|
Abstract:
|
Data collected from satellite-tracked floating drifters is pivotal to understanding our oceans. Drifter data provides better understanding of ocean currents and circulation, as well as the spread of objects in the ocean such as plastics, oil, and biological species such as plankton. In this talk I will highlight a number of recent advances in spatiotemporal methodology for handling what are very complicated but beautiful data structures.
Drifter data is particularly challenging because the data moves in time and space, referred to by Physicists as a "Lagrangian" perspective. We employ a number of techniques including splines, Markov transition matrices, windowed Fourier transforms, and fractional stochastic processes. We show how to answer questions such as: How do we best interpolate the raw noisy satellite data? What is the expected travel time and path between two arbitrary points in the ocean? Or what is the smoothness of a drifter's path, and what does this tell us about how fast particles spread in the ocean?
|
