From Phoenix Rapidfire to FireDST; probabilistic fire spread and impact — ASN Events

From Phoenix Rapidfire to FireDST; probabilistic fire spread and impact (#9)

Kevin Tolhurst 1 , Bob Cechet 2
  1. Department of Forest and Ecosystem Science, Melbourne University, Creswick, Victoria, Australia
  2. Bushfire CRC, Symonston, ACT, Australia

Over the last decade or so, researchers at the University of Melbourne and the Bushfire CRC have developed “Phoenix Rapidfire”, a computer program that predicts the direction, speed and intensity of bushfires. The software has been used as a key element for Victoria’s bushfire response over the last two fire seasons. As Phoenix Rapidfire predicts the behaviour of a fire, it can help determine which communities to warn, and where to allocate resources to minimise the impact.

Looking to the future, the Bushfire CRC has invested in the next generation computer simulation technology and is conducting a research effort to move beyond deterministic “best guess” predictions of fire spread (Phoenix Rapidfire) to a simulation system. As well as simulating a range of plausible scenarios for a particular event, such a system can also provide probabilistic information regarding both fire spread and risk in terms of impact to people and property. The development of the Fire Impact & Risk Evaluation Decision Support Tool (FireDST) is a multi-agency research effort which has provided a “proof of concept” simulation system that demonstrates technology that could be available to both planning and emergency managers in three to five years.

Probabilistic fire hazard, impact and risk products are in their infancy; however the research team has explored the feasibility and utility of such technology through three case studies. FireDST is a unique tool. Through the technology of parallel computing and ensemble forecasting, it is able to provide a large range of fire behaviour and impact simulations that consider the uncertainty within input parameters such as weather and the type and status of fuels. Through further consultation with the end user community and operational hardening of the simulation system, probabilistic forecasts of fire spread and impact could be a reality within the next five years.

#AFAC2013
Full Paper