Presentation at ANZ Disaster & Emergency Management Conference 2024
Dhirendra Singh presented the talk Predictive hazard and community behaviour modelling for disaster preparedness and resilience by authors Dhirendra Singh, Vincent Lemiale, Leorey Marquez, William Swedosh, Raymond Cohen from Data61’s Natural Systems Modelling Group at the ANZ Disaster & Emergency Management Conference in Gold Coast todayn. The talk’s abstract is below.
The use of computer simulations for forecasting behaviour of natural hazards like bushfires, floods, and cyclones, is well established within predictive functions of emergency agencies in Australia. These simulations can amongst other things help quantify community risk through the analysis of the geospatial intersection of hazard spread with exposed critical infrastructure (e.g., energy, transport, health) and populations (e.g., residential dwelling, vulnerable aged-care facilities). However, a simplifying assumption in this method is that the impacted population is motionless for the duration of the event. This makes this kind of risk assessment unsuitable for certain disaster preparation needs such as evacuation planning.
To solve this problem, the CSIRO has been building simulation technology that can model community behaviour down to the individual level, capturing the diversity of those impacted, where they might be at the time of impact, and what they might do in response to the threat given their circumstance and situation at the time the threat is recognised. By simulating individuals’ decisions and how their choices play out on the roads, in tandem with as well as in response to the approaching hazard, these predictive models can offer new detailed insights into localised congestion and emergency egress issues.
This presentation will give an overview of the simulation technologies that are being developed within the Natural Systems Modelling group in CSIRO’s Data61. These include SPARK for simulating bushfires, SWIFT for simulating floods, SAFER for community risk assessment against millions of bushfire ignitions, SEEKER for vehicle level simulation of community movements, and PiXIE for pedestrian and crowd movements. Example use cases will show how these simulation tools are helping in efforts to reduce disaster risk and improve community resilience.
Key Learnings to take away from the presentation:
- Simulation technology can represent community behaviour in emergencies down to the individual, capturing who is impacted, what they are doing, and what they will do in response to the threat.
- By simulating individuals’ decisions and how their choices play out, in tandem with as well as in response to hazard behaviour, these predictive models can offer new detailed insights into localised congestion and egress issues.
- CSIRO’s technology is helping emergency services and communities understand hazard behaviour (via SPARK, SWIFT) alongside community behaviour (via SAFER, SEEKER, PiXIE) towards reducing disaster risk and improving resilience.