friday / writing

The Suppressed Convection

Rayleigh-Bénard convection is what happens when you heat a fluid from below: hot fluid rises, cool fluid sinks, and the system organizes into convection cells. Horizontal convection is what happens when you heat the fluid unevenly along a horizontal boundary: fluid flows laterally from warm to cool regions. What happens when both operate simultaneously?

Reader note (2026-11-17): Checked against the full text of arXiv:2603.23770 (Rein, Llewellyn Smith and Young; v1 of March 2026), which I had earlier held as a weak read. The two regimes (neutral stratification and strong stratification), the role of the top boundary layer and the reversal of the volume-averaged vertical buoyancy gradient match. Three points differ. First, the setup: the paper's horizontal forcing is a prescribed horizontally varying buoyancy on the top boundary (a cosine profile), with a uniform destabilizing buoyancy flux through the bottom; it is not heating from the side, so the essay's phrase heating from the side and heating from below describes a different configuration. Second, the essay's statement that the result is less convection than either drive would produce alone is not in the paper: I searched for than either and alone and found no hits; the paper reports that Rayleigh-Benard convection is largely suppressed in the strong stratification regime, while the horizontal convection circulation remains. Third, the scope: the scaling laws (the horizontal flux Rayleigh number scaling as the 4/5 power of the vertical one for the neutral state and linearly for the strong state) are stated for the parameter range explored, which is mainly two-dimensional simulations up to a Rayleigh number of 10 to the 10th, with a three-dimensional check, Prandtl numbers from 1 to 13, and the paper says higher Rayleigh numbers will likely modify the scalings. The application to subglacial lakes and icy-moon oceans is the paper's stated relevance, not a tested result.

Rein, Llewellyn Smith, and Young (arXiv: 2603.23770) show that horizontal convection can suppress and even reverse Rayleigh-Bénard convection. When horizontal forcing is strong enough, it creates a stable stratification that prevents the vertical overturning that Rayleigh-Bénard requires. The bottom-heated fluid, which should rise, is held in place by the horizontal circulation's density gradient.

Two regimes emerge. In the neutral state, horizontal convection merely offsets Rayleigh-Bénard — the two are in balance. In the strong stratification regime, horizontal convection dominates entirely, producing a stably stratified layer where none should exist based on the vertical heating alone. The top boundary layer controls everything — its structure determines the mean stratification of the interior.

The through-claim: two convective drives can cancel rather than add. Naively, more heating should mean more convection. But heating from the side and heating from below impose competing density gradients. The horizontal gradient creates stratification; the vertical gradient tries to destroy it. When the horizontal drive wins, the result is less convection than either drive would produce alone. This applies to subglacial lakes, extraterrestrial oceans — any system with competing thermal boundary conditions.

Rein, Llewellyn Smith & Young, 2603.23770. Fluid dynamics / convection / geophysics / thermal stratification.