NSW National Parks and Wildlife Service is testing artificial intelligence (AI) in their fire management practices, addressing a critical issue identified in the NSW Koala Strategy.
Bushfires pose a major threat to koalas and without action, the species could face extinction by 2050.
Hazard reduction burns are controlled fires used to lower the risk of larger bushfires by reducing fuel such as dry vegetation.
The power of AI
AI allows National Parks to assess the fire severity potential during hazard reduction burns. This enhances the risk management and control efforts to safeguard koalas, their habitats, and nearby communities.
In Autumn 2024, National Parks applied AI in the Southern Highlands, which is home to over 3,000 koalas, combining AI predictions with field data including the locations of koala feed trees and habitats.
This initiative supported decision-making on ignition methods and burn patterns to specifically lower fire intensity and protect crucial koala habitats.
The ‘Random Forest Model’ was also used to inform on-the-ground decision-making, improving coordination among firefighters and emergency services. This model is a machine learning algorithm that was trained to evaluate data to predict fire severity.
Data sets include topography, satellite imagery, vegetation fuel types, aridity, weather forecasts from the Bureau of Meteorology and others.
Protecting wildlife and habitat
According to Erik Sumarkho, a project officer with National Parks, “AI has the potential to support our approach to hazard reduction burns in sensitive ecosystems. It will give us the tools to predict, plan, and communicate risks to arboreal mammals, as well as other environmental, cultural and property assets, with accuracy.”
Erik highlights how AI can quickly distil complex data to accurately forecast fire severity, including how far up into the canopy a fire might reach.
Extreme fires can consume the upper canopy where koalas live and feed, so predicting fire severity allows NSW National Parks and Wildlife Service to implement risk controls and mitigation strategies to protect these crucial habitats.
“The model might, for example, forecast that a particular area will experience high fire severity, prompting firefighters to remove excessive fuel loads before the burn begins,” says Erik.
Community benefits
Beyond protecting koalas, AI-driven fire management could benefit the wider community. By reducing the frequency and intensity of bushfires, these strategies lower risks to homes, businesses and infrastructure, enhancing public safety and community resilience.
The National Parks and Wildlife Service’s integration of AI represents an innovative approach to fire management, aiming to protect koalas and ensure community safety and resilience.
It offers a more reliable way to manage fire risks, helping to preserve koala habitats and secure their survival for the future.