Abstract <p>Although several prior reviews have addressed scientific visualization more broadly, none have specifically focused on interactive visualization within the MetOcean domain. As a result, many task- and system-level aspects relevant to MetOcean applications remain underexamined. This paper presents a taxonomy-guided review of interactive visualization approaches tailored for MetOcean data systems. We introduce a six-component taxonomy that integrates data characteristics, visualization objectives, user interaction, technological integration, visualization techniques, and design principles. Using this taxonomy, we conduct a dual analysis: a structured review of prior academic literature and a detailed mapping of nine validated MetOcean systems. The findings reveal consistent strengths in data representation and user interaction, but also highlight critical gaps in predictive analytics, real-time backend integration, user-centered design principles, and operational alignment with disaster response and recovery phases. The taxonomy enables a consistent and comparative synthesis across systems and helps surface underrepresented components, such as feature-based visualization, proactivity, transparency, and granularity. We conclude by outlining five open challenges and opportunities that offer a research roadmap for building next-generation, robust, and decision-aware MetOcean visualization platforms.</p> Graphical abstract <p></p>