Geostrophic Turbulent Scaling

• Geostrophic turbulent scaling is a framework that defines scaling laws for heat transport, flow speed, and convective lengths in rapidly rotating convection.
• It utilizes key parameters like Rayleigh, Ekman, and Prandtl numbers to characterize transitions into diffusivity-free regimes dominated by buoyancy and rotation.
• The approach is validated through lab experiments and DNS, offering insights for modeling planetary and stellar interiors and assessing the effects of boundary conditions and latitude.

Geostrophic turbulent scaling denotes the set of scaling relations that organize heat transport, flow speed, convective length scales, and spectra in rapidly rotating convection and quasi-geostrophic turbulence. In rotating Rayleigh–Bénard convection, radiatively driven convection, and related reduced models, these relations are commonly formulated in terms of the Rayleigh, Ekman, and Prandtl numbers, or in reduced combinations such as $\widetilde{Ra}=Ra\,Ek^{4/3}$ and $\mathcal{R}=Ra^{(P)}E^3/Pr^2$. In the most strongly rotation-dominated cases, the regime is described as “geostrophic turbulence”, “diffusivity-free”, or “ultimate”, indicating that bulk transport and velocity become governed primarily by buoyancy and rotation rather than molecular diffusivities; recent work further extends the framework to explicit latitude dependence, inverse-cascade condensates, boundary-sensitive behavior, and magnetically damped departures from the nonmagnetic laws (Hadjerci et al., 2024, Kannan et al., 24 Aug 2025).

1. Regime definition and control parameters

Geostrophic turbulent scaling is typically discussed for rapidly rotating thermal convection in which $Ek$ (or $E$) is small, $Ra$ is sufficiently supercritical, and turbulence coexists with strong rotational constraint. In the RRBC literature, the basic parameters are

$$Ra=\frac{g\alpha\Delta T H^3}{\nu\kappa},\qquad Ek=\frac{\nu}{2\Omega H^2},\qquad Pr=\frac{\nu}{\kappa},$$

with the reduced geostrophic Rayleigh number

$\widetilde{Ra}=Ra\,Ek^{4/3}.$

For radiatively driven convection, the corresponding flux-based control parameter is

$$Ra^{(P)}=\frac{\alpha g P H^4}{\rho C \kappa^2 \nu},$$

and the diffusivity-free reduced control parameter is

$\mathcal{R}=\frac{Ra^{(P)}E^3}{Pr^2}.$

The principal output observables are the Nusselt number $Nu$, Reynolds number $\mathcal{R}=Ra^{(P)}E^3/Pr^2$0, temperature fluctuation Rayleigh number $\mathcal{R}=Ra^{(P)}E^3/Pr^2$1, and the dimensionless horizontal convective scale $\mathcal{R}=Ra^{(P)}E^3/Pr^2$2 (Hadjerci et al., 2024).

The transition into the geostrophic regime is itself a scaling problem. Experiments in helium gas found that the transition to geostrophic turbulence is essentially Pr-independent and is described by

$\mathcal{R}=Ra^{(P)}E^3/Pr^2$3

with data collapse when heat transport is plotted as a function of $\mathcal{R}=Ra^{(P)}E^3/Pr^2$4 (Ecke et al., 2013). In direct simulations of classical RRBC, the transition is gradual and does not exactly coincide in $\mathcal{R}=Ra^{(P)}E^3/Pr^2$5 for different flow indicators; for the cases reported at $\mathcal{R}=Ra^{(P)}E^3/Pr^2$6 and $\mathcal{R}=Ra^{(P)}E^3/Pr^2$7, transitions occurred at roughly $\mathcal{R}=Ra^{(P)}E^3/Pr^2$8 and $\mathcal{R}=Ra^{(P)}E^3/Pr^2$9, respectively (Kunnen et al., 2014).

A further geometrical control parameter is latitude $Ek$0, defined by $Ek$1 at the equator and $Ek$2 at the pole. In the tilted-axis RRBC framework, latitude becomes the key parameter that couples local rotation–gravity geometry to convective transport and structure, enabling explicit regional scalings in spherical convection (Kannan et al., 24 Aug 2025).

2. Canonical transport and velocity laws

A central asymptotic statement of geostrophic turbulent scaling is that heat transport and velocity become diffusivity-free in the bulk. For rapidly rotating radiatively driven convection, asymptotic theory predicts

$Ek$3

or, in reduced variables,

$Ek$4

The same framework gives

$Ek$5

Re-expressed using the standard Rayleigh number, the heat-transport law is

$Ek$6

These relations were simultaneously validated for heat transport, temperature fluctuations, flow speed, and horizontal flow structure by combined laboratory experiments and DNS in radiatively driven convection (Hadjerci et al., 2024).

