Coriolis effects on horizontal motion

In September 1969, oceanographers tracking a free-floating buoy in the Baltic Sea watched it do something no current should do: instead of drifting in a line, it looped — fourteen graceful, nearly perfect circles, two kilometres across, one loop every fourteen hours, like a stylus tracing a spirograph on the sea. Nothing was steering it. No vortex, no eddy, no whirlpool. The buoy was simply moving over a rotating planet, and the Coriolis force — forever pushing sideways, never forward or back — bent its path into the purest orbit in geophysics: the inertial circle. That looping buoy is the cleanest showcase of what this article isolates: the Coriolis force projected onto the local horizontal plane. The Coriolis force article derives the full three-dimensional force $-2m\,\boldsymbol{\Omega}\times\mathbf{v}$ and its famous factor of 2; here we ask what it does to motion that stays flat — winds, currents, pucks, shells, rivers. The answer compresses into a single scalar, the Coriolis parameter $f = 2\Omega\sin\varphi$ : every horizontally moving body at latitude $\varphi$ has its velocity rotated at rate $f$ , clockwise in the Northern Hemisphere, counterclockwise in the Southern, regardless of which way it moves. One dial, three regimes. Fast and brief (a thrown ball): deflections of millimetres, ignore it. Intermediate (artillery shells, long sniper shots): metres — gunnery tables carry latitude corrections. Slow, vast and patient (the atmosphere, the ocean): the Coriolis force graduates from correction to ruler, locking winds along isobars in the geostrophic balance that organizes every weather map you have ever seen.

Key takeaways

For motion in the local horizontal plane, the entire three-dimensional Coriolis machinery collapses to one scalar: the Coriolis parameter $f = 2\Omega\sin\varphi$ , the rate at which the horizontal velocity vector is rotated.
The horizontal equations are $\dot{u} = f v$ , $\dot{v} = -f u$ : a pure rotation of the velocity — speed never changes, direction turns clockwise (Northern Hemisphere) at rate $f$ , rightward of the motion whatever the heading.
A free body loops around an inertial circle of radius $R = v/f$ and period $T_f = 2\pi/f = T_{\mathrm{sid}}/(2\sin\varphi)$ — the "half-pendulum-day" — exactly like a charge gyrating in a magnetic field.
Short flights deflect sideways by $\Delta s \approx \tfrac{1}{2} f v t^2$ : 6 mm for a thrown ball, tens of metres for long-range artillery.
When the Rossby number $\mathrm{Ro} = U/fL$ drops below 1, rotation rules: the geostrophic balance $f\,\hat{\mathbf{z}}\times\mathbf{v}_g = -\nabla_h p/\rho$ makes winds blow along isobars with low pressure on the left (N).
$f$ vanishes at the equator and is maximal at the poles — which is why hurricanes shun the equatorial band and why polar inertial circles close in 11.97 hours.
Key idea: project the rotating-frame equation onto the local horizontal plane and only $\Omega\sin\varphi$ survives — horizontal motion obeys $\dot{\mathbf{u}} = -f\,\hat{\mathbf{z}}\times\mathbf{u} + \mathbf{a}_{\mathrm{other}}$ with $f = 2\Omega\sin\varphi$ , a rotation of the velocity at rate $f$ , rightward in the North.

The core insight

Project the rotating-frame equation onto the local horizontal plane: only $\Omega\sin\varphi$ survives, so horizontal motion obeys $\dot{\mathbf{u}} = -f\,\hat{\mathbf{z}}\times\mathbf{u}+\mathbf{a}_{\mathrm{other}}$ with $f=2\Omega\sin\varphi$ — a rotation of the velocity at rate $f$ , rightward in the North.

Where this lives today

Every operational weather forecast is an integration of equation (3) with the pressure, friction, and moisture terms filled in — the horizontal Coriolis terms $fv$ and $-fu$ sit verbatim in the dynamical cores of ECMWF's IFS and the new kilometre-scale digital-twin models (DestinE) running in 2026. The geostrophic balance is so reliable above the friction layer that it works in reverse: satellites that only measure the height of the sea surface (the SWOT altimetry mission, flying since 2022, maps it to centimetres) hand oceanographers global current charts, because tilting sea surface means horizontal pressure gradient means geostrophic current, computable from $v_g = (g/f)\,\partial\eta/\partial n$ at every point away from the equator. The world's ocean-current atlases are, in a precise sense, photographs of the Coriolis balance. The inertial circle also earns its keep. Windfarm-wake meteorology must resolve inertial oscillations in the nocturnal boundary layer; hurricane-intensity models track how close the storm's Rossby number sits to unity as it spins up; and every drifting Argo float and iceberg-tracking buoy superposes a $2\pi/f$ wiggle on its mean drift that data pipelines routinely fit and remove — using exactly equation (6), latitude by latitude, as the buoys ride the currents of a rotating planet.

