Globally, the population is continuously increasing, as is the demand for food and energy. This growing need and limited possibilities for further exploiting on-land resources has led to an increased interest in the exploitation of the marine environment to fill the resource gap. In particular offshore areas are increasingly explored, despite challenges such as harsh environmental conditions characterised by strong variability and complex interactions between physical processes. While offshore sectors such as wind energy are already commercially established, others, such as offshore floating photovoltaic (FPV) systems and aquaculture are still in early stages of development. For both sectors, research has been conducted through small-scale field observations, laboratory experiments, or numerical modelling. However, these approaches often simplify or neglect statistical dependence among the environmental and structural variables, thereby hindering optimal design of offshore structures.

This dissertation explores the applicability of probabilistic models to explicitly capture such dependencies and to demonstrates how such models can enhance decision-making across different offshore technologies, with a particular focus on aquaculture and floating photovoltaic systems. Two copula-based multivariate modelling approaches with different complexity are applied. On the one hand, the Gaussian copula-based Bayesian Network (GCBN) offers a comparatively accessible and interpretable framework, as its graphical structure can be derived from the underlying physical processes. On the other hand, vine copula models are used to represent complex dependence patterns more flexibly. Unlike GCBNs, they are not restricted to a single copula family, instead, each pair of variables can be modelled using the copula that best captures their dependence. This flexibility, however, comes at the cost of higher model complexity and increased computational and practical challenges.... ... Read more