We present the advantages and challenges of the recently developed Lagrangian data assimilation (LaDA) method for forecasting the ocean coherent structures. Much data in the ocean is Lagrangian in nature, coming from ocean tracers such as surface drifters or sub-surface floats, or even satellite images of the ocean surface, such as color maps. Because key aspects of ocean phenomena are often best viewed in a Lagrangian light, such data can potentially play a major role in the ocean forecast.

The vital idea that differentiates LaDA from other approaches is that the position data is directly assimilated into the Eulerian model, rather than though a derived velocity approximation. Hence, LaDA can ncorporate Lagrangian information without sacrifice. The challenge of LaDA, however, is the filter divergence that can arise when the Lagrangian trajectories separate exponentially near the pivotal hyperbolic trajectory of the underlying ocean flow. Using Lagrangian analysis of the ocean flow dynamics, we present a basic concept how to overturn this challenge into an optimal deployment strategy for observing system design.