Geostrophy, Dynamic Topography and Eddies
CSIR NET Ocean Sciences: geostrophic balance, thermal wind, the dynamic method, altimetry and dynamic topography, Rossby radius, mesoscale eddies and rings.
- 19 min
On this page · 16 sections
- Prerequisites
- The Hook: Measuring a Current Without a Current Meter
- Scale Analysis: When Is the Flow Geostrophic?
- Geostrophic Balance
- Thermal Wind
- The Dynamic Method
- Dynamic Topography and Satellite Altimetry
- The Rossby Radius of Deformation
- Mesoscale Eddies
- Memory Tricks
- Don’t Confuse
- Common Mistakes
- Practice
- Quick Revision Table
- Key Takeaways
- What’s Next
Prerequisites
- T-S Diagrams, Mixing and Water Masses: density, potential density and specific volume.
- Ocean Currents, Waves, Tides and Productivity: the Part B statement of geostrophic flow.
- The Coriolis parameter \( f = 2\Omega\sin\phi \), with \( \Omega = 7.292 \times 10^{-5} \) s\(^{-1}\).
- Atmospheric geostrophy works the same way; if you studied it, the ocean version is a change of symbols.
The Hook: Measuring a Current Without a Current Meter
Scale Analysis: When Is the Flow Geostrophic?
The horizontal momentum equations (per unit mass) are
$$\frac{Du}{Dt} - fv = -\frac{1}{\rho}\frac{\partial p}{\partial x} + F_x, \qquad \frac{Dv}{Dt} + fu = -\frac{1}{\rho}\frac{\partial p}{\partial y} + F_y$$Compare the acceleration term (\( U^2/L \)) with the Coriolis term (\( fU \)): their ratio is the Rossby number
$$\boxed{Ro = \frac{U}{fL}}$$When \( Ro \ll 1 \), and friction \( F \) is small (away from the surface and bottom boundary layers), Coriolis balances the pressure gradient.
| Flow | U (m s\(^{-1}\)) | L (km) | Ro (approx.) |
|---|---|---|---|
| Gyre interior | 0.05 | 1000 | \( 10^{-3} \) |
| Mesoscale eddy | 0.3 | 100 | 0.03 |
| Gulf Stream at 35°N | 1 | 100 | 0.1 |
| Tidal flow in a 10 km strait | 1 | 10 | 1: not geostrophic |
| Equator (\( f \to 0 \)) | any | any | large: geostrophy fails |
Geostrophic Balance
$$\boxed{fv = \frac{1}{\rho}\frac{\partial p}{\partial x}, \qquad fu = -\frac{1}{\rho}\frac{\partial p}{\partial y}}$$The flow runs along isobars, not down the pressure gradient. In the northern hemisphere, with your back to the flow, high pressure is on the right; in the southern hemisphere, on the left. Flow round a high is clockwise in the NH (anticyclonic) and anticlockwise in the SH (still anticyclonic).
Surface geostrophic current from sea-surface slope
At the surface, the pressure gradient is set by the slope of the sea surface \( \eta \): \( \frac{1}{\rho}\frac{\partial p}{\partial x} = g\frac{\partial \eta}{\partial x} \). So
$$\boxed{v_s = \frac{g}{f}\frac{\partial \eta}{\partial x}, \qquad u_s = -\frac{g}{f}\frac{\partial \eta}{\partial y}}$$Inertial oscillations: what happens without a pressure gradient
If the wind stops and there is no pressure gradient, \( Du/Dt = fv \), \( Dv/Dt = -fu \): the water turns in circles (clockwise in the NH) with the inertial period \( 2\pi/f \). At 30° latitude this is 23.93 h (one sidereal day ÷ \( 2\sin\phi \)); at 45° it is 16.9 h. Inertial oscillations are the largest signal in many current-meter records after a storm, and they show up in a spectrum at frequency \( f \).
