Our people

The effect of fuel moisture content on the spread rate of bushfire in the presence of wind and slope.

Commenced: Feb. 2022 expected completion date Feb. 2026.

Understanding the interactive effect of live and dead fuel moisture content (FMC), wind, and terrain slope is a key challenge in wildfire science. This research employs the techniques of Computational Fluid Dynamics (CFD) to model flame propagation and bushfire spread, using a full three-dimensional simulation, with a Large Eddy Simulation (LES) turbulence model, combining the effect of fuel moisture content, wind-fire interaction, terrain slope, and radiative heat transfer. The project will improve the knowledge and ability to alleviate environmental problems and will contribute to the better management of bushfires and mitigate their hazards in real inhomogeneous landscapes that contain a mixture of live and dead fuels.

Understanding the geographical factors influencing pyroCb development and risk.

Commencement date: May 2021 Expected completion date: May 2025

Extreme wildfires can produce pyrocumulonimbus (pyroCb) storms, which impact not only the surface but also extend their influence high into the atmosphere. The overarching aim of this research project is to provide insights into the geographical patterns of pyroCb occurrence and practical guidance for wildfire management, now and into the future. The main research objectives are: (1) to better understand the drivers of pyroCb development and their impact mechanisms, to quantify the relative contributions of these drivers, and to investigate the geographical differences in these drivers and their impacts at the large scale; (2) to quantify pyroCb risk and develop models to predict the longer-term probability of pyroCb occurrence; (3) to understand the temporal and spatial patterns of pyroCb occurrences in Australia, investigate how these patterns might change under future climate projections, and to map the longer-term risks of pyroCb in Australia. This work forms part of an ongoing international collaboration on pyroCb research.

Stochastic modelling of bushfire spread

Commenced: August 2020, expected completion date August 2024

This project addresses the important problem of incorporating environmental uncertainty in bushfire propagation modelling.  To faithfully capture a range of uncertainties related to environmental inputs (in particular wind velocity) in bushfire propagation models, it is important to incorporate the stochastic nature of these inputs. In this research project, wind is modelled using stochastic processes in an attempt to achieve more reliable bushfire spread predictions compared to other probabilistic approaches. Two-dimensional bushfire propagation can be modelled using the level set method, and environmental uncertainty can be incorporated either directly into the bushfire spread model or by extending the propagation model to a stochastic level set approach. Stochastic bushfire simulations are compared to experimental fire spread data collected as part of broader international collaboration.

A comprehensive risk assessment of bushfire at wildland-urban interfaces in Australia

Commenced: May 2023, expected completion: November 2026

Bushfires pose significant threats to areas where settlements exist near wildland vegetation, known as wildland-urban interface or WUI. The WUI regions are a prime focus in bushfire management since they are a major site of ignition. The fires generated in the WUI have the potential to cause economic, social, and environmental damage as well as loss of lives. This project aims to conduct an integrated risk assessment of bushfires at WUI in Australia through spatially characterising and mapping the WUI across Australia and using fire simulation and demographics in WUI areas to quantify the risks. This study will facilitate in minimizing the existing challenges met by authorities in executing protective measures, and aid in planning optimal management approaches.

Simulation of ember storms

Commenced: February 2022, expected completion February 2026

Embers have been identified as the leading cause of house loss in previous fire events. Near the wildland–urban interface (WUI), areas where structures and developments meet wildland, embers pose a greater risk as small ember particles can flow over roads and terrain into properties, increasing the likelihood of structure damage. Although recent computational works have provided a significant insight into ember transport, the current models do not capture the near-ground behaviour and transportation of embers, and the entrainment (the process through which near-ground particles are introduced into the flow) of embers from the ground, both of which are key factors in ember storms. This project aims to improve current ember modelling to capture near-ground entrainment and relofting by analysing and adapting existing particle transport models and applying them to model ember storms using large eddy simulation (LES) in Nek5000, a popular spectral element solver. The implementation of these models will help develop more robust simulation techniques. This will ensure that authorities are better equipped to deal with wildfires at the WUI and formulate better mitigation strategies.

Investigation of dynamic fire behaviour models in an operational wildfire simulator

Commenced: November 2022 expected completion date: August 2029

The aim of my project is to simplify complex wildfire phenomena such as fire crossing a disruption (like a road or river) and the inception and propagation of crown fires and implement simplified models in a faster than real time two-dimensional simulator so that they could be used operationally (such as in Spark). This project will involve a review of existing one-dimensional models and complex 3D CFD models as well as conducting physical modelling using FDS (Fire Dynamics Simulator). The fundamentals of how fire spreads across the landscape in two dimensions will be investigated as a common theme between these phenomena.

Research Topic: Machine Learning-Aided Engineering Structural Safety Assessment and Design Optimization under Bushfires

 Start Date: 01/03/202 - Expected Completion Date: 01/03/2026

This research covers machine learning-aided structural safety assessment, uncertainty quantification, and structural design optimization for engineering infrastructures under bushfire attack. He and his team have been developing advanced structural safety assessment methods with the consideration of bushfires, and exploring the potential for engineering structures to improve the resistance of bushfires. 

Meteorological drivers of pyrocumulonimbus in Australia

Start date: September 2021 Expected completion date: September 2025.

This project aims to investigate specific meteorological factors responsible for the formation of pyrocumulonimbus (pyroCb/ firestorms) associated with bushfires in Australia. The presence of a pyroCb is an indication that a bushfire has become coupled with the atmosphere, and the resulting feedbacks between the fire at the surface and the storm above may lead to unpredictable and extremely dangerous conditions (both in terms of fire behaviour and storm behaviour) for anyone nearby.

Previous research has shown that meteorological factors of pyroCb formation may include—but are likely not limited to—the overspreading of a conditionally unstable airmass over ongoing fires with surface conditions favorable for extreme fire behavior, cold front and pre-frontal trough interactions with intensely burning fires, thunderstorm outflow boundary and sea breeze interactions with ongoing fires, and gravity wave interactions with ongoing fires. Nearly all previous research has focused on fires which resulted in pyroCb formation, while very little research has been conducted to differentiate fires which resulted in pyroCb formation from fires which did not.

This project will use the Weather Research and Forecasting-Fire model (WRF-Fire) to examine specific instances of pyroCb formation and non-formation in Australia. In particular, we are interested in events where one bushfire resulted in pyroCb formation, while another nearby fire did not. We aim to discover what differences (meteorological or otherwise) existed at the given locations of the fires and how that may have induced or inhibited pyroCb formation.

Time-series diffusion in wildfire simulation

Commenced: June 2023, expected completion date Jan 2027

 Diffusion based time-series generation applied in wildfire simulation. The research aims to develop new approaches to generate time-series data as wildfire propagation representation. Diffusion models are popular in generating high-resolution images with variations based on given conditions. I will explore the potential in diffusion models to generate time-series fire mask by learning satellite images. The model could be a alternative approach to simulate wildfire because diffusion models have inherent abilities to emulate resembling method for time-series forecasting by generate high-resolution data with variability.