Data analytics and machine learning enhance the capacity of the DST to provide accurate predictions, insights, and adaptive strategies, which are crucial for effective decision-making. As a first element of the proposed DST, metamodels are developed to estimate the accidental oil spill size. Then, a model for ranking of response technologies is developed. These two elements together achieve RO-1 shown in Figure 2.

Oil spill response planning involves a combination of technology selection, resource allocation, and facility location. A deterministic optimization model allows for a holistic approach, optimizing multiple conflicting objectives simultaneously to achieve the best overall outcome. Therefore, as the next element of the proposed DST, an optimization model of facility location and resource allocation of response recovery resources is developed, involving the effectiveness of mechanical recovery, dispersant use and in-situ burning techniques from PII. The spill size, estimated from PI, sets the demand for resources (a critical input parameter) for the optimization model in PIII and PIV. Finally, the deterministic optimization model is extended to incorporate uncertainty in the parameter estimation of several key parameters and to quantify them.

A graphical abstract of the DST is shown in Figure 3 as a visual aid to understand how the DST development is organized.