6.4 Eddies & Gyres
Mesoscale eddies (10β300 km, weeks to months) dominate ocean kinetic energy, containing roughly 90% of the total. They arise from baroclinic instability of mean currents, transport heat, salt, nutrients, and marine organisms across ocean basins, and are now routinely detected by satellite altimetry.
Mesoscale Eddies
Ocean eddies are the analogues of atmospheric weather systems. They form primarily through baroclinic instability (Eady, 1949; Charney, 1947), which extracts available potential energy from sloping isopycnals (density surfaces) and converts it to eddy kinetic energy. The characteristic length scale is set by the first baroclinic Rossby deformation radius:
$$L_d = \frac{NH}{f}$$
where $N$ is the buoyancy frequency ($\sim 10^{-3}$ s$^{-1}$),$H$ is the ocean depth, and $f$ is the Coriolis parameter. $L_d$ varies from ~200 km in the tropics to ~10 km at high latitudes. Eddies with scales near $L_d$ grow fastest.
Warm-Core Rings (Anticyclonic)
Elevated SSH, depressed thermocline. Rotate clockwise in the NH. Trap warm water from the subtropical side of western boundary currents. Gulf Stream warm rings can persist for 1β2 years.
Cold-Core Rings (Cyclonic)
Depressed SSH, raised thermocline. Trap cold, nutrient-rich water. Higher biological productivity. Rotate anticlockwise in the NH. Agulhas rings (anticyclonic) transport Indian Ocean water into the Atlantic.
10β300 km
Diameter range
~10,000
Active eddies at any time
~90%
Share of ocean kinetic energy
Derivation: Potential Vorticity Conservation
Step 1: Start from the Shallow Water Equations
Consider a single layer of incompressible fluid on a rotating planet. The momentum and continuity equations are:
$$\frac{Du}{Dt} - fv = -g\frac{\partial\eta}{\partial x}, \quad \frac{Dv}{Dt} + fu = -g\frac{\partial\eta}{\partial y}$$
$$\frac{\partial h}{\partial t} + \frac{\partial(hu)}{\partial x} + \frac{\partial(hv)}{\partial y} = 0$$
where $h$ is the layer thickness and $\eta$ is the free surface displacement.
Step 2: Derive the Vorticity Equation
Take $\partial/\partial x$ of the $v$-equation minus $\partial/\partial y$ of the $u$-equation. Define relative vorticity $\zeta = \partial v/\partial x - \partial u/\partial y$:
$$\frac{D}{Dt}(f + \zeta) = -(f + \zeta)\left(\frac{\partial u}{\partial x} + \frac{\partial v}{\partial y}\right)$$
The right-hand side represents vortex stretching/compression due to horizontal divergence.
Step 3: Use Continuity to Eliminate Divergence
From the continuity equation, the horizontal divergence is related to the rate of change of layer thickness:
$$\frac{\partial u}{\partial x} + \frac{\partial v}{\partial y} = -\frac{1}{h}\frac{Dh}{Dt}$$
Substituting into the vorticity equation:
$$\frac{D}{Dt}(f + \zeta) = \frac{(f + \zeta)}{h}\frac{Dh}{Dt}$$
Step 4: Obtain PV Conservation
Dividing both sides by $h$ and recognizing the material derivative of a ratio:
$$\frac{D}{Dt}\left(\frac{f + \zeta}{h}\right) = 0$$
This is Ertel's theorem for shallow water: the potential vorticity $q = (f + \zeta)/h$ is conserved following a fluid parcel in the absence of friction and diabatic effects. This is one of the most powerful constraints in geophysical fluid dynamics.
Step 5: Physical Consequences
If a parcel moves poleward ($f$ increases), it must either spin down ($\zeta$ decreases) or the layer must thicken ($h$ increases) to conserve $q$. For a barotropic ocean of depth $H$, if $\zeta \ll f$:
$$\frac{f}{H} = \text{const} \implies \text{flow follows lines of } f/H$$
Over a flat bottom, this confines flow to latitude circles. Topographic features ($H$ variations) can steer currents meridionally, acting as effective $\beta$-planes.
Eddy Kinetic Energy & Altimetric Detection
Eddy kinetic energy (EKE) is computed from the variance of velocity fluctuations:
$$\text{EKE} = \frac{1}{2}(u'^2 + v'^2)$$
where primes denote deviations from the time mean. Satellite altimeters (TOPEX/Poseidon, Jason, Sentinel-6) measure sea surface height (SSH) with ~2 cm accuracy, and geostrophic velocity anomalies are computed from SSH gradients:
$$u'_g = -\frac{g}{f}\frac{\partial \eta'}{\partial y}, \quad v'_g = \frac{g}{f}\frac{\partial \eta'}{\partial x}$$
EKE is highest in western boundary current regions (Gulf Stream, Kuroshio: ~500 cmΒ²/sΒ²), the ACC, and the Agulhas retroflection. Open-ocean EKE is typically 50β100 cmΒ²/sΒ².