Thermal Wind
Combine geostrophy with the hydrostatic equation \( \partial p/\partial z = -\rho g \) (Boussinesq, \( \rho_0 \) as reference):
$$\boxed{f\frac{\partial v}{\partial z} = -\frac{g}{\rho_0}\frac{\partial \rho}{\partial x}, \qquad f\frac{\partial u}{\partial z} = \frac{g}{\rho_0}\frac{\partial \rho}{\partial y}}$$The vertical shear of the geostrophic current is set by the horizontal density gradient. Rule (NH): looking downstream, if the current decreases with depth, the light (warm) water is on the right. For the Gulf Stream, the warm Sargasso Sea is on the right of the northeastward current and the flow weakens with depth.
Barotropic and baroclinic
| Barotropic | Baroclinic | |
|---|---|---|
| Density surfaces | parallel to pressure surfaces | inclined to pressure surfaces |
| Geostrophic velocity | same at all depths (no thermal wind shear) | varies with depth |
| What sets it | sea-surface slope alone | sea-surface slope plus internal density field |
In the real ocean the upper-ocean currents are largely baroclinic: the sea surface tilts one way, the thermocline tilts the other way, and the two pressure gradients nearly cancel at depth.
The reduced-gravity picture
Model the ocean as a light upper layer (density \( \rho_1 \), thickness \( h \)) over a deep, motionless layer (\( \rho_2 \)). For the deep layer to have no pressure gradient, a rise \( \eta \) of the sea surface must be compensated by a deepening of the interface:
$$\boxed{\Delta h = \frac{\rho_1}{\rho_2 - \rho_1}\,\Delta\eta \approx \frac{g}{g'}\,\Delta\eta}, \qquad g' = g\frac{\Delta\rho}{\rho_0}$$With \( \Delta\rho/\rho_0 = 0.002 \), a 0.5 m rise of sea level means the thermocline is about 250 m deeper. Sea surface and thermocline slope in opposite directions, the thermocline about 500 times more. This is why a sea-level high is also a thick, warm upper layer.
The Dynamic Method
Before altimetry, currents were computed from hydrographic stations.
Specific volume anomaly and dynamic height
The specific volume anomaly is \( \delta = \alpha(S, T, p) - \alpha(35, 0, p) \), the specific volume minus that of a standard ocean at the same pressure. The geopotential anomaly (dynamic height anomaly) of pressure surface \( p_1 \) relative to a deeper reference \( p_2 \) is
$$\Delta\Phi = \int_{p_1}^{p_2}\delta\, dp$$Units: J kg\(^{-1}\) = m\(^2\) s\(^{-2}\). One dynamic metre = 10 J kg\(^{-1}\), chosen so that a dynamic metre is close to a geometric metre of height (since \( g \approx 9.8 \)). Warm, fresh (light) water has large \( \delta \), so a light column stands taller.
Relative geostrophic velocity between two stations
For stations A and B a distance \( L \) apart, the geostrophic velocity normal to the section at \( p_1 \), relative to that at \( p_2 \), is
$$\boxed{v_1 - v_2 = \frac{\Delta\Phi_B - \Delta\Phi_A}{fL}}$$(This is the Helland-Hansen form.) Direction: in the NH, the flow at the upper level relative to the lower has the station with larger dynamic height on its right.
graph LR
A[CTD stations: S, T, p] --> B[Specific volume anomaly delta]
B --> C[Integrate to dynamic height relative to reference level]
C --> D[Difference between stations]
D --> E[Divide by f L: relative velocity]
E --> F{Know velocity at reference level?}
F -->|Assume zero: level of no motion| G[Absolute velocity, approximate]
F -->|Use altimetry, floats, LADCP| H[Absolute velocity]The reference-level problem
Hydrography gives only relative velocities. The classical fix was a level of no motion, a deep level assumed at rest (often 1000–2000 dbar, or the bottom). This works poorly where deep currents are strong (e.g. the Antarctic Circumpolar Current, deep western boundary currents). Modern fixes use direct velocity measurements (current meters, lowered ADCP, Argo float drift at parking depth) or satellite altimetry to supply the reference.
