Driven around the bend: Spatial evolution and controls on the orientation of helical bend flow in a natural submarine gravity current

Submarine channel systems transport vast amounts of terrestrial sediment into the deep sea. Understanding the dynamics of the gravity currents that create these systems, and in particular, how these flows interact with and form bends, is fundamental to predicting system architecture and evolution. Bend flow is characterized by a helical structure and in rivers typically comprises inwardly directed near-bed flow and outwardly directed near-surface flow. Following a decade of debate, it is now accepted that helical flow in submarine channel bends can exhibit a variety of structures including being opposed to that observed in rivers. The new challenge is to understand what controls the orientation of helical flow cells within submarine flows and determines the conditions for reversal. We present data from the Black Sea showing, for the first time, the three-dimensional velocity and density structure of an active submarine gravity current. By calculating the forces acting on the flow, we evaluate what controls the orientation of helical flow cells. We demonstrate that radial pressure gradients caused by across-channel stratification of the flow are more important than centrifugal acceleration in controlling the orientation of helical flow. We also demonstrate that nonlocal acceleration of the flow due to topographic forcing and downstream advection of the cross-stream flow are significant terms in the momentum balance. These findings have major implications for conceptual and numerical models of submarine channel dynamics, because they show that three-dimensional models that incorporate across-channel flow stratification are required to accurately represent curvature-induced helical flow in such systems. Key Points First data of the 3-D velocity and density structure of a submarine gravity flow Internal pressure gradients, not centrifugal forces control flow structure Flow stratification and downstream advection must be included in future models.