Hybrid and Multimodal Underwater Communication Architectures for Subsea Operations: A Survey and Design Perspective

Reliable underwater communication is essential for marine sensing, offshore resource exploration, subsea infrastructure monitoring, and autonomous underwater vehicle operations. However, underwater wireless networks are constrained by signal attenuation, variable propagation conditions, limited bandwidth, restricted energy availability, node mobility, and intermittent connectivity. As a result, single-modality systems are often unable to meet the diverse requirements of emerging Internet of Underwater Things and underwater ad hoc sensor-network applications. Underwater acoustic communication provides long-range connectivity but is limited by low data rates and high latency. By contrast, underwater optical, electromagnetic, magnetic-induction, and fibre-optic links can offer complementary advantages in terms of data rate, latency, energy efficiency, reliability, and deployment suitability under specific environmental conditions. This paper presents a structured survey and design-oriented synthesis of hybrid and multimodal underwater communication architectures. It examines recent advances in acoustic, optical, electromagnetic, magnetic-induction, and fibre-optic communication, and considers how these modalities can be integrated across the physical, datalink, network, and system levels. Attention is given to multimodal architecture selection, cross-layer protocol design, ad hoc and IoUT network constraints, readiness assessment, and enabling technologies such as software-defined networking, edge computing, and AI-assisted channel-aware adaptation. In contrast to surveys that mainly address individual underwater communication technologies, this work compares multimodal design choices in relation to range, throughput, latency, reliability, energy efficiency, environmental sensitivity, and deployment maturity. The analysis identifies practical modality combinations for representative underwater networking scenarios and highlights open research challenges, including field validation, standardisation, interoperability, adaptive wavelength selection, distributed modality selection, and scalable multimodal routing.