Potential Vorticity & Quasi-Geostrophic Theory
Potential vorticity (PV) is conserved following fluid parcels in the absence of friction and diabatic processes. For a shallow water layer:
$$q = \frac{f + \zeta}{h}$$
where $\zeta = \partial v/\partial x - \partial u/\partial y$ is the relative vorticity and $h$ is the layer thickness. PV conservation $(Dq/Dt = 0)$ constrains fluid motion and underpins much of large-scale ocean dynamics.
Quasi-Geostrophic (QG) Vorticity Equation
The QG framework filters fast gravity waves and focuses on the slower, geostrophically balanced motions:
$$\frac{\partial}{\partial t}\nabla^2\psi + J(\psi, \nabla^2\psi) + \beta\frac{\partial\psi}{\partial x} = 0$$
where $\psi$ is the streamfunction and $J$ is the Jacobian (advection) operator. This equation supports Rossby waves (westward propagating) and describes eddyβmean flow interactions.
Turbulent Cascades
In 2-D and quasi-geostrophic turbulence, energy undergoes an inverse cascadeto larger scales (feeding the mean circulation), while enstrophy($\zeta^2$) cascades to smaller scales. This is opposite to 3-D turbulence and explains why ocean eddies tend to merge and grow. Submesoscale dynamics (1β10 km) involve a forward energy cascade and are important for vertical exchange and phytoplankton productivity.
Derivation: Rossby Wave Dispersion Relation
Step 1: Linearize the QG Vorticity Equation
Start from the quasi-geostrophic potential vorticity equation with no mean flow. Linearize about a state of rest by setting $\psi = \psi'$ (small perturbation):
$$\frac{\partial}{\partial t}\left(\nabla^2\psi' + \frac{f_0^2}{N^2}\frac{\partial^2\psi'}{\partial z^2}\right) + \beta\frac{\partial\psi'}{\partial x} = 0$$
Step 2: Assume Wave Solutions
Seek plane wave solutions of the form $\psi' = \hat{\psi}\,e^{i(kx + ly - \omega t)}\cos(mz)$, where $k$ and $l$ are horizontal wavenumbers and $m$ is the vertical wavenumber. Substituting:
$$-i\omega\left(-k^2 - l^2 - \frac{f_0^2 m^2}{N^2}\right)\hat{\psi} + ik\beta\hat{\psi} = 0$$
Step 3: Dispersion Relation
Solving for $\omega$ and defining the Rossby deformation radius $L_d = NH/(n\pi f_0)$ for vertical mode $n$:
$$\omega = \frac{-\beta k}{k^2 + l^2 + L_d^{-2}}$$
The phase speed in the $x$-direction is $c_x = \omega/k = -\beta/(k^2 + l^2 + L_d^{-2})$, which is always negative (westward). Long waves ($k^2 + l^2 \ll L_d^{-2}$) travel at the maximum speed $c_{max} = -\beta L_d^2$.
Step 4: Group Velocity
The group velocity determines energy propagation. Taking $\partial\omega/\partial k$:
$$c_{gx} = \frac{\partial\omega}{\partial k} = \frac{\beta(k^2 - l^2 - L_d^{-2})}{(k^2 + l^2 + L_d^{-2})^2}$$
For short waves ($k^2 > l^2 + L_d^{-2}$), the group velocity is eastward -- energy propagates opposite to phase. This has implications for how the ocean adjusts to wind changes: long Rossby waves carry the signal westward.
Derivation: Baroclinic Instability Criterion (Eady Model)
Step 1: QG Equations with a Mean Shear
Consider a mean zonal flow with constant vertical shear $\bar{U}(z) = \Lambda z$ (the Eady basic state). The linearized QG potential vorticity equation becomes:
$$\left(\frac{\partial}{\partial t} + \bar{U}\frac{\partial}{\partial x}\right)\left(\nabla^2\psi' + \frac{f_0^2}{N^2}\frac{\partial^2\psi'}{\partial z^2}\right) = 0$$
Note: the interior PV gradient vanishes in the Eady model (no $\beta$ term), and all the dynamics come from the boundary conditions.
