---
title: Subcritical Baroclinic Instability
url: https://www.emergentmind.com/topics/subcritical-baroclinic-instability
type: topic
---

# Subcritical Baroclinic Instability

Subcritical baroclinic instability is a nonlinear baroclinic amplification process in rotating, radially stratified disks in which finite-amplitude vortices extract available potential energy from an unfavorable entropy gradient, provided that thermal transfer is neither absent nor excessively rapid. In the protoplanetary-disk literature, the term “subcritical” denotes the requirement of a finite perturbation amplitude: the background flow alone is not linearly unstable in the usual infinitesimal-perturbation sense, but once a vortex exists, baroclinic torque can reinforce it through a feedback between entropy anomalies and vorticity generation. Within this usage, the instability is primarily a vortex amplification and sustainment mechanism rather than a universal primary vortex-formation mechanism, and its efficacy depends sensitively on the radial entropy gradient, thermal relaxation or diffusion, three-dimensional stratification, and competition with elliptic or magnetohydrodynamic instabilities [1602.09063; 1011.0497; 1601.01921].

## 1. Definition and conceptual scope

In protoplanetary-disk models, the SBI is defined by three necessary ingredients: a radially unstable entropy stratification, a thermal transfer process, and finite-amplitude perturbations that seed initial vortices. It is therefore explicitly nonlinear and does not arise as a linear instability of the background flow alone [1602.09063]. In local compressible shearing-box models, this subcriticality is tied to the fact that perturbations must overcome strong Keplerian shear; finite-amplitude density or entropy perturbations are introduced, and any subsequent vortex growth is attributed to baroclinic feedback rather than to a conventional linear eigenmode [1011.0497].

The defining baroclinic ingredient is the misalignment of pressure and density gradients, written in compressible inviscid form as
$$
\nabla P \times \nabla \rho \neq 0.
$$
This misalignment supplies a vorticity source. In the disk interpretation used for SBI, a vortex advects fluid parcels through the background entropy gradient, thermal transfer produces a phase shift between thermodynamic and kinematic anomalies, and the resulting baroclinic torque amplifies anticyclonic vorticity when the radial entropy stratification is unstable [1602.09063].

The term should be distinguished from several adjacent but non-identical usages of “baroclinic instability.” In one rapidly rotating plane-layer problem with imposed thermal wind, baroclinic instability occurs for negative critical Rayleigh number and is described as subcritical with respect to convection; that is a smooth transition from convective to shear-driven modes, not the finite-amplitude vortex-feedback process of disk SBI [1111.0186]. In terrestrial-exoplanet climate theory, “baroclinic criticality” is a marginality parameter,
$$
\xi = s\frac{a}{H},
$$
used to characterize how close an atmosphere is to baroclinic instability; that framework does not discuss subcritical baroclinic instability explicitly and does not introduce finite-amplitude triggering, hysteresis, or metastable branches [1908.02661]. A recurrent source of confusion is therefore terminological: “subcritical” in SBI refers to nonlinear triggering and sustainment of vortices, whereas in other baroclinic contexts it may refer to onset below a convective threshold or to proximity to a critical state.

## 2. Thermodynamic criteria and local diagnostics

A standard local diagnostic for the disk SBI is the radial Brunt–Väisälä frequency. In the two-dimensional compressible disk formulation,
$$
N_r^2=-{ \frac1{\gamma\rho} \frac{\partial P}{\partial r}\frac{\partial}{\partial r}\ln\left(\frac{P}{\rho^\gamma}\right)},
$$
with the Schwarzschild criterion interpreted as \(N_r^2>0\) for stable stratification and \(N_r^2<0\) for unstable stratification [1602.09063]. For an equilibrium power-law disk with
$$
T_D \propto r^{-q}, \qquad \Sigma_D \propto r^{-p},
$$
the same work derives
$$
N_r^2= \frac{(p+q)(p(\gamma-1)-q)}{\gamma r^2}T,
$$
so that radial stratification is unstable when
$$
p< \frac{q}{\gamma-1}.
$$
A related formulation in vertically shearing disks writes the radial Brunt–Väisälä frequency as
$$
N_R^2=-{1 \over C_{\rm p} \rho}{\partial P \over \partial R}{\partial S \over \partial R},
$$
with the entropy per unit mass
$$
S=C_{\rm p} \ln(P^{1/\gamma}\rho^{-1}),
$$
and summarizes the negative midplane entropy-gradient condition as
$$
q+p(1-\gamma) < 0.
$$
In that framework the SBI again requires \(N_R^2<0\) together with thermal relaxation on a timescale comparable to orbital time [1601.01921].