The physical picture

Stand at the North Pole and slide a puck across frictionless ice toward a flag planted in the distance. In inertial space the puck's physics is trivial — a dead-straight line. But the ice, the flag, and you rotate counterclockwise underneath it, once per sidereal day. Relative to the ground the puck therefore appears to curve rightward, missing the flag to its right. Now walk south to latitude $\varphi$ : the ground still rotates under the puck, but only the component of Earth's spin about your local vertical — the fraction $\sin\varphi$ of it — rotates your horizontal world within itself. The apparent turning rate of every horizontal velocity drops to $f = 2\Omega\sin\varphi$ .

Figure

The two faces of horizontal Coriolis motion — a deflected shot and a closed inertial circle

Because the Coriolis force is always perpendicular to the velocity, it can turn the velocity but never change the speed. Free horizontal motion is therefore uniform circular motion in the rotating frame — the inertial circle — in exact mathematical analogy with a charged particle gyrating in a uniform magnetic field. The Coriolis parameter $f$ plays the cyclotron frequency $qB/m$ , and the local vertical $\hat{\mathbf{z}}$ plays the field direction. Everything you know about cyclotron orbits transfers: radius proportional to speed, period independent of speed, sense of rotation fixed by the sign of the "field" — which here flips at the equator. The factor $\sin\varphi$ is pure geometry, and it is worth seeing without algebra. Earth's angular velocity vector $\boldsymbol{\Omega}$ points out of the North Pole. At latitude $\varphi$ it makes an angle $90°-\varphi$ with the local vertical, so it splits into a vertical piece $\Omega\sin\varphi$ — which spins your horizontal plane within itself like a slow turntable — and a horizontal piece $\Omega\cos\varphi$ pointing north, which tips the plane over toward tomorrow but cannot rotate anything within it. Horizontal deflection of horizontal motion reads only the turntable part. (The $\cos\varphi$ part couples horizontal and vertical motion — the eastward deflection of falling bodies and the Eötvös effect — which the Coriolis force article treats; here it is a spectator.) Magnitudes set the regime, and they are worth memorizing as reflexes. At mid-latitudes $f \approx 10^{-4}\ \mathrm{s^{-1}}$ : a full turn of the velocity takes $2\pi/f \approx 17$ hours. A footballer's 30 m/s pass, airborne for 2 seconds, drifts sideways $\tfrac12 f v t^2 \approx 6$ millimetres — invisible beneath spin and wind. A 10 m/s breeze, left alone for a day, would loop a circle 100 km across. The Coriolis force never competes on strength; it competes on stamina. Slow, large, persistent flows have no choice but to obey it.

The horizontal equations

Start from the rotating-frame equation of motion with the centrifugal term already absorbed into effective gravity (that is the standing convention for Earth-frame mechanics — see the effective-gravity article):

\mathbf{a}_{\mathrm{rot}} = \mathbf{g}_{\mathrm{eff}} - 2\,\boldsymbol{\Omega}\times\mathbf{v}_{\mathrm{rot}} + \frac{\mathbf{F}_{\mathrm{other}}}{m} \qquad (1)

where: $\mathbf{a}_{\mathrm{rot}}, \mathbf{v}_{\mathrm{rot}}$ — acceleration and velocity measured in the ground frame; $\mathbf{g}_{\mathrm{eff}}$ — true gravity plus centrifugal acceleration, defining the local plumb-line vertical; $\mathbf{F}_{\mathrm{other}}$ — genuine forces (pressure gradients, friction, normal forces). The Euler term is absent because $\dot{\boldsymbol{\Omega}} \approx 0$ for Earth. Everything horizontal in this article flows from projecting equation (1) onto the local level plane.

Derivation AProjecting the Coriolis force onto the local horizontal plane

We are deriving the horizontal equations of motion (3) — the entire dynamical content of this article — by resolving the full three-dimensional Coriolis acceleration $-2\boldsymbol{\Omega}\times\mathbf{v}$ in a local east-north-up basis and keeping only the in-plane components. The strategy is brute-force but transparent: write $\boldsymbol{\Omega}$ in the local basis, expand one cross product, and sort the six resulting terms by what they do. The payoff is seeing exactly where the $\sin\varphi$ goes and where the $\cos\varphi$ terms hide. The step to watch is the last one: dropping the vertical-velocity terms is a physical approximation about the motion being level, not an algebraic identity — for strongly vertical motion those discarded terms are the whole story.

