The predictability of the ocean currents and the Congo River plume within the Angola Basin was investigated using the Regional Ocean Modelling System (ROMS) with data assimilation (4D-Var). Firstly, the impact of assimilating a novel remote-sensing data set, satellite-derived ocean currents (OSCAR) as compared to the more conventional satellite sea surface height (SSH) on ocean current predictability was assessed. In comparing 17 simulated and observed drifters throughout January-March 2013 using four different metrics, it was found that OSCAR assimilation only improves the Lagrangian predictability of ocean currents as much as altimetry assimilation. The impact of combining the aforementioned remote-sensing observations (OSCAR or SSH) with drifters was then investigated throughout the same period to assess whether this combination could improve upon assimilating the drifters alone on ocean current predictability. It was found that the addition of drifters significantly improves the Lagrangian predictability of the ocean currents in comparison to either altimetry or OSCAR as expected. More surprisingly, the assimilation of either SSH or OSCAR with the drifter velocities does not significantly improve the Lagrangian predictability compared to the drifter assimilation alone, even degrading predictability in some cases. Additionally, a new metric denoted the crossover time was formulated using the drifters, defined as the time it takes for a numerical model to equal the performance of persistence. In addition to ROMS, a global ocean model was also evaluated to demonstrate and quantify the metric fully. Finally, the impact of assimilating a recently available advanced version of a satellite salinity product (SMOS), on the Congo River plume was investigated. With some metrics specifically focusing on validating the Congo River plume, it was found that the assimilation of SMOS improved the representation of the plume within the model as well as the modelled salinity fields.