Impact of Data Assimilation and Resolution on Modeling the Gulf Stream Pathway

In this paper, we study, in detail, one important aspect of our ongoing work on global ocean prediction. Modeling the behavior of western boundary currents, like the Gulf Stream, has been a long-standing issue. Recent modeling results suggest that the abyssal currents play an important role in determining the pathway of the Gulf Stream. The present-generation ocean prediction model fails to adequately reproduce the Gulf Stream pathway, and does not generate a vigorous abyssal circulation. Here we use twin simulations, at different horizontal resolutions and with and without data assimilation, to study their effects on modeling the observed Gulf Stream pathway. Increasing the resolution of the model improves the strength of the abyssal circulation, but still fails to predict the Gulf Stream pathway. Surprisingly, assimilating sea level anomalies and upper ocean profiles produces a robust abyssal circulation and a Gulf Stream pathway similar to the observed pathway. However, the Gulf Stream with assimilation is weaker than observed and, unlike most regions of the deep ocean, there is no skill in the 14-day forecasts. All simulations were performed under our current Department of Defense (DoD) High Performance Computing Modernization Program (HPCMP) Challenge Project, Eddy Resolving Global Data Assimilation Including Tides, or our earlier HPC Challenge Project, Global Ocean Prediction using HYCOM.