Australian Bushfire Intelligence with AI-Driven Environmental Analytics

TL;DR

This study tackles the problem of predicting high-risk bushfire zones in Australia by integrating NASA FIRMS fire events, Meteostat daily weather, and NDVI from Google Earth Engine into a spatio-temporal framework. It systematically compares multiple ML models (XGBoost, Random Forest, MLP, LightGBM) and culminates in a two-class ensemble that prioritizes recall, achieving approximately $87\%$ accuracy and a ROC-AUC of $0.77$, outperforming three-class formulations. The work demonstrates the value of multi-source data fusion and ensemble learning for operational bushfire risk prediction, while highlighting practical constraints such as data alignment, class imbalance, and the need for additional indices. The replication package enables open reproduction and provides a baseline pipeline for extending the approach with drought indices, other vegetation metrics, and human activity data to further enhance decision-support in disaster management.

Abstract

Bushfires are among the most destructive natural hazards in Australia, causing significant ecological, economic, and social damage. Accurate prediction of bushfire intensity is therefore essential for effective disaster preparedness and response. This study examines the predictive capability of spatio-temporal environmental data for identifying high-risk bushfire zones across Australia. We integrated historical fire events from NASA FIRMS, daily meteorological observations from Meteostat, and vegetation indices such as the Normalized Difference Vegetation Index (NDVI) from Google Earth Engine for the period 2015-2023. After harmonizing the datasets using spatial and temporal joins, we evaluated several machine learning models, including Random Forest, XGBoost, LightGBM, a Multi-Layer Perceptron (MLP), and an ensemble classifier. Under a binary classification framework distinguishing 'low' and 'high' fire risk, the ensemble approach achieved an accuracy of 87%. The results demonstrate that combining multi-source environmental features with advanced machine learning techniques can produce reliable bushfire intensity predictions, supporting more informed and timely disaster management.

Figures (5)

• Figure 1: Research Methodology.
• Figure 2: Monthly Bushfire Point Counts (2015–2023).
• Figure 3: Fire Radiative Power (FRP) vs Daily Precipitation plot (2015-2023).
• Figure 4: Correlation Heatmap of the Two-class Classification Technique with Ensemble Model.
• Figure 5: ROC-AUC Curve for the Final Two-class Ensemble Model.