Learning latent dynamics for partially observed chaotic systems.

This paper addresses the data-driven identification of latent representations of partially observed dynamical systems, i.e., dynamical systems for which some components are never observed, with an emphasis on forecasting applications and long-term asymptotic patterns. Whereas state-of-the-art data-driven approaches rely in general on delay embeddings and linear decompositions of the underlying operators, we introduce a framework based on the data-driven identification of an augmented state-space model using a neural-network-based representation. For a given training dataset, it amounts to jointly reconstructing the latent states and learning an ordinary differential equation representation in this space. Through numerical experiments, we demonstrate the relevance of the proposed framework with respect to state-of-the-art approaches in terms of short-term forecasting errors and long-term behavior. We further discuss how the proposed framework relates to the Koopman operator theory and Takens' embedding theorem.

Rafting behaviour of seabirds as a proxy to describe surface ocean currents in the Balearic Sea.

Spatio-temporal variability of surface geostrophic mesoscale currents in the Balearic Sea (western Mediterranean) is characterized from satellite altimetry in combination with in-situ velocity measurements collected, among others, by drifting buoys, gliders and high-frequency radar. Here, we explore the use of tracking data from living organisms in the Balearic Sea as an alternative way to acquire in-situ velocity measurements. Specifically, we use GPS-tracks of resting Scopoli's shearwaters Calonectris diomedea, that act as passive drifters, and compare them with satellite-derived velocity patterns. Results suggest that animal-borne GPS data can be used to identify rafting behaviour outside of the breeding colonies and, furthermore, as a proxy to describe local sea surface currents. Four rafting patterns were identified according to the prevailing driving forces responsible for the observed trajectories. We find that 76% of the bird trajectories are associated with the combined effects of slippage and Ekman drift and/or surface drag; 59% are directly driven by the sea surface currents. Shearwaters are therefore likely to be passively transported by these driving forces while resting. The tracks are generally consistent with the mesoscale features observed in satellite data and identified with eddy-tracking software.

Diel vertical migration of European hake Merluccius merluccius and associated temperature histories: insights from a pilot data-storage tagging (DST) experiment.

A pioneering experiment of archival tagging of European hake Merluccius merluccius provided evidence of a diel vertical migration pattern which was analysed using an automatic method, developed and validated through time-frequency analyses. Frequent vertical movements across the thermocline were observed and fish experienced rapid temperature changes >7° C. These tagging-recapture data also suggested a homing behaviour to feeding grounds.