Observing System Design for Lagrangian Data Assimilation:
Targeting Lagrangian Coherent Structures
Abstract:
The use of Lagrangian floats/ drifters in the ocean has become more widespread
due to their ability to track coherent flow features (e.g.\ meddies) and
their relevance in understanding mesoscale transport. More recently, it
has become apparent that such data can be valuable in the forecasting of
the ocean. To realize this goal, we have developed a method for
assimilating Lagrangian data into a shallow-water equation ocean model.
The method employs an augmented state vector that includes Lagrangian
drifter coordinates and uses an Ensemble Kalman Filter to evolve the
error correlations between the drifters and the flow in our model.
We have shown that the method can track the `true' system in our twin
experiment configuration provided the assimilation time interval is of the
order of the Lagrangian integral time-scale. The success of the method is
intrinsically tied to the launch sites chosen for the drifters. Given
this, a directed launch strategy is needed for the methodology to be
useful in operational forecasts. In this work, we employ a dynamical
systems approach to extract Lagrangian coherent structures of the flow. We
show that by targeting the Lagrangian centers residing within the energy
dominated eddies, and the Lagrangian hyperbolic saddle points, an optimum
drifter deployment is realized for the performance of the method.