Thermal transfer is central because the baroclinic feedback depends on a finite thermodynamic lag. With relaxation, the temperature evolves toward a reference profile \(T_0\) on timescale \(\tau\), as in
$$
{\partial \rho E \over \partial t} + \vec{\nabla}.\vec{V}(\rho E+P) =\rho \vec{V} .\vec{\nabla}\Phi - {\rho \over \gamma-1} {T-T_0 \over \tau},
$$
or equivalently through the source term
$$
Q = \frac{\gamma}{\gamma-1}\rho \frac{(T-T_0)}{\tau}.
$$
With diffusion, the source term is
$$
Q = \frac{\gamma}{\gamma-1}\frac{\rho}{Pe}\Delta (T-T_0),
$$
where
$$
Pe = \frac{H(r_*)\, r_*\, \Omega(r_*)}{\kappa}.
$$
The common result is that thermal adjustment must be intermediate: too short a thermal timescale suppresses the phase relation needed for vortex growth, while adiabatic or excessively long thermal evolution removes effective amplification [1602.09063; 1601.01921].

This criterion is often summarized qualitatively as follows. Too short \(\tau\) prevents vortex formation by the SBI; too long \(\tau\) does not allow vortices to be amplified; the optimal \(\tau\) is of order one orbital period [1601.01921]. A plausible implication is that SBI should be understood as a resonance-like thermodynamic feedback rather than as a purely kinematic consequence of adverse entropy stratification.

## 3. Nonlinear amplification cycle and two-dimensional phenomenology

The amplification cycle in SBI is explicitly vortex-mediated. A finite perturbation first produces a vortex. As gas parcels execute azimuthal and radial excursions around the vortex, thermal transfer modifies their temperature and entropy relative to the surroundings. In an unstable entropy gradient, the parcel is accelerated or decelerated in such a way that net vorticity increases; in a stable entropy gradient, the same cycle operates with opposite sign and the vortex decays [1602.09063]. Local shearing-box work expresses the same idea by stating that the vortex creates a local entropy gradient around itself, that this gradient opposes the background entropy gradient that made the vortex, and that the resulting baroclinic torque reinforces the vortex [1011.0497].

In two-dimensional compressible simulations with thermal relaxation, the evolution separates into two stages. First, random density bumps of amplitude \(\sim 10\%\) rapidly form many small vortices, almost independent of stratification. Second, once vortices exist, their subsequent amplification or decay depends on the sign of \(N_r^2\): stable stratification yields gradual decay, whereas unstable stratification yields amplification over hundreds of orbits [1602.09063]. The corresponding local shearing-box calculations likewise find slow growth over hundreds of orbits, with the strongest sustained growth when the thermal relaxation time is comparable to the orbital time [1011.0497].

The dependence on thermal transfer model is structurally important. Under thermal relaxation, the strongest growth occurs for intermediate \(\tau\), roughly \(\tau \sim \Omega_K^{-1}\). Very short cooling (\(\tau=0.1\)) yields no effective vortex formation, and the adiabatic case (\(\tau=\infty\)) yields vortices that do not amplify [1602.09063]. Under heat diffusion, amplification appears only when diffusion is not too weak, with \(Pe \gtrsim 2\times 10^4\) and an optimal range around \(Pe \approx 10^5\); at \(Pe=\infty\) there is no diffusion and no amplification [1602.09063].

The vortex morphology also depends on the thermal model. With relaxation, amplified vortices remain approximately Kida-like, with Rossby number and aspect ratio related by
$$
Ro = -\frac{3}{4}\frac{\chi+1}{\chi(\chi-1)}.
$$
During amplification the Rossby number increases in magnitude, the vortex size grows, the aspect ratio \(\chi\) decreases, and mergers can produce a single dominant structure [1602.09063]. With diffusion, by contrast, hollow vortices form: an annular vorticity enhancement develops, the central vorticity weakens relative to the rim, and the core becomes turbulent while the outer rim remains coherent. The paper attributes this to spatially nonuniform thermal feedback, with amplification in the periphery and damping in the core [1602.09063].

