Scale- and Process-Jumps in a Multimodel Project on Hurricane Impacts at the Seabed
The multi-institution project "Shelf-Slope Sediment Exchange in the Northern Gulf of Mexico..." used a chain of models to examine the effect of hurricanes on the seabed in the Gulf of Mexico and their effect on infrastructure. The components were ROMS (Rutgers U) for weather, wave climate and currents; CSTMS (VIMS) for sediment transport, WBMsed for river discharges and hurriSlip for sediment failures and gravity ignitions (U Colorado), LES/RANS-TURBINS (U California, Santa Barbara) for downslope turbidity currents. The project was an ambitious testing of the ability of these models to connect, and achieved most of the goals set out. The test case was 3 years of oceanographic data for Louisiana-Mississippi-Alabama including storms Dolly, Gustav, and Ike.
Connection between the models was a challenge on several levels. (i) Physical dimensionalities of the models were between 3D, 2.5D, 2D, and pointwise 0D. Model resolutions also varied. (ii) The models also reported daily, 3-hourly, or event- based responses. However different spatial and temporal scales in nature are observed to show different variances on aspects like sediment, and event return-time scalings. These are an added, often unforeseen difficulty for assembling multi-models,and validating them with data. (iii) Important process-jumps are involved in the total chain of sediment transport. Perhaps the most riveting are seabed mass-failures, and turbidity-current ignitions. Actually, separate modules often requiring energy considerations are required to handle such sharp phase-changes - sections of metastable deposits on the seafloor becoming slides, and of advecting suspended sediment layers becoming turbulent gravity flows.