Abstract's details
Pathways, impacts and fate of marine debris generated by the 2011 tsunami in Japan, derived from a synthesis of numerical models and observational reports
CoAuthors
Event: 2018 Ocean Surface Topography Science Team Meeting
Session: Application development for Operations
Presentation type: Type Oral
Contribution: PDF file
Abstract:
March 11, 2011 tsunami devastated the east coast of Japan and produced millions of tons of marine debris that drifted across the North Pacific. The extraordinary amount and unusual composition of tsunami debris allowed to trace its drift across the ocean and arrivals on remote shores and helped to better understand the pathways of floating marine debris. A suite of ocean models synthesized with these observations produced most complete picture of the debris dynamics, pathways, and fate. While observations allowed to optimize such model parameters as windage and source distribution, the models filled large gaps in sparse observations and produced estimates of the total budgets. For example, the study suggests that the original number of boats lost to the tsunami was about 1,000 and about 100 of these boats may be still floating in the ocean.
Detailed analysis of model fluxes on the US/Canada West Coast and their comparison with reports of tsunami boats from the same region revealed serious difficulties that even best OGCMs (Ocean General Circulation Models), such as the HYCOM (Hybrid Coordinate Model, operated by the US Navy and used for coordination of such operational activity as oil spill response), have serious problems with reproducing even main peaks in observations.
The best correspondence was achieved in a simple diagnostic model (SCUD), whose coefficients were optimized using historical data of satellite altimetry, scatterometry and trajectories of Lagrangian floats (Figure). In this presentation we discuss methods available to numerically study the drift of marine debris (particles versus tracer), to calibrate/validate models using sparse observational data and to maximize the utility of satellite and model products in various applications.
Figure. Model fluxes on the North America west coast between 40 and 51N as function of time and windage. Rows (from top to bottom) correspond to SCUD (a and f), SCUD-HYCOM (b and g), MOVE (c and h), FORA (d and i), and GNOME (e and j) models. Left column shows original model fluxes and right column same fluxes smoothed in time with a Gaussian 1.5-month half-width filter. White dots and lines mark peaks in model fluxes for different windages. Vertical magenta lines mark five main peaks in observations. Horizontal grey lines mark the optimal windage parameter values, derived from model comparison with boat reports, and span over the periods of the comparison. Color scale is strongly nonlinear and model flux units are fraction of the released tracer per a year. White dashed line in (d) illustrates faster drift and earlier arrival of higher windages.
Detailed analysis of model fluxes on the US/Canada West Coast and their comparison with reports of tsunami boats from the same region revealed serious difficulties that even best OGCMs (Ocean General Circulation Models), such as the HYCOM (Hybrid Coordinate Model, operated by the US Navy and used for coordination of such operational activity as oil spill response), have serious problems with reproducing even main peaks in observations.
The best correspondence was achieved in a simple diagnostic model (SCUD), whose coefficients were optimized using historical data of satellite altimetry, scatterometry and trajectories of Lagrangian floats (Figure). In this presentation we discuss methods available to numerically study the drift of marine debris (particles versus tracer), to calibrate/validate models using sparse observational data and to maximize the utility of satellite and model products in various applications.
Figure. Model fluxes on the North America west coast between 40 and 51N as function of time and windage. Rows (from top to bottom) correspond to SCUD (a and f), SCUD-HYCOM (b and g), MOVE (c and h), FORA (d and i), and GNOME (e and j) models. Left column shows original model fluxes and right column same fluxes smoothed in time with a Gaussian 1.5-month half-width filter. White dots and lines mark peaks in model fluxes for different windages. Vertical magenta lines mark five main peaks in observations. Horizontal grey lines mark the optimal windage parameter values, derived from model comparison with boat reports, and span over the periods of the comparison. Color scale is strongly nonlinear and model flux units are fraction of the released tracer per a year. White dashed line in (d) illustrates faster drift and earlier arrival of higher windages.