Step 2: Boundary Conditions (Temperature)
At the upper ($z = H$) and lower ($z = 0$) rigid boundaries, the buoyancy equation provides:
$$\left(\frac{\partial}{\partial t} + \bar{U}\frac{\partial}{\partial x}\right)\frac{\partial\psi'}{\partial z} + \Lambda\frac{\partial\psi'}{\partial x} = 0 \quad \text{at } z = 0,\,H$$
Step 3: Normal Mode Analysis
Assume solutions $\psi' = \hat{\psi}(z)\,e^{ik(x - ct)}$. In the interior, the PV equation gives $\hat{\psi}(z) = A\cosh(\mu z) + B\sinh(\mu z)$ where $\mu = kN/(f_0)$. Applying the boundary conditions yields an eigenvalue problem for the phase speed $c$:
$$c = \frac{\Lambda H}{2} \pm \frac{\Lambda H}{2}\sqrt{\left(\frac{\tanh(\mu H/2)}{\mu H/2} - 1\right)\left(1 - \frac{\mu H/2}{\tanh(\mu H/2)}\right)}$$
Step 4: Instability Criterion
The expression under the square root is negative (giving complex $c$ and hence growing modes) when $\mu H < \mu_c H \approx 2.399$, i.e., when:
$$kL_d < \frac{2.399}{\pi} \approx 0.76 \quad \Longrightarrow \quad \lambda > 2\pi L_d / 0.76 \approx 8.3\,L_d$$
Disturbances with wavelengths longer than about $4L_d$ are baroclinically unstable. The Rossby deformation radius $L_d = NH/f_0$ thus sets the characteristic eddy scale.
Step 5: Maximum Growth Rate
The maximum growth rate occurs at $\mu H \approx 1.61$ and is:
$$\sigma_{max} = 0.31\,\frac{f_0}{N}\,\Lambda = 0.31\,\frac{f_0}{N}\frac{\partial\bar{U}}{\partial z} = 0.31\,\frac{|\nabla_H b|}{N}$$
For typical mid-latitude ocean conditions ($\Lambda \sim 10^{-3}$ s$^{-1}$, $N/f_0 \sim 20$), the e-folding time is $1/\sigma_{max} \sim 1$β2 weeks, explaining why mesoscale eddies develop on timescales of weeks to months.
Python: Rossby Deformation Radius, QG Vortex & Eddy Tracking
Python: Rossby Deformation Radius, QG Vortex & Eddy Tracking
Python!/usr/bin/env python3
Click Run to execute the Python code
Code will be executed with Python 3 on the server
Fortran: Barotropic Vorticity Equation Solver (Rossby Waves)
This program solves the barotropic vorticity equation on a $\beta$-plane to demonstrate the propagation of Rossby waves and the evolution of vortical structures.
Fortran: Barotropic Vorticity Equation Solver (Rossby Waves)
Fortran============================================================
Click Run to execute the Fortran code
Code will be compiled with gfortran and executed on the server
Eddy Transport & Notable Ring Systems
Mesoscale eddies play a crucial role in the global transport of heat, salt, nutrients, and biological organisms. The eddy-induced transport is parameterised in coarse-resolution models but explicitly resolved in eddy-resolving simulations. Key ring systems include:
Gulf Stream Rings
The Gulf Stream sheds ~15β20 rings per year as it meanders and pinches off. Warm-core rings (anticyclonic) form on the slope-water (northern) side, trapping Sargasso Sea water. Cold-core rings (cyclonic) form on the southern side, trapping cold, nutrient-rich slope water. These rings can persist for 1β2 years and transport significant quantities of heat ($\sim 10^{20}$ J per ring) and biota across the Gulf Stream front.
Agulhas Rings
Large anticyclonic eddies (~200β300 km diameter) shed from the Agulhas Current retroflection south of Africa. They carry warm, salty Indian Ocean water into the South Atlantic, contributing to the inter-ocean exchange that helps maintain the AMOC. Each ring transports ~0.5β1 Sv and ~0.045 PW of heat. About 5β6 rings are shed per year.
Submesoscale Dynamics (1β10 km)
Below the mesoscale, submesoscale fronts and eddies (O(1β10 km), hours to days) develop from frontogenesis and mixed-layer instabilities. These motions have Rossby numbers $Ro = \zeta/f \sim O(1)$, meaning they are only marginally geostrophic. Submesoscale dynamics drive intense vertical velocities (up to 100 m/day), which are critical for nutrient supply and phytoplankton bloom initiation. They bridge the gap between the geostrophic mesoscale and the fully three-dimensional turbulent microscale.
Key Equations Summary
Eady Growth Rate (Baroclinic Instability)
$$\sigma_{max} = 0.31\,\frac{f}{N}\frac{\partial U}{\partial z} = 0.31\,\frac{|\nabla_H b|}{N}$$
Maximum growth rate for baroclinic instability with e-folding time ~1β2 weeks for typical ocean conditions.
Rossby Wave Phase Speed
$$c_R = -\frac{\beta}{k^2 + l^2 + L_d^{-2}}$$
Always westward. Long waves are fastest; at mid-latitudes $c_R \sim 2$β5 cm/s.