Instabilities and layering of a heated laboratory anticyclone

We generate and observe large (∼1 m diameter) anticyclonic vortices in a stably stratified rotating tank that are neutrally buoyant. A nonaxisymmetric ($m=2$) instability quickly develops, leading to subvortices and fine-structured shed filaments. When heated with respect to their surroundings the vortex lenses behave much the same way but with two conspicuous additions. First, prominent early, is that the circumferential edge appears serrated with cusplike features from lateral intrusions, surmised to be due to thermal convection. Second, a stepped upper layer develops due to thermohaline diffusive convection. We see no evidence of layering in the nonheated cases, indicating that viscohaline double diffusion appears to be inoperative, likely prevented by a turbulent background. These observations are considered with respect to previous laboratory work as well as to geophysical vortices that are thermally distinct from their environs, such as Meddies.

Figure 1

Schematic of the 13 m diameter Coriolis II rotating platform (a) along with a top view (b) and side view [(c), seen 20 s after injection] of a typical anticyclonic vortex lens approximately 1.3 m in diameter. The vertical shaft affixes a plunging conductivity/temperature probe, which may be seen in the foreground.

Figure 2

LIF data comparing example heated (experiment 37) and unheated (experiment 46) cases with respect to volume. The black line indicates the influx volume, which remains constant after 750 s. The resulting vortex volumes show significant entrainment. Although data is sparse, more mixing may be inferred in the heated case.

Figure 3

Representative PIV profiles of the azimuthal velocity across the eddy shortly after the injection ceases, $t=771$ s (experiment 37, injection volume 360 L). Panel (a) is a horizontal cut at the vortex center ($z=37$ cm) and (b) is a vertical cut at the radius of maximum velocity ($r=65$ cm).

Figure 4

Top view of vortices early in their evolution. The edge of the heated vortex appears serrated at higher $z$ (a) with convective cusps on the scale ∼3–5 cm [closeup (b)]. At lower $z$ (c), and for the unheated case at all $z$ [e.g. (d)], the circumference appears relatively smooth. Note in (c) and (d) the presence of background satellite vortices, shed from earlier episodes of $m=2$ instability.

Figure 5

Cross section of a typical $m=2$ shedding event, with time increasing for frames going downward. The development of a distinct fissure occurs in about 40 s (top three panels), while minutes later the peripheral material is more clearly separated. Scale is in cm.

Figure 6

Representative side view images of the vortices at intermediate times. A heated vortex lens (b) shows a diffuse layer that is a few cm thick directly above it (see arrow), which the unheated case (a) lacks. Note that the images do not necessarily show the entire vortex cross section as they are not exactly nor equally centered in the vertical laser sheet.

Figure 7

Probe traces of density and temperature comparing earlier and later times for the unheated (blue) vs heated (red) cases. As the heated lens evolves, the rms decreases while a staircase pattern in density is seen to emerge near $z=40-45$ cm.

Figure 8

Probe trace detail showing temporal development of the heated profiles.