Dynamic Topography and Satellite Altimetry
Sea-surface height: what the altimeter sees
A radar altimeter measures the height of the sea surface above a reference ellipsoid to a few centimetres. That height has two parts:
$$\text{SSH} = N_{\text{geoid}} + \eta_{\text{dynamic}}$$| Component | Size | Cause |
|---|---|---|
| Geoid height \( N \) | tens of metres, up to about 100 m | the gravity field (a level surface); a resting ocean would follow it |
| Mean dynamic topography (MDT) | about 1–2 m peak to peak across the globe | time-mean geostrophic currents |
| Sea-level anomaly (SLA) | centimetres to tens of cm | eddies, seasonal cycle, ENSO, Rossby and Kelvin waves |
| Removed as corrections | — | tides, atmospheric pressure (inverse barometer), wet troposphere, ionosphere |
Dynamic topography is the height of the sea surface above the geoid. Because the geoid is an equipotential surface, slopes of dynamic topography (and only those) drive surface geostrophic currents.
- Sea-level anomaly (SSH minus its long-term mean at each point) needs no geoid; it shows the variable currents and eddies. This is what most altimeter maps show.
- Mean dynamic topography needs an accurate geoid at the scale of the currents; satellite gravity missions supply it. Mean sea surface minus geoid gives MDT.
Pattern of the mean dynamic topography
- Highs at the centres of the subtropical gyres (western sides especially), with clockwise flow round them in the NH and anticlockwise in the SH.
- Lows in the subpolar gyres and around Antarctica; the steepest gradient is across the Antarctic Circumpolar Current and the western boundary currents.
- The western tropical Pacific stands higher than the eastern, sustained by the trade winds; the gradient drives the Equatorial Undercurrent at depth (see the gyres and currents page).
The Rossby Radius of Deformation
The natural horizontal scale on which rotation balances stratification is the Rossby radius of deformation. For the first baroclinic mode in a two-layer approximation,
$$\boxed{R_d = \frac{\sqrt{g'H}}{|f|} = \frac{c}{|f|}}$$where \( c \) is the internal (baroclinic) gravity-wave speed, typically 1–3 m s\(^{-1}\).
| Latitude | \( R_d \) with \( c = 3.16 \) m s\(^{-1}\) |
|---|---|
| 10° | about 125 km |
| 30° | about 43 km |
| 45° | about 31 km |
\( R_d \) shrinks poleward because \( f \) grows. Features smaller than \( R_d \) feel mainly gravity; larger ones feel mainly rotation. Fronts, boundary currents and eddies have widths of order \( R_d \): the ocean’s “weather” is an order of magnitude smaller in scale than the atmosphere’s, because the ocean’s \( g' \) is so small.
Near the equator \( f \to 0 \), and the relevant scale is the equatorial deformation radius \( \sqrt{c/\beta} \), a few hundred kilometres (about 370 km for \( c = 3.16 \) m s\(^{-1}\)).
Long Rossby waves
The variation of \( f \) with latitude, \( \beta = \partial f/\partial y = 2\Omega\cos\phi/R_E \), supports Rossby waves. Long baroclinic Rossby waves propagate westward with speed
$$c_R \approx -\beta R_d^2$$At 30° (\( \beta = 1.98 \times 10^{-11} \) m\(^{-1}\) s\(^{-1}\), \( R_d = 43 \) km), \( c_R \approx 3.7 \) cm s\(^{-1}\), about 3 km per day: crossing a 5000 km basin takes about four years. They are faster near the equator. These waves carry the ocean’s adjustment to wind changes westward and set the time scale for gyre spin-up.
Mesoscale Eddies
Satellite altimetry showed that the ocean is full of eddies, with diameters of roughly 50–200 km (a few times \( R_d \)), lifetimes of weeks to more than a year, and current speeds often larger than the mean flow. Eddy kinetic energy exceeds mean kinetic energy over most of the ocean.
Formation
- Baroclinic instability: sloping isopycnals of a front or current release available potential energy; the fastest-growing disturbances have scales near \( R_d \). The dominant source.
- Barotropic instability: horizontal shear of a jet.
- Wind forcing, flow past islands and capes, and current meanders that pinch off (rings).
