Over the past decades, various operational drift forecast models were developed for trajectory prediction of objects lost at sea for search and rescue operations. Most of these models are now based on a stochastic, Monte Carlo definition of the object’s initial position and its time-evolving search area through computation of an ensemble of equally probable trajectories (Breivik [1]). Uncertainties in environmental forcing, mainly surface currents and wind, as well as the uncertainties inherent in the simplified computation of leeway speed and direction relative to the wind are also accounted for through this ensemble-based approach. Accuracy of the drift forecast obviously depends to a large extent on the quality of the environmental forecast data provided by numerical weather prediction models and ocean models, but it also depends on the level of uncertainty associated with the estimation of the drift properties (leeway) of the objects themselves. The present work mostly focuses on this second aspect of the problem. Drift properties of objects can be described by means of their downwind and crosswind leeway coefficients, according to the definition of leeway as stated by Allen [2, 3]. Assessment of the leeway coefficients is based on a direct method, which requires measurements acquired during field tests. Such field experiments basically entail deploying one or more objects at sea and simultaneously recording the environmental parameters (namely wind speed and motion of the object relative to the ambient water masses, i.e., its leeway) as well as the object’s position while adrift for periods ranging from several hours to several days. Using this method, a large database providing leeway coefficients for more than sixty object classes ranging from medical waste to a person-in-water to small fishing vessels was compiled over the years by the United States Coast Guard (Allen [2]). More recently additional trials were conducted, which allowed evaluation of new objects, including 20-ft shipping containers. We present in this paper the methods and analysis procedures for field determination of leeway coefficients of typical search-and-rescue objects. As an example we present the case study of a 20-ft container and discuss results obtained from a drift forecast model assessing sensitivity of such a model to the quality of environmental data as well as uncertainty levels of some reference parameters.

This content is only available via PDF.
You do not currently have access to this content.