Development of a Climate Risk Index for Australia Using Multi-Source Remote Sensing Data and Weighted Composite Analysis

Climate change posed significant challenges to Australia, characterized by diverse ecosystems and recurrent hazards such as bushfires, droughts, and heatwaves, yet a comprehensive spatially explicit risk assessment framework remained lacking. The primary objective of this study was to develop a Climate Risk Index (CRI) to identify and prioritize high-risk areas across Australia using multi-source remote sensing data. Data were gathered from 2017 to 2024, processed via Google Earth Engine, by merging 36 sub-region GeoTIFFs incorporating Palmer Drought Severity Index, wildfire burn fraction, land surface temperature, Normalized Difference Vegetation Index, and population density across 4040 non-null population pixels. Four weighting approaches—Random Forest, combined Random Forest-manual, manual, and Principal Component Analysis—were applied, with the Random Forest method achieving a testing accuracy of 0.983 and identifying 2022 high-risk pixels (CRI>0.75). Contrary to expectations of southeastern dominance, the main finding revealed a concentration of high-risk pixels mainly in central and western Australia, driven by aridity and thermal stress. This suggested the Random Forest-weighted CRI’s adaptability to diverse climatic hazards. The most important implication was the urgent need for targeted drought and heatwave mitigation in these regions, supported by the framework’s potential for real-time updates to guide resilient policy as climate patterns evolve.