## 4. Three-dimensional stratification, vortex survival, and interaction with other instabilities

Three-dimensional disk structure changes the SBI from a vertically uniform picture to a layered one. In stratified disks, the radial Brunt–Väisälä frequency becomes height dependent,
$$
N_r^2 = -\frac{RT}{\gamma} \left( \frac{\partial \ln \rho_e}{\partial r} - \frac{q}{r} \right) \left( (1-\gamma)\frac{\partial \ln \rho_e}{\partial r} - \frac{q}{r} \right),
$$
so the disk can contain stable and unstable vertical layers simultaneously [1602.09063]. Two topologies are identified: a mid-plane unstable layer when \(\beta(\gamma-1)-q<0\), and an intermediate unstable layer when \(\beta(\gamma-1)-q>0\). When the unstable layer contains the mid-plane, the 3D SBI develops first there as in 2D, producing large-scale vortices that are subsequently stretched into stable layers, creating long-lived columnar vortical structures extending through the disk height [1602.09063].

These columnar vortices coexist with thinner internal vortex layers associated with baroclinic critical layers. The observed rays in the \(r\)-\(z\) plane are interpreted through
$$
\omega - m\Omega_D(r,z)= \pm N_z(r,z),
$$
and, assuming the mid-plane vortex is the forcing source,
$$
m[\Omega_D(r_v,0)-\Omega_D(r,z)] = \pm \sqrt{\gamma-1}\,\Omega_K(r)\,\frac{z}{H(r)}.
$$
The reported ray inclinations agree with the theoretical critical-layer slopes for \(m=16,24,32\), and the associated off-midplane vortices are treated as real physical structures rather than numerical artifacts [1602.09063].

The major destructive competitor in 3D is the elliptic instability. In unstratified 3D tests, a baroclinic vortex with \(\chi=4\) develops the expected vertical wavy motions but is not destroyed when baroclinic amplification continues to inject vorticity; instead it reorganizes into a structure with a turbulent core [1602.09063]. In vertically shearing disk simulations, the survival time is strongly aspect-ratio dependent. For a Keplerian disk, the elliptic-instability growth rate is written
$$
\Gamma={3 \over 2} \Omega \sqrt{-\left({4 \over 3}- {\chi \over {\chi-1}\right)\left({4 \over 3}- {1 \over {\chi(\chi-1)}\right)} ,
$$
with vertical modes dominating for \(1.5 < \chi < 4\), vortices with \(4 < \chi < 5.9\) stable in an unstratified flow, and a growth rate for \(\chi > 6\) about 50 times smaller than for \(\chi < 4\) [1601.01921]. This makes large-\(\chi\) vortices much easier to sustain.

A key synthesis emerges from simulations in which the vertical shear instability produces the initial vorticity perturbations. There, vortices form indirectly when VSI-generated axisymmetric vorticity bands create extrema in
$$
\mathcal{L}={\Sigma \over \omega_z}{\left(P \over \Sigma^{\gamma}\right)}^{2/\gamma},
$$
destabilize through the Rossby wave instability, and roll up into discrete vortices. The SBI is not the primary formation mechanism in that scenario; rather, it becomes important when steep outwardly decreasing entropy profiles and suitable thermal relaxation allow long-lived vortices to resist elliptic destruction [1601.01921]. This suggests a two-step organization in some disk models: VSI or another process seeds vortices, while SBI amplifies and maintains them if the baroclinic stratification is sufficiently unfavorable.

## 5. Relation to other baroclinic instabilities

The label “subcritical baroclinic instability” is not interchangeable with every baroclinic instability discussed in geophysics, astrophysics, or planetary fluid dynamics. The literature represented here spans several distinct mechanisms.

