Home » Publications » TsuDat: GeoNode and Disaster Modelling

When tsunamis struck Indonesia and Japan it drove home the need to plan and model the risks of catastrophic coastal inundation for every coastal nation. In Australia, surrounded by a ring of active faults, there has long been expertise in tsunami effects modelling. Unfortunately, there were far more Australian sites requiring tsunami expertise than there were tsunami experts available to do the modelling!

The solution hinged on the ability to distribute the knowledge of the tsunami experts to local planners. The first iteration of that effort was TsuDAT 1.0, desktop software that allowed flood model experts to access pre-generated tsunami model data directly. This increased access to the wider population of flood modellers, but still required expert flood modelling input. What was truly needed was a way for non-modellers—emergency managers, planning officials, land-use staff—to directly create detailed tsunami inundation hazard maps for coastal communities.

In partnership with GeoScience Australia and AIFDR, OpenGeo has delivered TsuDAT 2.0, a web-based system that links tsunami model data with local elevation data and inundation models. The new system is accessible to planners with just a few hours training.

Nick Horspool, of Geoscience Australia's Risk and Impact Analysis Group, worked on the TsuDAT project team. “There is no doubt that Tsudat will be a vital tool to help countries address the risk of tsunamis”, says Horspool, “TsuDAT will help identify communities at risk and develop tsunami evacuation plans and preparedness based on the results.”

Using the new TsuDAT system, anyone can create a tsunami hazard map with just a few hours training. Any kind of event can be simulated: from a small event with a 50% probability over 10 years to a large event with a 10% probability over 1000 years.

To create a hazard map using the TsuDAT simulation client, a local planner will:

1. Create a tsunami scenario. This selects likelihood of a tsunami, and thus it's size (small and frequent, or big and rare).
2. Outline the inundation simulation area and upload custom elevation data. Municipalities often have very fine engineering-grade elevation information which is good for flood modelling.
3. Set the simulation parameters, such as the tide level (a wave at high tide will cause more flooding than at low tide).
4. Save the simulation parameters for future tinkering, and run the simulation.

The simulation is then passed to a compute queue, and run. The planner receives an e-mail when the simulation is complete, then returns to the site to view the hazard map.

TsuDat benefited greatly by leveraging Anuga, originally developed and deployed in response to the tsunami, for it's computations and hydrodynamic modelling. In order to allow TsuDAT to grow and service the needs of potentially hundreds of computer-intensive model runners, OpenGeo developed a systems architecture that leverages the GeoNode data sharing system, and the HPC and data-intensive cloud computing services provided by Australia's peak High performance computing organization, NCI - National Computational Infrastructure - based at the Australian National University (ANU).

Building on GeoNode allowed the team to focus on the specific TsuDAT simulation application. All the map sharing, user management, and map visualization functionality needed for TsuDAT was already available in GeoNode.

By leveraging the NCI cloud, multiple organizations can easily access the service, or stand up their own versions without the hassle of acquiring and managing computer infrastructure. As simulation demand grows, more virtual machines can be provisioned to the TsuDAT application cluster at NCI and released when demand wanes. The TsuDAT application is one of the first users of the newly provisioned OpenStack cloud at NCI.

“We're really proud of the work we did with the TsuDAT team.” said Jeff Johnson, OpenGeo's team lead, “It feels great to be a part of a project that is going to make a big difference for improving public safety throughout the Pacific over the next several years.”

Other Case Studies