Cyclonic and anticyclonic eddies
| Anticyclonic eddy | Cyclonic eddy | |
|---|---|---|
| Rotation (NH) | clockwise | anticlockwise |
| Rotation (SH) | anticlockwise | clockwise |
| Sea-surface height | high (both hemispheres) | low (both hemispheres) |
| Thermocline / isopycnals | depressed (bowl) | domed (raised) |
| Core | warm-core | cold-core |
| Vertical motion in eddy pumping at formation/intensification | downwelling | upwelling |
| Biology | usually nutrient-poor centre | nutrients raised into the euphotic zone; often higher chlorophyll |
| Westward drift tendency | slight equatorward deflection | slight poleward deflection |
Gulf Stream rings
When a Gulf Stream meander pinches off:
- North of the stream it traps warm Sargasso water in the cold slope water: a warm-core ring, rotating clockwise (anticyclonic).
- South of the stream it traps cold slope water in the warm Sargasso Sea: a cold-core ring, rotating anticlockwise (cyclonic).
The same logic gives Kuroshio rings and Agulhas rings, large anticyclones shed at the Agulhas retroflection that carry warm, salty Indian Ocean water into the South Atlantic.
Subsurface eddies
Meddies are lens-shaped anticyclonic eddies of Mediterranean Water at about 1000 m in the North Atlantic, warmer and saltier than their surroundings, with almost no surface signal. Similar lenses of Red Sea Water occur in the Arabian Sea.
Eddies near India
- The Great Whirl, a large anticyclone off Somalia during the southwest monsoon (see the Somali Current section).
- The Lakshadweep (Laccadive) High and Low off southwest India: a sea-level high in the northeast monsoon and a low in the southwest monsoon, linked to coastally trapped Kelvin waves and westward Rossby waves.
- Many eddies in the Bay of Bengal, cyclonic ones raising nutrients and chlorophyll in an otherwise stratified bay.
Gradient balance in eddies
In a circular eddy the centrifugal term \( v^2/r \) can matter:
$$\frac{v^2}{r} + fv = g\frac{\partial\eta}{\partial r}$$For a given pressure gradient, flow around a low (cyclone) is slower than geostrophy predicts and flow around a high (anticyclone) is faster, because \( v^2/r \) opposes the pressure-gradient force in a cyclone and adds to it in an anticyclone. An anticyclone also has a maximum possible pressure gradient, \( g\,\partial\eta/\partial r \le f^2 r/4 \). For most ocean eddies \( Ro \) is small and plain geostrophy is a good first estimate.
Detecting eddies
Eddies are tracked in SLA maps as closed contours; the Okubo–Weiss parameter \( W = s_n^2 + s_s^2 - \omega^2 \) (strain squared minus relative vorticity squared) is negative inside eddy cores, where rotation dominates strain. Drifters looping round a centre and Argo floats trapped in an eddy confirm them.
Memory Tricks
Don’t Confuse
| This | vs | That |
|---|---|---|
| Geoid: equipotential surface, tens of metres of relief | Dynamic topography: SSH above the geoid, about 1–2 m | |
| Sea-level anomaly: variable part, no geoid needed | Mean dynamic topography: needs a geoid | |
| Dynamic height: relative to a reference pressure | Absolute velocity: needs a reference velocity | |
| Barotropic: velocity uniform with depth | Baroclinic: shear set by density gradient | |
| Inertial period \( 2\pi/f \): 23.93 h at 30° | Sidereal day \( 2\pi/\Omega \): 23.93 h at every latitude | |
| Warm-core ring north of the Gulf Stream, anticyclonic | Cold-core ring south of it, cyclonic | |
| Rossby number \( U/fL \) | Rossby radius \( c/f \) |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not from any CSIR NET paper.
Level 1: Foundation
- State the geostrophic relation between surface current and sea-surface slope.
- What is the inertial period at the North Pole (\( \phi = 90° \))?
- Is an SSH high in the southern hemisphere a cyclonic or anticyclonic eddy?
Answers
- \( v_s = (g/f)\,\partial\eta/\partial x \), \( u_s = -(g/f)\,\partial\eta/\partial y \).