Step 1. Set up the local basis at latitude $\varphi$ : $\hat{\mathbf{x}}$ east, $\hat{\mathbf{y}}$ north, $\hat{\mathbf{z}}$ up. Earth's spin axis lies in the north–up plane, so

\boldsymbol{\Omega} = \Omega\left(\cos\varphi\,\hat{\mathbf{y}} + \sin\varphi\,\hat{\mathbf{z}}\right)

Why no east component: the angular velocity vector points from the South Pole to the North Pole; viewed from any point on the surface that direction has a northward tilt and an upward tilt, but no east–west lean.

Step 2. Expand the Coriolis acceleration for a general velocity $\mathbf{v} = u\,\hat{\mathbf{x}} + v\,\hat{\mathbf{y}} + w\,\hat{\mathbf{z}}$ :

-2\boldsymbol{\Omega}\times\mathbf{v} = -2\Omega\left[(w\cos\varphi - v\sin\varphi)\,\hat{\mathbf{x}} + u\sin\varphi\,\hat{\mathbf{y}} - u\cos\varphi\,\hat{\mathbf{z}}\right]

Why six terms collapse to four: the basis cross products $\hat{\mathbf{y}}\times\hat{\mathbf{x}} = -\hat{\mathbf{z}}$ , $\hat{\mathbf{y}}\times\hat{\mathbf{z}} = \hat{\mathbf{x}}$ , $\hat{\mathbf{z}}\times\hat{\mathbf{x}} = \hat{\mathbf{y}}$ , $\hat{\mathbf{z}}\times\hat{\mathbf{y}} = -\hat{\mathbf{x}}$ kill the parallel pairings.

Step 3. Sort the terms by their job. The $\sin\varphi$ terms take horizontal velocity to horizontal acceleration — in-plane deflection. The $w\cos\varphi$ term takes vertical velocity to horizontal (eastward) acceleration — that is the falling-stone deflection. The $u\cos\varphi$ term takes eastward velocity to vertical acceleration — the Eötvös effect that makes eastbound ships fractionally lighter. Neither $\cos\varphi$ term deflects horizontal motion within the plane.

Step 4. Specialize to level motion, $w \approx 0$ , and read off the horizontal components:

a_x = 2\Omega\sin\varphi\,v = f\,v, \qquad a_y = -2\Omega\sin\varphi\,u = -f\,u \qquad (2)

where: $f \equiv 2\Omega\sin\varphi$ — the Coriolis parameter, the single scalar that survives the projection. In vector form the pair reads $\mathbf{a}_h = -f\,\hat{\mathbf{z}}\times\mathbf{v}_h$ : a rotation generator acting on the velocity.

Step 5. Append the other horizontal forces to get the working equation of this whole subject:

\dot{u} = f\,v + \frac{F_x}{m}, \qquad \dot{v} = -f\,u + \frac{F_y}{m} \qquad (3)

where: $u, v$ — eastward and northward velocity components; $F_{x,y}$ — real horizontal forces per unit mass (pressure gradient, friction). With no real forces the motion is a pure velocity rotation; with steady forces the rotation competes against them — both regimes are solved below. Verified: at the pole ( $\varphi = 90°$ ) the $\cos\varphi$ cross-couplings vanish and $f = 2\Omega$ — pure in-plane looping; at the equator $f = 0$ and horizontal motion feels no in-plane deflection. Matches Goldstein §4.10 and Marion–Thornton ch. 10.

The Coriolis parameter deserves its own display, because meteorology and oceanography treat it as a fundamental constant of the local sky:

f = 2\,\Omega\sin\varphi \qquad (4)

where: $\Omega = 7.292\times10^{-5}\ \mathrm{rad/s}$ — Earth's sidereal spin rate (one turn per 86,164 s); $\varphi$ — geographic latitude, positive north. The values to carry in your head: $f = 1.03\times10^{-4}\ \mathrm{s^{-1}}$ at $45°$ , $0.73\times10^{-4}$ at $30°$ , $1.46\times10^{-4}$ at the pole, and exactly zero at the equator — with a sign flip across it that reverses every handedness in this article. Meteorologists call $f$ the planetary vorticity: it is the spin about the local vertical that the planet hands, free of charge, to every parcel of air sitting still on the ground.

Free motion: inertial circles

With no real horizontal force, equation (3) says the velocity vector rotates rigidly. The cleanest way to see it — and to solve it in two lines — is the complex-velocity trick that this corner of mechanics shares with the Foucault pendulum analysis.

