Abstract
Western Boundary Currents and the eddies they shed are high priorities for numerical estimation and forecasting due to their economic, ecological and dynamical importance. However, the rapid evolution, complex dynamics and baroclinic structure that is typical of eddies and the relatively sparse sampling in western boundary currents leads to significant challenges in understanding the 3-dimensional structure of these boundary currents and mesoscale eddies. Here, we use Observing System Simulation Experiments (OSSEs) to explore the impact of assimilating synthetic subsurface temperature observations at a range of temporal resolutions, to emulate expendable bathythermograph transects with different repeat frequencies (weekly to quarterly). We explore the improvement in the representation of mesoscale eddies and subsurface conditions in a dynamic western boundary current system, the East Australian Current, with a data-assimilating regional ocean model. A characterisation of the spatial and temporal ocean variability spectrum demonstrates the potential for undersampling and aliasing by a lower sampling frequency. We find that assimilating subsurface temperature data with at least a weekly repeat time best improves subsurface representation of this dynamic, eddy-rich region. However, systemic biases introduced by the data assimilation system hinder the ability of the model to produce more accurate subsurface representation with fortnightly or monthly sampling. Removal of this bias may improve subsurface representation in eddy-rich regions with fortnightly or even less frequent observations. These results highlight the value of both increased subsurface observation density in regions of dynamic oceanography as well as continued development of data assimilation techniques in order to optimise the impact of existing observations.