- About 11.97 h: \( 2\pi/(2\Omega) \), half a sidereal day.
- Anticyclonic (rotating anticlockwise in the SH).
Level 2: Part B Style
Practice Q1. At 30°N, sea level rises by 20 cm toward the east over 100 km. The surface geostrophic current is about:
(a) 0.27 m s\(^{-1}\) northward (b) 0.27 m s\(^{-1}\) southward (c) 0.13 m s\(^{-1}\) eastward (d) 0.27 m s\(^{-1}\) westward
Answer: (a)
\( v = \frac{g}{f}\frac{\partial\eta}{\partial x} = \frac{9.81}{7.29\times 10^{-5}}\times\frac{0.2}{10^5} = 0.269 \) m s\(^{-1}\), positive: northward. Check with the rule: facing north, the high (east) is on the right.
Practice Q2. In a two-layer ocean with \( \Delta\rho/\rho_0 = 0.0025 \), the sea surface rises by 0.3 m across a current. The interface (thermocline) moves:
(a) up by 120 m (b) down by 120 m (c) down by 75 m (d) down by 0.75 m
Answer: (b) down by 120 m
\( \Delta h = \Delta\eta/(\Delta\rho/\rho_0) = 0.3/0.0025 = 120 \) m, and it deepens where the surface rises, so the lower layer feels no pressure gradient.
Practice Q3. A ring forms north of the Gulf Stream by a meander pinching off. It is:
(a) cold-core, cyclonic (b) warm-core, anticyclonic (c) warm-core, cyclonic (d) cold-core, anticyclonic
Answer: (b)
A northward meander loops around Sargasso water and pinches off into the cold slope water: warm core, clockwise rotation (anticyclonic in the NH), SSH high.
Practice Q4. If the baroclinic gravity-wave speed is 2.5 m s\(^{-1}\) everywhere, the first baroclinic Rossby radius at 60° latitude compared with that at 15° latitude is about:
(a) 0.30 times (b) 0.5 times (c) 2 times (d) 3.3 times
Answer: (a)
\( R_d \propto 1/\sin\phi \): \( \sin 15°/\sin 60° = 0.259/0.866 = 0.30 \).
Level 3: Part C Style
Practice Q5. Two hydrographic stations, A (west) and B (east), are 80 km apart at 25°N. The dynamic height of the sea surface relative to 1500 dbar is 1.85 dyn m at A and 1.61 dyn m at B. Assuming 1500 dbar is a level of no motion:
- The surface geostrophic velocity normal to the section is about 0.49 m s\(^{-1}\).
- The flow is southward.
- If the true velocity at 1500 dbar were 0.05 m s\(^{-1}\) northward, the true surface velocity would be about 0.44 m s\(^{-1}\) southward.
- The lighter upper-ocean water lies to the east.
Which are correct?
(a) 1, 2 and 3 only (b) 1 and 4 only (c) 2 and 4 only (d) 1, 2, 3 and 4
Answer: (a) 1, 2 and 3
\( \Delta\Phi = (1.61 - 1.85) \times 10 = -2.4 \) J kg\(^{-1}\) (B minus A). \( f = 2\Omega\sin 25° = 6.16\times 10^{-5} \) s\(^{-1}\).
\( v = \frac{-2.4}{6.16\times 10^{-5}\times 8\times 10^4} = -0.487 \) m s\(^{-1}\): about 0.49 m s\(^{-1}\), southward (1, 2 correct). Rule check: facing south, west (A, the higher station) is on the right. Consistent.
3: The computed velocity is relative to 1500 dbar. Adding a northward 0.05: \( -0.487 + 0.05 = -0.437 \), about 0.44 southward. Correct.
4: Larger dynamic height means a lighter (warmer or fresher) column, so the light water is at A, to the west. Wrong.
Practice Q6. An altimeter map at 20°N shows a closed circular SSH anomaly of +25 cm at the centre, falling to 0 at a radius of 60 km. Argo profiles inside show the 20 °C isotherm 60 m deeper than outside.
