Investigating Data Assimilation Methodologies for Improved Swell Prediction Accuracy

Intense westerly winds blowing across the Southern Ocean over almost unlimited fetches generate long, fast-moving swell waves that propagate into the Indian, Pacific, and Atlantic Oceans. Despite the significant developments in numerical wave models over the past four decades, swell prediction remains impaired by biases in height and arrival time. These biases are associated with limited understanding of physical processes active in swell propagation and dissipation, and their incorporation in numerical models, as well as inaccuracies in hydrodynamic and atmospheric forcings. However, the increasing number and type of wave observations in the ocean coupled with new analysis techniques create avenues for enhancing wave prediction models. Here, we explore the use of data assimilation methodologies to improve swell prediction on Australia’s Northwest Shelf. Observations are utilised spanning the Southern Ocean (i.e., swell generation zone), the Indian Ocean (i.e., swell propagation path), and the Northwest Shelf. Observational data includes measurements from a large network of drifting buoys and the newly launched China France Oceanography Satellite (CFOSAT). Buoy and satellite data will be assimilated into the WaveWatch-III wave model in various test cases using different data assimilation frameworks over selected swell events. The model’s performance will be evaluated for each scenario and the role of data assimilation in improved swell prediction accuracy will be discussed in detail.