A Fire Behaviour Modelling Platform

Developing Spark, a fire behaviour modelling platform and bridge from bushfire data acquisition and processing to operational use
A firefighter silhouetted against a bushfire
Who will benefit
Bushfire researchers and government agencies, who can now use the freely available web interface to test new types of data processing, analytics and models and test operationalisation

The Challenge

It is challenging to translate bushfire research into operational use. Key questions include:

  • How can new research be tested with existing wildfire models?
  • How can a new wildfire model be tested in an operational system?
  • How do models interact with each other?
  • How can we demonstrate predictive improvement?

An open, web-based sandbox for researchers to trial new data and models is needed.

The Approach

This project: 

  • identified the fuel attributes known to influence fire behaviour processes – these were key inputs to the fire behaviour platform and part of the validation database provided by the Aggregating and Harmonising Fuel Data on a National Scale project by Terrestrial Ecosystem Research Network (TERN) 
  • detailed the standard protocols to collect information on the key fuel attributes on the field and with remote sensing data
  • determined the suitability of the data sources identified by TERN in their project to represent those key attributes and validate remote sensing products.
  • developed a pipeline for the validation of remote sensing fuel attributes with on-ground data.

The Outcomes

The Spark fire behaviour modelling platform has been developed as a bridge from data acquisition, processing and modelling to operation and assessing bushfire risk. Spark is a proving and testing ground for new data, models and analytics. A collaborative, open environment for researchers, agencies and stakeholders, it is open for use by stakeholders for research as well as operational evaluation. 

Researchers affiliated with an AAF research institution can now access Spark using their institution’s credentials. Access Spark.

The project has created a model where there is a clearly articulated process for translating bushfire research into operational use from a predictive modelling perspective.

Who Will Benefit

This project helps researchers develop and evaluate new ember attack/spot fire models and fire suppression models for a range of fuel/vegetation types. Comparison with relevant fire history data enables evaluation of the models’ accuracy. 

This project also allows emergency agencies to provide feedback through Spark to researchers in relation to the effectiveness of the model for operational use.

View a demonstration of Spark by James Hilton of CSIRO, project lead for Spark, at the May 2023 Bushfire Data Challenges Forum.

The Partners

  • Australasian Fire and Emergency Service Authorities Council (AFAC)
  • SA Department of Environment and Water
  • NSW Rural Fire Services
  • Australian National University (ANU)
  • University of Queensland (UQ)
  • Terrestrial Ecosystem Research Network (TERN)
  • Victoria Department of Environment, Land, Water and Planning (Vic DELWP)

Contact the ARDC

  • This field is for validation purposes and should be left unchanged.