Evaluating AlphaEarth Foundations Embeddings for Wildfire Susceptibility Mapping

2026-08-12

AI Summary

This study addresses the limitations of traditional wildfire susceptibility mapping, which relies on complex multi-source data fusion and task-specific feature engineering, resulting in constrained generalizability. For the first time, it systematically evaluates the effectiveness of off-the-shelf geospatial embeddings from AlphaEarth Foundations (AEF) as a replacement for conventional physical variables in wildfire modeling. Using satellite-based fire observations and ROC-AUC as the evaluation metric, the proposed approach achieves an AUC exceeding 0.92 in Victoria, Australia. When transferred to the climatically similar Canberra region, it demonstrates a performance gain of approximately 4%, in stark contrast to traditional models, which suffer an average decline of 25%. These results validate the high representational fidelity and strong near-domain transfer potential of AEF embeddings.

Abstract

Wildfire susceptibility mapping typically relies on physical variables assembled from multiple remote-sensing, climate, and geospatial products. AlphaEarth Foundations (AEF) provides analysis-ready geospatial embeddings that may reduce this dependence on heavy harmonisation and task-specific feature engineering, but their value for wildfire susceptibility mapping has not been systematically evaluated. Using Victoria, Australia (2017-2025), as a case study, we show that AEF embeddings can reconstruct commonly used variables in wildfire susceptibility analysis with high accuracy. In downstream susceptibility models trained on satellite-derived fire occurrence data, embedding-based susceptibility models achieve ROC-AUC values above 0.92 and consistently identify high wildfire susceptibility across eastern Victoria, particularly Gippsland and the north-eastern uplands, with additional localized hotspots in central and northwestern Victoria. A key feature of AEF embeddings is their strong near-region transferability within climatically similar regions. When embedding-based models trained in Victoria are applied to Canberra and Western Sydney-Blue Mountains, ROC-AUC improves by around 4% at Canberra and declines by around 2% at Western Sydney-Blue Mountains, compared with a mean decrease of approximately 25% for physical-variable models. These findings provide practical guidance for using AEF embeddings and lay a foundation for scalable wildfire susceptibility mapping workflows for downstream users such as government agencies and (re)insurers.

Problem

• wildfire susceptibility mapping
• geospatial embeddings
• feature engineering
• transferability
• remote sensing

Innovation

• geospatial embeddings
• wildfire susceptibility mapping
• transferability
• feature engineering reduction
• AlphaEarth Foundations