Climate Vulnerability Assessment of Infrastructure Using Edge‑AI Integrated IoT Systems: A Survey

Abstract

Climate change is amplifying extremes that directly threaten critical infrastructure. Timely, spatially resolved vulnerability assessment is indispensable for adaptation planning and operational resilience. Cloud‑first analytics alone struggle with bandwidth, latency, privacy, and continuity constraints in fast‑evolving hazards. This survey synthesizes advances at the intersection of climate vulnerability assessment, internet‑of‑things (IoT) sensing, and edge artificial intelligence (edge‑AI). We ground the assessment problem in contemporary climate risk evidence and definitions, propose an end‑to‑end framework linking hazard–exposure–vulnerability constructs to IoT/edge data flows, and review methods spanning sensing architectures, communication standards, on‑device learning (TinyML, model compression, federated learning), spatio‑temporal learning over sensor networks, and digital‑twin integration. Representative deployments in flood monitoring, structural health monitoring, and wildfire detection illustrate how edge‑AI reduces detection latency, preserves operation under degraded connectivity, and improves data stewardship—capabilities aligned with the needs of climate adaptation and risk‑informed asset management [1]–[3], [9], [10].