Abstract
Laboratory experiments were performed to investigate the aggregate behavior of two identical parallel line turbulent plumes that were discharged to either a homogeneous water column or a two-layer fluid. These studies were motivated by the processes that occur at winter polar leads (long narrow cracks in the ice pack); fluid motions resulting from the refreezing of leads are often modeled as line plumes, and the interaction between such plumes signifies the interaction between the leads.
In both cases, the plumes initially descended as though the other plume was not present. In the former case (called the “direct merging” case), the interaction between the plumes was initiated at a nondimensional timescale of the order Ted/(X/Q1/30)≈3.2, where 2X is the distance between the plumes and Q0 is the buoyancy flux per unit length. The point of confluence between the plumes gradually rose to a quasi-steady depth ze. The ratio z/X was found to be a function of the plume Reynolds number R=Q1/30X/v, where v is the kinematic viscosity; at large R. it appeared that ze/X≈1.0. In the latter case (termed the “indirect merging” case) the plume heads first impinged on the interface and then split into gravity currents that spread on the interface. The counterflowing gravity currents between the plumes collided with each other, merged, deflected upward, and rose to a height he≈0.7D, before the two plumes initiated merging; here D is the depth of the upper fluid layer. In this case also, the confluence point rose to a quasi-steady depth. The results of the experiments were used to estimate the typical interaction timescales between thin polar leads; conditions under which such estimates are valid are also discussed.
