Mt Bundey training area fire simulations
Introduction
This project was developed by the Darwin Center for Bushfires Research (DCBR) with support from the Defence Support Group (DHCB) to assist in fire management planning and communication on the defence estate. The primary purpose of this project was to develop fire spread simulation models and visualisation tools to support thinking about fire management operations. The simulation models provide a virtual lab to explore fire management scenarios and to facilitate discussion between land managers and range users. This project ran through the 2019-2020 fincial year. Outputs from this work are described below.
The way in which fires move can be very complex and hard to predict particularly in the vast tropical savannas of northern Australia. ‘Incendiary’ is a fire spread simulation that allows us to explore the effects of a range of weather and landscape variables on fire behaviour which can assist thinking about strategic fire management. Of particular interest in Northern Australia is the application of early dry season mitigation burns to reduce fuel loads and thus reduce the impact of hot late dry season wildfires.
Project outputs
Three products where produced for this project:
-
A large 3D printed landscape of Mt Bundey training area as a base for projecting landscape information and simulations.
-
A fire propagation simulation model illustrating the key landscape and fire weather variables the effect fire spread.
-
A 'rate of spread' simulation using predictive modeling algorithms and real-time weather feeds. This was developed as an addtional experimental proof-of-concept application.
3D Landscapes
For participatory planning and training activities, the simulation models are designed to be projected onto 3D printed landscapes. This 3D terrain surface helps in understanding topographic and vegetation effects on fire spread. This technique, developed by DCBR, has been used extensively to support fire management palling across northern Australia (Fisher, 2019).
Images of the Mt Bundey projection augmented landscape is shown to the right:
This technique is described in more detail in this video:
Fisher, R., Heckbert, S., Villalobos, J. M. L., & Sutton, S. (2019). Augmenting physical 3D models with projected information to support environmental knowledge exchange. Applied Geography, 112, 102095.
Fire propagation simulation
The model currently uses the following variables to determine fire spread:
-
a grass vegetation map derived from Landsat Satellite imagery and a 'time since burnt' layer (from NAFI) to produce a fuel load variable. The vegetation layer shows grass, low cover and mangrove types.
-
an elevation layer (SRTM-DEM) is used to determine slope in relation to fire spread direction.
-
a topographic wetness layer, derived from the DEM, is used to represent differential landscape curing.
-
Fire danger as a value from 1 (wet season) to 10 (late dry season). This combines the influence of curing and temperature on fire spread.
-
Wind speed from none (no wind influence) to strong. Wind speed increases the directionality and likelihood of a pixel igniting.
-
wind direction (the direction a fire will spread)
To Use the model:
Click the drop incendiaries button and use the cursor to ignite some initial pixels. Change curing, wind direction and wind speed to set your fire scenario. Use the variable-wind button to allow the model to randomly change the wind speed as the model runs. Use the View drop list to display one of a range of landscape layers.
The video below shows and example of the model being run.
Fire rate of spread model
These simulations were based on formulas that estimate the Grassland Fire Danger Index (GFDI) and the rate of spread in km/h. The GFDI equation was derived by Noble et al. (1980) from the Mk5 slide rule. The model implements the equations (Eq. 3,4,5 and 6) of the CSIRO Grassland Fire Spread Model of Cheney et al. (1998). It’s a quasi-empirical model for predicting the rate of spread of grassland fires in undisturbed and cut/grazed pastures.
The rate of spread is combined with a Netlogo based cost distance analysis model (Fisher 2018) to produce a spread ellipse from a point of ignition.
The model also incorporates a live feed of weather data from the Mt Bundy BOM station that can be used to calculate the current rate of spread and run the model.
Cheney, N. P., et al. (1998). "Prediction of Fire Spread in Grasslands." International Journal of Wildland Fire 8(1): 1-13.
Noble, I., et al. (1980). "McArthur's fire‐danger meters expressed as equations." Australian journal of ecology 5(2): 201-203.
Prediction of grassfire rate of spread according to the CSIRO Grassland Fire Spread Meter (Cheney et al. 1998)
Fisher, R., Heckbert, S., Villalobos, J. M. L., & Sutton, S. (2019). Augmenting physical 3D models with projected information to support environmental knowledge exchange. Applied Geography, 112, 102095.
Modeling Software
These models were created in Netlogo modeling and simulation software. This is free open-source
software that can be downloaded and installed on any computer to run the models described.
Spread simulation model downloads
Download the zipped models here:
For more information about this project contact: Rohan.fisher@cdu.edu.au
