Laboratory observations and simple models of slow oscillations in cooled salt-stratified bodies

Abstract

A laboratory basin of saltwater with a freshwater layer at the top is exposed to cooling from above and inflow of freshwater from a layer that lies alongside. It exhibits oscillations of temperature, salinity, and layer depth when certain parameters are set. A simple analytical model illustrates the basic mechanism. Upon cooling, the layer gradually becomes denser. Ultimately, a density inversion at the interface between the layer and deeper fluid causes a gravitational instability. The layer is mixed with the deeper water and it vanishes. The emplacement of a new thin layer from freshwater inflow at the top repeats the cycle. Layered numerical models duplicate this cycle. Examples incorporate transport of heat out of the top and into the bottom of the layer. In addition, there is salinity transport up through the bottom from turbulent mixing. They reproduce the qualitative structure of the experimental data, but the many simplifications in the model do not permit precise quantitative agreement.

Convection driven by temperature and composition flux with the same diffusivity

“…The pulsations differ from the oscillations at marginal stability in double diffusion (Veronis 1964, Turner 1973 and with multiple boundary conditions, (Welander 1989, Tsitverblit 2007, which occur for Ra > 0 with stabilising salinity present and in some thermohaline laboratory experiments (te Raa 2001, Whitehead et al 2005. They are also unlike the binary fluid oscillations (Matura and Lücke 2006) that are travelling waves.…”

“…A chamber exposed to a temperature difference and a flux of salt water pumped steadily into a fresh water environment recovers the abrupt transitions and hysteresis of the original Stommel box model and a slightly more complex chamber produces a severely limited hysteresis range (Whitehead 2009). In other experiments, spontaneous oscillations back and forth between temperature and salinity driven flows occur (Whitehead et al 2005).…”

A Simple Conceptual Model for the Self‐Sustained Multidecadal AMOC Variability

“…However, Stommel's Two‐Box Model itself does not have self‐sustained multidecadal AMOC oscillation solutions. The revised Stommel's Three‐Box Model (R. Zhang et al., 2002), Four‐Box Model (Colin de Verdière, 2007), or Six‐Box Model (Colin de Verdière et al., 2006) has been combined with Welander‐type heat–salt oscillators (Welander, 1982; Whitehead et al., 2005) to explain ocean model simulated self‐sustained millennial‐scale AMOC relaxation oscillations (e.g., Colin de Verdière et al., 2006; Whitehead, 2018; Winton & Sarachik, 1993; R. Zhang et al., 2002) in the paleo context. The revised Stommel's Four‐Box Model has been employed to explain the stochastic forced damped multidecadal AMOC oscillations (Griffies & Tziperman, 1995).…”