Spatially Tracking Wave Events in Partitioned Numerical Wave Model Outputs

Abstract

Numerical wave models can output partitioned wave parameters at each grid point using a spectral partitioning technique. Because these wave partitions are usually organized according to the magnitude of their wave energy without considering the coherence of wave parameters in space, it can be difficult to observe the spatial distributions of wave field features from these outputs. In this study, an approach for spatially tracking coherent wave events (which means a cluster of partitions originated from the same meteorological event) from partitioned numerical wave model outputs is presented to solve this problem. First, an efficient traverse algorithm applicable for different types of grids, termed breadth-first search, is employed to track wave events using the continuity of wave parameters. Second, to reduce the impact of the garden sprinkler effect on tracking, tracked wave events are merged if their boundary outlines and wave parameters on these boundaries are both in good agreement. Partitioned wave information from the Integrated Ocean Waves for Geophysical and other Applications dataset is used to test the performance of this spatial tracking approach. The test results indicate that this approach is able to capture the primary features of partitioned wave fields, demonstrating its potential for wave data analysis, model verification, and data assimilation.