**5. Conclusions**

A model-data workflow was developed to numerically represent sediment fluxes from fluvial sources on the inner continental shelf to the continental slope. The workflow is perhaps one of the more complex ever attempted for the problem of routing sediment from coastal sources to deep-sea sinks. The range of components (Figure 1) included: (1) database frameworks for sediment texture and bathymetry of the continental shelf and slope environments; (2) hydrology framework to simulate the discharge of water and sediment for multiple (fifteen) rivers geographically distributed along the northern Gulf of Mexico; (3) an ocean modeling framework that combined output from a spectral wave-action model with ocean circulation simulations, as driven by winds, tides and solar radiation; and tuned to the seafloor environments where bottom boundary layer dynamics can be su fficiently represented including the resuspension, transport and deposition of sediment; (4) a seafloor geotechnical modeling framework able to capture the strengthening and weakening of seafloor deposits, under both ambient ocean conditions, and high intensity, short-lived hurricanes; (5) a gravity flow generator able to determine the location(s) and sediment volume(s) displaced; and (6) a high-resolution CFD model able to simulate the development of a turbidity current, including the bottom shear stresses likely to impact o ffshore infrastructure. The immersed boundary *RANS* approach, in conjunction with the multiple successive streamwise modules, appears to be well suited to perform the Gulf of Mexico turbidity current simulations over the realistic length and time scales.

The workflow was exercised to explore the conditions that trigger episodes of sediment flux on the continental slope where gas and oil infrastructure exist. Several one-way nested grids from coarse to fine were developed to simulate the hydrodynamic circulation, sediment transport, sediment failure, sediment liquefaction, and turbidity currents in the northern Gulf of Mexico. A full Gulf of Mexico *ROMS* domain was run to provide boundary conditions to a higher-resolution grid that better resolved bathymetric features, river runo ff, and sediment transport. Ocean hydrodynamic simulations covered the period from 1 January 2000, to 31 December 2005 (spinup), and from 1 January 2006, to 31 December 2012. It allowed us to characterize sediment transport scenarios during diverse forcing events (river discharge, storms, and multiple hurricanes). We ran focused suspended sediment transport solutions from 1 October 2007, to 30 September 2008, a time period that saw very active tropical storms and major hurricanes crossing the study area.

The suspended sediment model indicated that episodic suspended transport down the Mississippi and De Soto Canyons was fed principally by sediment fluxes generated by wave resuspension on the shelf. During the two hurricanes modeled (Ike and Gustav), suspended sediment fluxes were predominantly seaward in the vicinity of the Mississippi and De Soto Canyons. Peak suspended sediment fluxes coexisted with the occurrence of the highest wave-induced bed stresses on the continental shelf, but showed increasingly long delays relative to this timing with distance down the canyon or continental slope. While hurricane conditions only lasted for two brief episodes during the one-year model run, they accounted for about 30% of the sediment delivered from the continental shelf to the slope. Delivery of sediment directly from settling from the freshwater river plume at the canyon head or over the continental slope provided a more gradual source of sediment delivery for the study period from 1 October 2007, to 30 September 2008. Plume delivery and transport during moderate-intensity frontal passages accounted for 70% of the total sediment delivered to the continental slope during the study period.

The workflow applied a newly developed ignitions model, which was used to explore some particular mechanisms for creating turbidity currents as an additional, and perhaps the major, transportation of sediments to the slope and into channelized features there. Modeling of the flows explored physical constraints on the flow velocities and forces.

On the continental slope, turbidity currents can be triggered by slope failure when storm-driven supply forces accumulation of sediment in deeper water and steeper slopes. These appeared intense enough to both erode sediment along the path of the turbidity current and to damage offshore infrastructure. Modeling efforts in the future should explore more two-way coupling along with workflows such as developed here, and take advantage of observational methods for developing model parameterizations and confirming model estimates.

**Author Contributions:** Author contributions summarized as: conceptualization, H.G.A., G.A., C.K.H., E.H.M., J.S.; methodology, all; software, H.G.A., J.J.B., S.C., E.W.H.H., C.J.J., T.A.K., S.R.; formal analysis, J.J.B., S.C., T.A.K., C.J.J., S.R.; writing—original draft preparation, C.K.H., J.S.; writing—review and editing, all; visualization, H.G.A., S.C., C.J.J., T.A.K., S.R.; supervision, H.G.A., G.A., C.K.H., E.H.M., J.S.; project administration, G.A.; funding acquisition, H.G.A., C.K.H., E.H.M., J.S. Author contributions can also be summarized by the modeling components on which each contributed: workflow (G.A., J.S.); ROMS (H.G.A.); CSTMS (C.K.H., J.J.B., T.A.K.); WBMsed (J.S., S.C.); HurriSlip (E.W.W.H., C.J.J.); TURBINS (E.H.M., S.R.). All authors have read and agreed to the published version of the manuscript.

**Funding:** J.S. was supported by U Colorado, CSDMS and BOEM funding, H.A., J.B., C.H., E.H., C.J., S.R.,T.K., and G.A. were supported by BOEM funding. E.M. and S.R. were supported by BOEM and NSF. S.C. was supported by the U Alabama and NSF. J.B. and T.K. were partially supported by VIMS. J.B. was partially supported by the U.S. Geological Survey, Coastal and Marine Hazards and Resources Program.

**Acknowledgments:** The authors appreciate technical support from D. Robertson (Rutgers U.) and D. Forrest (VIMS). Use of trade, firm or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government. The authors acknowledge William & Mary Research Computing for providing computational resources and technical support that have contributed to the results reported within this paper. The authors appreciate input from reviewers (two anonymous, and Mr. R.C. Mickey, USGS) that helped improve the manuscript. This is contribution number 3925 of the Virginia Institute of Marine Science, William & Mary.

**Conflicts of Interest:** The authors declare no conflict of interest.