Experimental heat-transport measurements in the same radiative configuration found a master-curve collapse for $Ek$7 versus $Ek$8 in the rapidly rotating regime $Ek$9. Over approximately 1.5 decades in $E$0, the measured exponents were $E$1 for $E$2 and $E$3 for $E$4, while DNS gave $E$5, all consistent with the theoretical geostrophic-turbulence prediction $E$6 (Bouillaut et al., 2022).

Small-scale quasi-geostrophic convective turbulence without depth-invariant large-scale vortices approaches the same diffusivity-free forms at large $E$7: $E$8 At the same time, this setting shows that transport exponents can approach diffusion-free behavior even when the dynamically dominant length scales remain viscously controlled. This distinction between transport scaling and structural scaling is one of the recurrent themes in the subject (Oliver et al., 2023).

3. Latitude-dependent heat transport and length-scale scaling

The most explicit regional formulation of geostrophic turbulent scaling is the latitude-dependent theory developed for tilted RRBC and validated against spherical-shell DNS. At high latitudes, where the Taylor–Proudman constraint yields vertically aligned columns, the convective length scale perpendicular to the rotation axis obeys

$E$9

near onset, and

$Ra$0

above onset. The corresponding Nusselt-number laws are

$Ra$1

or equivalently

$Ra$2

near onset, and

$Ra$3

or equivalently

$Ra$4

above onset. At low latitudes, where the rotation axis is nearly horizontal and columns orient more horizontally, the above-onset scaling becomes

$Ra$5

with $Ra$6 the equatorial Nusselt number, which does not follow a simple $Ra$7, $Ra$8 scaling and must be obtained from DNS (Kannan et al., 24 Aug 2025).

Regime	$Ra$9 scaling	$$Ra=\frac{g\alpha\Delta T H^3}{\nu\kappa},\qquad Ek=\frac{\nu}{2\Omega H^2},\qquad Pr=\frac{\nu}{\kappa},$$0 scaling
High latitude, near onset	$$Ra=\frac{g\alpha\Delta T H^3}{\nu\kappa},\qquad Ek=\frac{\nu}{2\Omega H^2},\qquad Pr=\frac{\nu}{\kappa},$$1	$$Ra=\frac{g\alpha\Delta T H^3}{\nu\kappa},\qquad Ek=\frac{\nu}{2\Omega H^2},\qquad Pr=\frac{\nu}{\kappa},$$2
High latitude, above onset	$$Ra=\frac{g\alpha\Delta T H^3}{\nu\kappa},\qquad Ek=\frac{\nu}{2\Omega H^2},\qquad Pr=\frac{\nu}{\kappa},$$3	$$Ra=\frac{g\alpha\Delta T H^3}{\nu\kappa},\qquad Ek=\frac{\nu}{2\Omega H^2},\qquad Pr=\frac{\nu}{\kappa},$$4
Low latitude, above onset	$$Ra=\frac{g\alpha\Delta T H^3}{\nu\kappa},\qquad Ek=\frac{\nu}{2\Omega H^2},\qquad Pr=\frac{\nu}{\kappa},$$5	based on $$Ra=\frac{g\alpha\Delta T H^3}{\nu\kappa},\qquad Ek=\frac{\nu}{2\Omega H^2},\qquad Pr=\frac{\nu}{\kappa},$$6, $$Ra=\frac{g\alpha\Delta T H^3}{\nu\kappa},\qquad Ek=\frac{\nu}{2\Omega H^2},\qquad Pr=\frac{\nu}{\kappa},$$7

The crossover latitude is $$Ra=\frac{g\alpha\Delta T H^3}{\nu\kappa},\qquad Ek=\frac{\nu}{2\Omega H^2},\qquad Pr=\frac{\nu}{\kappa},$$8. Flow structures measured by the ratio $$Ra=\frac{g\alpha\Delta T H^3}{\nu\kappa},\qquad Ek=\frac{\nu}{2\Omega H^2},\qquad Pr=\frac{\nu}{\kappa},$$9 confirm the transition from vertically dominated polar convection to horizontally dominated equatorial convection. High-resolution DNS in both planar tilted RRBC and spherical-shell geometries show excellent quantitative agreement with the predicted $\widetilde{Ra}=Ra\,Ek^{4/3}.$0, $\widetilde{Ra}=Ra\,Ek^{4/3}.$1, and $\widetilde{Ra}=Ra\,Ek^{4/3}.$2 laws, with deviations observed only near the crossover or outside the theory’s validity range. The resulting interpretation connects local planar RRBC turbulence to global spherical convection and provides a quantitative framework for regional thermal transport in planetary and stellar interiors (Kannan et al., 24 Aug 2025).

