Geospatial Foundation Models to Enable Progress on Sustainable Development Goals

2025-05-30

AI Summary

Geospatial foundation models (FMs) hold significant promise for Earth observation, yet their practical utility and alignment with the United Nations Sustainable Development Goals (SDGs) remain systematically unassessed. To address this gap, we introduce SustainFM—the first SDG-aligned benchmarking framework for geospatial FMs, covering all 17 SDGs and emphasizing critical tasks such as asset forecasting and environmental hazard detection. Our method establishes a multidimensional evaluation paradigm centered on transferability, cross-domain generalization, energy efficiency, and ethical robustness, shifting focus from “model-centric” to “impact-centric” assessment. Leveraging models including SatMAE and Earthformer, SustainFM integrates multi-source remote sensing and socioeconomic data, incorporating zero-shot transfer and carbon-aware inference. Experiments demonstrate that SustainFM substantially outperforms conventional benchmarks; identifies energy efficiency and robustness as pivotal metrics; and releases an open, reproducible benchmark suite—filling a critical void in SDG-aligned geospatial model evaluation.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Systems and Infrastructure for Web, Mobile and WoT: Sustainability and carbon-aware systems for Web, mobile, and WoT
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Foundation Models (FMs) are large-scale, pre-trained AI systems that have revolutionized natural language processing and computer vision, and are now advancing geospatial analysis and Earth Observation (EO). They promise improved generalization across tasks, scalability, and efficient adaptation with minimal labeled data. However, despite the rapid proliferation of geospatial FMs, their real-world utility and alignment with global sustainability goals remain underexplored. We introduce SustainFM, a comprehensive benchmarking framework grounded in the 17 Sustainable Development Goals with extremely diverse tasks ranging from asset wealth prediction to environmental hazard detection. This study provides a rigorous, interdisciplinary assessment of geospatial FMs and offers critical insights into their role in attaining sustainability goals. Our findings show: (1) While not universally superior, FMs often outperform traditional approaches across diverse tasks and datasets. (2) Evaluating FMs should go beyond accuracy to include transferability, generalization, and energy efficiency as key criteria for their responsible use. (3) FMs enable scalable, SDG-grounded solutions, offering broad utility for tackling complex sustainability challenges. Critically, we advocate for a paradigm shift from model-centric development to impact-driven deployment, and emphasize metrics such as energy efficiency, robustness to domain shifts, and ethical considerations.

Problem

• Assessing geospatial FMs' real-world utility for sustainability goals
• Evaluating FMs beyond accuracy for responsible sustainability applications
• Shifting from model-centric to impact-driven deployment of FMs

Innovation

• Geospatial Foundation Models for sustainability tasks
• Benchmarking framework with diverse SDG-aligned tasks
• Impact-driven deployment with energy efficiency metrics