquake occurs. JATWC personnel keep an eye on a DST-generated timer once a threat has been predicted, to help them meet their alert-issue deadlines. The DST also gives them the flexibility to continue adding earthquake and wave height information so that the threat levels can be refined as the event unfolds.

“For all undersea earthquakes of magnitude 6.5 or higher we are obliged to issue an alert irrespective of whether or not a tsunami has been detected,” said Miao.

Automated Warnings

The DST selects the likely scenario and impact zone within seconds of receiving the seismic inputs. The DST user can immediately begin issuing automatically-generated text warnings to individuals, partner organizations, news media, and emergency services in the areas that may be affected via email, text messages, and Twitter. The DST also taps into the geospatial visualization tools in IDL to create dynamic web maps showing where the tsunami is likely to form and where it might hit.

“The JATWC web pages are our true comprehensive source of all information that people need to get a full picture of what is happening in their state and anywhere in the Indian Ocean,” said Miao.

As part of the 2018 DST upgrade, NV5 leveraged IDL’s geospatial capabilities to enable the tool to generate maps of the impact zone as GIS layers that can be pushed directly into the GIS software environments of participating agencies and partners. This color-coded visual map constantly updates with easy-to-understand graphics as the DST changes its tsunami predictions based on new information.

The Bureau is already looking ahead to the next iteration of the DST. According to Miao, future editions could make use of real-time computer modeling which would allow the tool to become more dynamic. The tsunami threat level would be determined using on-demand computer models instead of pre-computed model scenarios.