Evaluating probabilistic forecasts for maritime engineering operations

Abstract Maritime engineering relies on model forecasts for many different processes, including meteorological and oceanographic forcings, structural responses, energy demands. Understanding the performance evaluation of such forecasting models is crucial in instilling reliability maritime operations. Evaluation metrics that assess point accuracy forecast (such as root-mean-squared error) are commonplace, but with increased uptake probabilistic methods may not consider full distribution. The statistical theory proper scoring rules provides a framework which to score compare competing forecasts, it seldom appealed applications. This translational paper presents underlying principles rules, develops simple panel be used robustly evaluate demonstrates this an application surface winds at asset Australia’s North–West Shelf. Where appropriate, we relate common requirements by industry. case study from body work was undertaken quantify value resulting operational product clear demonstration downstream impacts data science can have