Oil Spill Trajectories from HF Radars: Applied Dynamical Systems Methods vs. a Lagrangian Stochastic Model

Brian M Emery1,2, Libe Washburn1,3, Igor Mezic4, Sophie Loire4, Hassan Arbabi4, Carter Ohlmann5 and Jack Harlan6, (1)Marine Science Institute, University of California, Santa Barbara, Santa Barbara, CA, United States, (2)Department of Mechanical Engineering, University of California Santa Barbara, Santa Barbara, CA, United States, (3)Department of Geography, University of California Santa Barbara, (4)Department of Mechanical Engineering, University of California Santa Barbara, (5)University of California Santa Barbara, Earth Research Institute, Santa Barbara, CA, United States, (6)U.S. Integrated Ocean Observing System, NOAA, Washington DC, DC, United States
Abstract:
We apply several analysis methods to HF radar ocean surface current maps to investigate improvements in trajectory modeling. Results from a Lagrangian Stochastic Model (LSM) are compared with methods based on dynamical systems theory: hypergraphs and Koopman mode analysis. The LSM produces trajectories by integrating Eulerian fields from the HF radar, and accounts for sub-grid scale velocity variability by including a random component based on the Lagrangian decorrelation time. Hypergraphs also integrate the HF radar maps in time, showing areas of strain, strain-rotation, and mixing, by plotting the relative strengths of the eigenvalues of the gradient of the time-averaged Lagrangian velocity. Koopman mode analysis decomposes the velocity field into modes of variability, similarly to EOF or a Fourier analysis, though each Koopman mode varies in time with a distinct frequency. Each method simulates oil drift from a the oil spill of May, 2015 that occurred within the coverage area of the HF radars, in the Santa Barbara Channel near Refugio Beach, CA. Preliminary results indicate some skill in determining the transport of oil when compare to publicly available observations of oil in the Santa Barbara Channel. These simulations have not shown a connection between the Refugio spill site and oil observations in the Santa Monica Bay, near Los Angeles CA, though accumulation zones shown by the hypergraphs correlate in time and space with these observations. Improvements in the HF radar coverage and accuracy were observed during the spill by the deployment of an additional HF radar site near Gaviota, CA. Presently we are collecting observations of oil on beaches and in the ocean, determining the role of winds in the oil movement, and refining the methods. Some HF radar data is being post-processed to incorporate recent antenna calibrations for sites in Santa Monica Bay. We will evaluate effects of the newly processed data on analysis results.
Back to: New Insights in Coastal Oceanography from High-Frequency Radar Observations Posters