Introducing a simple convex hull method to calibrate diffusion coefficients in Lagrangian particle models
Yang Song,Ayumi Fujisaki-Manome,Christopher H. Barker,Amy MacFadyen,Dan Titze,James Kessler,Jia Wang
DOI: https://doi.org/10.1016/j.oceaneng.2024.119926
IF: 5
2024-12-01
Ocean Engineering
Abstract:Calibrating Lagrangian particle models (LPMs) is crucial for simulating oil spill trajectories, enabling emergency responses, and mitigating harmful effects on the environment. However, there is a lack of rigorous observation data on oil spill events, especially in water bodies with no serious spills but perceived increasing risk, hampering LPM application. This study utilized trajectory data from three drifters in Lake Erie, collected in 2016, 2018, and 2019. We then proposed a simple and efficient convex hull method to identify the smallest envelope of the simulated particle cloud, covering all drifter trajectories while simultaneously controlling over-diffusion. This approach was employed to calibrate the General NOAA Operational Modeling Environment oil spill model (an LPM) and determine optimal diffusion coefficients. Based on 362 one-day-long drifter trajectory simulation cases, the recommended diffusion coefficient for particle trajectory simulation in Lake Erie was determined to be 14 m 2 s −1 . The calibrated model demonstrated the ability to effectively capture the movement direction of drifters and control the discrepancy between simulated and observed trajectories. Meanwhile, it underestimated drifter travel distances due to errors in input currents, the primary force driving drifter movement. These results promote LPM application for particle trajectory simulations.
engineering, civil, ocean, marine,oceanography