Detecting flow features in scarce trajectory data using networks derived from symbolic itineraries: an application to surface drifters in the North Atlantic

• Published in: Nonlinear Processes in Geophysics
• Main Authors: D. Wichmann, C. Kehl, H. A. Dijkstra, E. van Sebille
• Format: Article in Journal/Newspaper
• Language: English
• Published: Copernicus Publications 2020
• Subjects: Science; Q; Physics; QC1-999; Geophysics. Cosmic physics; QC801-809; North Atlantic
• Online Access: https://doi.org/10.5194/npg-27-501-2020; https://doaj.org/article/ce6a5163905b46a5ae76d4070ba9775f

The basin-wide surface transport of tracers such as heat, nutrients and plastic in the North Atlantic Ocean is organized into large-scale flow structures such as the Western Boundary Current and the Subtropical and Subpolar gyres. Being able to identify these features from drifter data is important for studying tracer dispersal but also for detecting changes in the large-scale surface flow due to climate change. We propose a new and conceptually simple method to detect groups of trajectories with similar dynamical behaviour from drifter data using network theory and normalized cut spectral clustering. Our network is constructed from conditional bin-drifter probability distributions and naturally handles drifter trajectories with data gaps and different lifetimes. The eigenvalue problem of the respective Laplacian can be replaced by a singular value decomposition of a related sparse data matrix. The construction of this matrix scales with O ( N M + N τ ) , where N is the number of particles, M the number of bins and τ the number of time steps. The concept behind our network construction is rooted in a particle's symbolic itinerary derived from its trajectory and a state space partition, which we incorporate in its most basic form by replacing a particle's itinerary by a probability distribution over symbols. We represent these distributions as the links of a bipartite graph, connecting particles and symbols. We apply our method to the periodically driven double-gyre flow and successfully identify well-known features. Exploiting the duality between particles and symbols defined by the bipartite graph, we demonstrate how a direct low-dimensional coarse definition of the clustering problem can still lead to relatively accurate results for the most dominant structures and resolve features down to scales much below the coarse graining scale. Our method also performs well in detecting structures with incomplete trajectory data, which we demonstrate for the double-gyre flow by randomly removing data points. We finally ...