- The eddy is anticyclonic and rotates clockwise.
- The average geostrophic speed between centre and edge is about 0.8 m s\(^{-1}\).
- The deeper isotherm is consistent with a warm-core eddy.
- A sea-level rise of 25 cm with a 60 m isotherm deepening implies \( \Delta\rho/\rho_0 \) of about 0.004 in a reduced-gravity model.
Which are correct?
(a) 1 and 3 only (b) 1, 2 and 3 only (c) 1, 2, 3 and 4 (d) 2 and 4 only
Answer: (c) all four
- SSH high → anticyclone → clockwise in the NH. Correct.
- \( f = 2\Omega\sin 20° = 4.99\times 10^{-5} \) s\(^{-1}\). \( v = \frac{9.81}{4.99\times 10^{-5}}\times\frac{0.25}{6\times 10^4} = 0.82 \) m s\(^{-1}\). Correct.
- Depressed isotherms: warm core. Correct.
- \( \Delta\rho/\rho_0 = \Delta\eta/\Delta h = 0.25/60 = 0.0042 \). Correct. (A real eddy has continuous stratification, but the two-layer estimate is the expected Part C reasoning.)
Practice Q7. Using \( c = 2.8 \) m s\(^{-1}\), \( \Omega = 7.292\times 10^{-5} \) s\(^{-1}\) and Earth radius 6371 km, how long does a long first-mode baroclinic Rossby wave take to cross 4000 km at 20°N?
(a) about 5 months (b) about 1 year (c) about 1.9 years (d) about 9 years
Answer: (c) about 1.9 years
\( f = 2\Omega\sin 20° = 4.99\times 10^{-5} \) s\(^{-1}\); \( R_d = 2.8/4.99\times 10^{-5} = 56.1 \) km. \( \beta = 2\Omega\cos 20°/R_E = 2.15\times 10^{-11} \) m\(^{-1}\) s\(^{-1}\).
\( c_R = \beta R_d^2 = 2.15\times 10^{-11}\times(5.61\times 10^4)^2 = 0.0677 \) m s\(^{-1}\), about 5.9 km per day.
Time \( = 4\times 10^6/0.0677 = 5.9\times 10^7 \) s; dividing by \( 3.156\times 10^7 \) s per year gives 1.9 years.
The point for Part C: baroclinic adjustment of a basin takes years, and is faster at low latitudes.
Quick Revision Table
| Idea | Formula or fact |
|---|---|
| Rossby number | \( Ro = U/(fL) \); geostrophic if \( Ro \ll 1 \) |
| Geostrophy | \( fv = \rho^{-1}\partial p/\partial x \); high pressure on right (NH) |
| Surface current | \( v_s = (g/f)\,\partial\eta/\partial x \) |
| Inertial period | \( 2\pi/f \); 23.93 h at 30° |
| Thermal wind | \( f\,\partial v/\partial z = -(g/\rho_0)\,\partial\rho/\partial x \) |
| Reduced gravity | interface moves \( \Delta\eta/(\Delta\rho/\rho_0) \), opposite sign |
| Dynamic height | \( \Delta\Phi = \int\delta\,dp \); 1 dyn m = 10 J kg\(^{-1}\) |
| Relative velocity | \( v_1 - v_2 = (\Delta\Phi_B - \Delta\Phi_A)/(fL) \) |
| SSH | geoid (tens of m) + dynamic topography (1–2 m) |
| Rossby radius | \( R_d = c/f \); tens of km at mid-latitudes |
| Rossby wave | westward, \( c_R = \beta R_d^2 \) |
| Anticyclonic eddy | SSH high, warm core, depressed thermocline |
| Cyclonic eddy | SSH low, cold core, domed thermocline, upwelling |
| Gulf Stream rings | warm-core north, cold-core south |
Key Takeaways
What’s Next
- Wind-Driven Circulation: Ekman, Sverdrup, Stommel and Munk: what drives the geostrophic interior and why boundary currents are on the west.
- Gyres, Boundary and Equatorial Currents, Somali Current: the observed currents.
- Back to the Physical Oceanography overview.