| System | Mechanism in cited work | Relation to SBI |
|---|---|---|
| Rapidly rotating plane layer with thermal wind [1111.0186] | Smooth transition from convective modes to shear-driven baroclinic modes; instability can occur for \(Ra<0\) | Subcritical with respect to convection, not finite-amplitude vortex SBI |
| Two-layer Philips disk model [1203.1686] | Mixed barotropic-baroclinic instability from phase locking of Rossby waves at PV defects | Linear Rossby-wave interaction, not canonical nonlinear SBI |
| Differentially rotating stellar radiation zone [1308.6046] | Linear baroclinic instability of global \(r\)- and \(g\)-mode families at very small \(q\) | Baroclinic wave instability in a stratified star, not finite-amplitude disk-vortex feedback |
| Terrestrial exoplanet atmospheres [1908.02661] | Baroclinic criticality \(\xi\) as a marginality parameter controlling heat transport | Criticality theory; paper does not discuss SBI explicitly |
| Stratified thermal upper-ocean flow [2403.18084] | Spectral and Lyapunov stability of a baroclinic jet with vertical curvature; short-wave growth vanishes | Stratification modifies instability but does not establish a general SBI mechanism |
| Rotating annulus with stable salinity stratification [1604.08109] | Local baroclinic onset in shallow double-diffusive layers (“barostrat instability”) | Local threshold instability, not finite-amplitude subcritical bifurcation |

These distinctions matter because the common ingredient—baroclinicity—does not determine the bifurcation structure. The mixed barotropic-baroclinic instability in a two-layer disk model is linear and arises from interacting Rossby waves trapped at potential-vorticity defects [1203.1686]. The stellar radiation-zone instability is also linear, with two unstable disturbance families corresponding to Rossby waves and internal gravity waves, and its onset occurs at very small rotation inhomogeneity \(q \sim 10^{-4}\) in the upper solar radiation zone [1308.6046]. By contrast, the disk SBI is defined by finite-amplitude triggering, thermodynamic phase lag, and vortex sustainment.

A related misconception concerns stratification. In stratified thermal shallow-water theory, one study finds that stratification changes the unstable region, slows Rossby waves because \(R_S<R\) when \(S>0\), and makes the short-wave growth rate vanish, yet does not support the generality of earlier numerical evidence for universal suppression of submesoscale activity [2403.18084]. This is adjacent to SBI only in the broad sense that both problems concern how baroclinicity and stratification alter nonlinear vortex dynamics.

## 6. Astrophysical role, limitations, and common misunderstandings

In protoplanetary disks, the strongest astrophysical case for SBI lies in low-ionization dead zones. Unmagnetized or very weakly coupled local compressible models show that the instability can generate and sustain large-scale vortices, with measured transport
$$
\alpha \approx 5\times10^{-3},
$$
slow growth on the order of hundreds of orbits, and vortices growing to the sonic scale [1011.0497]. The same work concludes that a baroclinically unstable dead zone should be characterized by large-scale vortices whose cores are elliptically unstable yet sustained by baroclinic feedback.

The limitation is magnetic coupling. Once magnetic fields are included, the magnetorotational instability grows much faster than the BI and dominates saturation, while the vortex core is disrupted by the magneto-elliptic instability. The combined MRI+BI state is reported to be virtually identical to MRI-only turbulence, and vortex excitation and self-sustenance by the baroclinic instability are described as viable only in low ionization, that is, the dead zone [1011.0497]. This sharply confines the regime of direct astrophysical relevance.

A second limitation is that SBI is not a universal primary vortex generator. In disks susceptible to the vertical shear instability, vortices can be generated through VSI-induced vorticity bands and subsequent Rossby wave instability, with SBI acting only as a sustainment mechanism when entropy gradients and cooling are favorable [1601.01921]. In the outer protoplanetary disk, where VSI is most likely to occur, this same study concludes that vortices formed by the VSI are likely to be short-lived overall, because the entropy gradients inferred observationally are usually too weak for the SBI to maintain them efficiently [1601.01921]. This suggests that long-lived SBI-supported vortices should be treated as conditional rather than generic features of outer disks.

A final misunderstanding is to equate any baroclinic threshold problem with SBI. The rotating-annulus “barostrat instability” shows that a globally stable full-depth configuration can become locally unstable in shallow convective layers once stratification reduces the effective vertical scale, but the authors explicitly do not frame that result as subcritical onset in the finite-amplitude sense [1604.08109]. Likewise, baroclinic criticality theory for exoplanet atmospheres is about self-organization near marginal instability, not about nonlinear finite-amplitude triggering [1908.02661]. The specific content of SBI remains narrower and more technical: a nonlinear, thermally mediated, vortex-based instability of baroclinic disk flows whose persistence depends on the balance between vorticity production and competing 3D or MHD destruction mechanisms.

Source: https://www.emergentmind.com/topics/subcritical-baroclinic-instability