Stochastic trajectory predictions from ADCP velocity measurements

Stochastic modeling is commonly applied for analysis and uncertainty quantification of Lagrangian tracking. Among a wide range of applications, a few examples include tracking of surface drifters, pollutants, or identification of coherent structures. In this work, we employ a Lagrangian stochastic model (LSM) to predict drifter trajectories based on limited, stochastic ocean velocity measurements from an acoustic Doppler current profiler. This study is motivated by a recent field experiment in which column profiles of ocean velocity data were assimilated in real time by shipboard ADCP, and stochasticity of the Eulerian data originates from measurement errors. From this sparse and stochastic data, we applied a random flight model with first order Markov process that is commonly used in turbulent flows. The model parameters were tuned to satisfy a tensorial and coordinate invariant Lagrangian integral time scale obtained by autocorrelation of an ergodic process. The stochastic model prediction of Lagrangian trajectories was then compared with actual trajectories of released near-surface and sub-surface drifters and dye plumes that were tracked as part of the field experiment. We demonstrate that despite the sparsity and stochasticity of the ADCP data, there is generally good agreement between the stochastic model prediction and the trajectory observations measured using drifters and dye plumes.