Focussed, point sources:

Plumes and jets are differentiated based on a comparison of the buoyancy and momentum of the fluid that forms them. If the non-dimensional flux Richardson number (Ri = Q*B^0.5*M^(-5/4), where M=Q*W, and W is a characteristic vertical velocity) is less than one, jet-like behavior is expected, whereas flows with Richardson numbers greater than one are termed plumes (McDuff 1995). McDuff provides estimates of typical plume parameters which I utilized in the design of laboratory apparatuses. In particular, he suggests that the Ricardson numbers of black smoker and diffuse flow are 0.30 and 27, respectively.

My most thorough examination of a focussed flow at Ri=0.3 was in a cubical tank large enough to ignore edge effects on the velocity field (at least until the plume reached the boundaries), filled with uniform density fluid. In an effort to observe (what I expected) to be radial entrainment, I covered the tank with cellophane and left it for 4 hours. During the experiment, I measured how much fluid entered the tank through a vertical nozzle with an inner diameter of 0.4cm, and thus was able to calculate an average Q=1.3cm^3/s. During most runs, this flow rate varied as the source fluid reservoir drained, but this Q was held nearly constant by suspending the reservoir (a bucket) from a spring. With this arrangement, the reservoir level remained at the same height above the surface of the experimental tank, providing approximately constant pressure head, and therefore Q.

Horizontal entrainment was observed by dropping 1mm glass beads (coated with the fluorescent green dye Fluorescein) into the initially calm tank, initiating the focussed source, and videotaping the deflection of the tracer profiles. Although the animated gif shown here does not do justice to the visual power of the technique, a light sheet (from cross-hair slits on an overhead projector) was scanned horizontally or vertically through the fluid. The vertical sheet, centered on the plume source in this sequence of images, was anticipated to illuminate the radial deflection of the tracers as ambient fluid was entrained by the flow.

At Ri=0.3 in the tank, the focussed flow was observed to be laminar upon entering the fluid, entraining fluid only within a few millimeters of the laminar surface; yet at and beyond the depth where the laminar flow first undulates and then suddenly becomes turbulent and plume-like, the apparent radius of entrainment is on the order of centimeters. This is clear evidence that turbulent processes control the entrainment dynamics, and that appropriate models of black smokers must account for the difficulty of simulating at small scales the role of turbulence present in natural flows. Trivett(1994) notes that "scaling the velocities up to field conditions yield unrealistically large vertical velocities (up to 20m/s, or 20 times the black smoker velocities)," and cautions that the entrainment by such sources may be unrealistically large. This is a concern to bear in mind in the next section in which I attempt to simulate both diffuse and focussed flows in the same tank.

Another major result was that the plume developed a vortical circulation, and the entrainment trajectories were not simply radial (except in an average sense). While the circulation is beautiful, it complicates measurement of radial entrainment velocity profiles. The effect may be minimized by ensuring that the tank is initially stagnant, and that the source fluid enters the tank plumb vertically so that no angular momentum relative to the plume axis is added cumulatively to the tank during the experiment.