4. Reynolds scaling, spectra, and force balances

Velocity scaling in geostrophic turbulence remains less settled than heat transport. At constant $\widetilde{Ra}=Ra\,Ek^{4/3}.$3 and $\widetilde{Ra}=Ra\,Ek^{4/3}.$4, stereoscopic PIV measurements in water found

$\widetilde{Ra}=Ra\,Ek^{4/3}.$5

These exponents are steeper than in nonrotating convection, and comparison with the theoretical viscous–Archimedean–Coriolis and Coriolis–inertial–Archimedean relations showed that both matched the data equally well. The same experiments reported convergence toward diffusion-free velocity scaling as $\widetilde{Ra}=Ra\,Ek^{4/3}.$6 decreases, but not full convergence at the accessible parameters (Madonia et al., 2022).

The spectral phenomenology in that regime contains both geostrophic organization and classical turbulent signatures. Kinetic-energy spectra exhibited a dominant peak at $\widetilde{Ra}=Ra\,Ek^{4/3}.$7 in the horizontal-velocity spectra, associated with a quadrupolar vortex filling the cross-section, while at larger $\widetilde{Ra}=Ra\,Ek^{4/3}.$8 a scaling range developed with exponent close to $\widetilde{Ra}=Ra\,Ek^{4/3}.$9. This coexistence indicates a quasi-two-dimensional large-scale organization superposed with increasingly developed small-scale turbulence (Madonia et al., 2022).

Reduced-model simulations clarify why transport laws and local balances need not align. In quasi-geostrophic convection with depth-invariant flows suppressed, the Taylor microscale does not vary significantly with increasing $$Ra^{(P)}=\frac{\alpha g P H^4}{\rho C \kappa^2 \nu},$$0, the integral length scale grows only weakly, and the computed length scales remain $$Ra^{(P)}=\frac{\alpha g P H^4}{\rho C \kappa^2 \nu},$$1 with respect to the linearly unstable critical wavenumber. The interior dynamics do not exhibit a point-wise CIA force balance. Instead, the dominant interior balance is between horizontal advection, vortex stretching, and the vertical pressure gradient, with a secondary viscous–buoyancy balance whose rms ratio approaches unity as $$Ra^{(P)}=\frac{\alpha g P H^4}{\rho C \kappa^2 \nu},$$2 increases. The proposed interpretation is a “non-local” CIA balance in which buoyancy is dominant within the thermal boundary layers and spatially separated from the interior Coriolis and inertial forces (Oliver et al., 2023).

Wave–geostrophic coupling provides another route into geostrophic structure. Near-resonant triadic interaction involving two inertial waves and a geostrophic mode yields a growth rate proportional to $$Ra^{(P)}=\frac{\alpha g P H^4}{\rho C \kappa^2 \nu},$$3 at low $$Ra^{(P)}=\frac{\alpha g P H^4}{\rho C \kappa^2 \nu},$$4, crossing over to a $$Ra^{(P)}=\frac{\alpha g P H^4}{\rho C \kappa^2 \nu},$$5 law at larger $$Ra^{(P)}=\frac{\alpha g P H^4}{\rho C \kappa^2 \nu},$$6, in excellent agreement with DNS (Reun et al., 2020).

5. Boundary conditions, inverse cascades, and condensates

Boundary conditions strongly affect which geostrophic turbulent scaling laws are realized. In DNS of RRBC with $$Ra^{(P)}=\frac{\alpha g P H^4}{\rho C \kappa^2 \nu},$$7, $$Ra^{(P)}=\frac{\alpha g P H^4}{\rho C \kappa^2 \nu},$$8 or $$Ra^{(P)}=\frac{\alpha g P H^4}{\rho C \kappa^2 \nu},$$9, and $\mathcal{R}=\frac{Ra^{(P)}E^3}{Pr^2}.$0, stress-free plates produced post-transition heat-transport scaling consistent with

$\mathcal{R}=\frac{Ra^{(P)}E^3}{Pr^2}.$1

whereas no-slip plates yielded a significantly shallower fixed-$\mathcal{R}=\frac{Ra^{(P)}E^3}{Pr^2}.$2 scaling with $\mathcal{R}=\frac{Ra^{(P)}E^3}{Pr^2}.$3–$\mathcal{R}=\frac{Ra^{(P)}E^3}{Pr^2}.$4 in $\mathcal{R}=\frac{Ra^{(P)}E^3}{Pr^2}.$5. The transition occurred at roughly the same parameter values for both boundary conditions, but the geostrophic-regime phenomenology diverged sharply: stress-free cases exhibited a large-scale barotropic vortex and an inverse energy cascade, while no-slip cases did not (Kunnen et al., 2014).

