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

Approved

1. Relevance and Originality

This paper provides a timely and well scoped survey on the use of edge AI integrated IoT systems for assessing climate vulnerability of infrastructure assets. The relevance of the topic is very strong, particularly in the context of increasing climate extremes and the operational limitations of cloud centric analytics during hazard events. The originality lies in the paper’s systems level perspective, where climate risk theory, sensing technologies, edge computation, governance, and decision support are connected into a coherent end to end narrative. While the work does not introduce new algorithms, its integrative framing offers meaningful value to researchers and practitioners alike.

2. Methodology

The survey methodology is clearly bounded by explicit inclusion and exclusion criteria, which helps maintain focus and coherence throughout the manuscript. The paper systematically organizes literature across sensing, communication, edge learning, and digital twin integration. That said, the methodological rigor would benefit from additional clarity on how sources were selected within the stated period. For example, indicating whether the review followed a structured protocol or keyword driven search strategy would enhance reproducibility and transparency.

3. Validity and Reliability

The paper demonstrates a high level of reliability by grounding its discussion in authoritative assessments, standards documents, and peer reviewed studies. The use of IPCC definitions, NIST frameworks, and well established IoT and AI standards strengthens the trustworthiness of the conclusions. The manuscript also appropriately acknowledges uncertainty, non stationarity, and governance risks. A minor limitation is that performance outcomes reported across different studies are not always normalized, which may make direct comparison challenging for some readers.

4. Clarity and Structure

The manuscript is logically organized and easy to follow despite the breadth of topics covered. Each major section builds naturally on the previous one, moving from conceptual foundations to architectural considerations and real world applications. The use of figures, tables, and highlighted takeaways improves comprehension. However, some technical sections could be streamlined by reducing repetition across standards discussions and by grouping related protocols more concisely.

5. Results and Analysis

The analytical contribution of the paper lies in its synthesis of reported deployment outcomes rather than original experimentation. This synthesis is well executed, especially in highlighting reduced latency, improved resilience under degraded connectivity, and governance aligned deployment practices. The evaluation scaffold proposed in the later sections is particularly valuable and could serve as a reference framework for future empirical studies. Expanding the discussion on limitations of edge AI under extreme sensor sparsity or hardware failure scenarios would further balance the analysis.