Data Science and Signal Processing for Drifter Data

Drifter deployments continue to be a popular observational method for understanding ocean currents and circulation, with numerous recent regional deployments, as well as the continued growth of the Global Drifter Program. Drifter data is however highly heterogenous, prone to measurement error, and captures an array of physical processes that are difficult to disentangle. Moreover, the data is “Lagrangian” in that each drifter moves through space and time, thus posing a unique statistical and physical modelling challenge. In this presentation we discuss a number of recent data science advances from time series analysis and signal processing which contribute to solving these challenges.

We start by overviewing some novel techniques for preprocessing and interpolating noisy GPS data using smoothing splines and non-Gaussian error structures. We then examine how the data can be uniquely visualised and interpreted using time-varying spectral densities. Finally we highlight some parametric stochastic models which separate physical processes such as diffusivity, inertial oscillations and tides from the background flow. Core to our methodology is computational efficiency and flexible implementation. We show how our suite of methodological tools can estimate parameters and test hypotheses from a range of physical models, and provide new ways of collating and visualising the rich information content of drifter data.

Back to: Data Science for Modern Oceanography: Statistics, Machine Learning, Visualization, and More II