Domain anisotropy selects the form of the large-scale condensate. In square periodic domains, rapidly rotating RB convection produces a domain-filling barotropic vortex dipole. In rectangular domains, the dipole is replaced by a Kolmogorov-like array of alternating unidirectional jets with embedded vortices and a weak meandering transverse jet. The dipole-to-jet transition occurs at $\mathcal{R}=\frac{Ra^{(P)}E^3}{Pr^2}.$6. The associated spectra separate into a steep barotropic branch,

$\mathcal{R}=\frac{Ra^{(P)}E^3}{Pr^2}.$7

and a shallower baroclinic branch,

$\mathcal{R}=\frac{Ra^{(P)}E^3}{Pr^2}.$8

showing that the large-scale condensate and the small-scale turbulence obey distinct scaling laws even within the same flow (1711.01685).

Large-scale friction changes the condensate cutoff law. DNS of rotating RB convection with a linear friction term $\mathcal{R}=\frac{Ra^{(P)}E^3}{Pr^2}.$9 show that, contrary to the classical Kraichnan–Leith–Batchelor prediction $Nu$0, the large-scale-vortex radius follows

$Nu$1

The explanation is that the barotropic manifold displays

$Nu$2

over the upscale-transfer range, not the canonical $Nu$3 spectrum, and shell-to-shell transfer is strongly nonlocal. Circulation statistics further support scale-invariant coarse-grained vorticity through

$Nu$4

This replaces the classical friction-cutoff estimate by one tied to condensation-dominated, nonlocal inverse transfer (Ding et al., 1 Jun 2026).

6. Extensions beyond nonmagnetic RRBC

Geostrophic turbulent scaling extends beyond nonmagnetic plane-layer convection, but the extensions are not always simple continuations of the canonical laws. In liquid-metal rotating magnetoconvection with liquid gallium at $Nu$5 and $Nu$6, both rotating convection and rotating magnetoconvection follow geostrophic turbulent scaling when the local interaction parameter satisfies $Nu$7. For $Nu$8, Lorentz forces dominate local inertia and the flow enters a magnetically damped regime with

$Nu$9

while heat transfer is enhanced, a behavior attributed to increased coherence of vertically aligned magnetostrophic convective flow (Liu et al., 3 Sep 2025).

In SQG turbulence, the direct cascade has the time-averaged spectral law

$\mathcal{R}=Ra^{(P)}E^3/Pr^2$00

but temporal flux imbalance produces a subleading non-equilibrium correction. For SQG, the instantaneous correction scales as

$\mathcal{R}=Ra^{(P)}E^3/Pr^2$01

and DNS at $\mathcal{R}=Ra^{(P)}E^3/Pr^2$02 support both the $\mathcal{R}=Ra^{(P)}E^3/Pr^2$03 equilibrium law and the steeper $\mathcal{R}=Ra^{(P)}E^3/Pr^2$04 correction (Valadão et al., 2024). At still higher Reynolds number, SQG direct-cascade simulations recover the Kolmogorov spectrum at $\mathcal{R}=Ra^{(P)}E^3/Pr^2$05, with a converged Kolmogorov constant $\mathcal{R}=Ra^{(P)}E^3/Pr^2$06, while the Lyapunov exponent scales anomalously as

$\mathcal{R}=Ra^{(P)}E^3/Pr^2$07

rather than the dimensional $\mathcal{R}=Ra^{(P)}E^3/Pr^2$08 expectation (Valadão et al., 10 Apr 2025).

Atmospheric observations provide complementary cascade diagnostics. Offshore wind-speed spectra show a low-frequency regime consistent with geostrophic or 2D turbulence, with

$\mathcal{R}=Ra^{(P)}E^3/Pr^2$09

and the third-order structure function changes sign at the crossover frequency $\mathcal{R}=Ra^{(P)}E^3/Pr^2$10. After applying Taylor’s hypothesis locally, the third-order law is linear in separation for negative $\mathcal{R}=Ra^{(P)}E^3/Pr^2$11 and cubic for positive $\mathcal{R}=Ra^{(P)}E^3/Pr^2$12,

$\mathcal{R}=Ra^{(P)}E^3/Pr^2$13

with the crossover occurring at about the same separation distance of $\mathcal{R}=Ra^{(P)}E^3/Pr^2$14 km as in aircraft measurements (Sim et al., 2022).

Taken together, these results indicate that geostrophic turbulent scaling is not a single universal exponent set, but a family of asymptotic and crossover laws. The exponents depend on the observable under consideration, the forcing protocol, the presence or suppression of large-scale vortices, boundary conditions, latitude, magnetic damping, and whether the system is diagnosed through transport, spectra, or predictability. A plausible implication is that reliable extrapolation to planetary and stellar interiors requires both the correct scaling exponent and the correct structural regime in which that exponent is realized.