Go deeper · Derivation B · 2 minInertial circles from the constant parameter horizontal equations

Derivation BInertial circles from the constant parameter horizontal equations

The goal is the complete trajectory of a free particle under horizontal Coriolis deflection alone: we will show the path is a closed circle of radius $R = v_0/f$ traversed clockwise (Northern Hemisphere) with period $T_f = 2\pi/f$ , independent of speed. The strategy is to pack the two coupled real equations into one complex equation for $w = u + iv$ , which turns "rotate a vector at rate $f$ " into "multiply by $e^{-ift}$ " — and then to integrate once more for position. Watch the sign in the exponent: it encodes the clockwise sense, and it flips with the hemisphere.

Step 1. Combine the free equations $\dot{u} = fv$ , $\dot{v} = -fu$ into one. Multiply the second by $i$ and add:

\frac{\mathrm{d}}{\mathrm{d}t}(u + iv) = f v - i f u = -i f\,(u + iv)

Why this works: multiplication by $-i$ is a 90° clockwise rotation in the plane, which is precisely what the force $-f\hat{\mathbf{z}}\times\mathbf{v}$ does to the velocity.

Step 2. Solve the first-order equation with $w = u+iv$ and initial speed $v_0$ at heading $\theta_0$ :

w(t) = v_0\,e^{i(\theta_0 - f t)} \qquad (5)

where: the modulus $|w| = v_0$ is constant — speed is conserved, as it must be for a force that does no work — while the heading decreases at the uniform rate $f$ : the velocity compass needle sweeps clockwise.

Step 3. Integrate once for position $\zeta = x + iy$ :

\zeta(t) = \zeta_0 + \frac{v_0}{-i f}\left(e^{i(\theta_0 - ft)} - e^{i\theta_0}\right)

Why the geometry is now readable: $\zeta(t)$ traces a circle of radius $|v_0/f|$ whose centre $\zeta_c = \zeta_0 + (v_0/(-if))\,(-e^{i\theta_0}) = \zeta_0 + (v_0/f)\,e^{i(\theta_0 - \pi/2)}$ sits at distance $v_0/f$ exactly 90° to the RIGHT of the launch velocity — the body forever falls toward a point beside it.

Step 4. Read off the orbit parameters:

R = \frac{v_0}{f}, \qquad T_f = \frac{2\pi}{f} = \frac{T_{\mathrm{sid}}}{2\sin\varphi} \qquad (6)

where: $R$ — inertial-circle radius, proportional to speed; $T_f$ — the inertial period, independent of speed, only of latitude. $T_f$ is called the half-pendulum-day: the Foucault pendulum's precession period is $T_{\mathrm{sid}}/\sin\varphi$ , and the inertial circle closes in exactly half that. At $45°$ : a 10 m/s drift gives $R = 10/(1.03\times10^{-4}) \approx 97$ km and $T_f \approx 16.9$ h.

Step 5. Check the polar limit. At $\varphi = 90°$ , $T_f = \pi/\Omega = 43{,}082\ \mathrm{s} = 11.97$ h — half a sidereal day. In inertial space the puck moves in a straight line while the ground turns under it at $\Omega$ ; the relative motion closes its loop in half the rotation period because both the velocity rotation ( $-2\Omega$ from Coriolis in the frame) and the frame's own sweep conspire — the same factor-of-2 bookkeeping that lives inside $f$ itself. Verified: identical structure to a cyclotron orbit $\dot{\mathbf{v}} = (q/m)\,\mathbf{v}\times\mathbf{B}$ with $f \leftrightarrow qB/m$ ; speed conservation follows from $\mathbf{F}\perp\mathbf{v}$ . Numbers match Marion–Thornton ch. 10 examples.

Inertial circles are not a textbook idealization — the ocean actually draws them. After a storm dumps momentum into the surface mixed layer, moored current meters and drifting buoys routinely record velocity vectors sweeping clockwise (N) with periods matching $2\pi/f$ at the mooring's latitude to within a percent. The famous 1969 Baltic record shows fourteen consecutive loops. Meteorological balloons show the same signature in nocturnal jets. When a geophysicist sees a velocity spectrum with a sharp peak at frequency $f$ , she reads it instantly: something kicked the fluid, and the planet has been steering it in circles ever since. For short times ( $ft \ll 1$ ) the circle is invisible; only its first sliver matters. Expanding the exact solution to second order in $ft$ gives the workhorse deflection formula:

\Delta s \approx \tfrac{1}{2}\,f\,v\,t^{2} \qquad (7)

where: $\Delta s$ — sideways displacement (to the right in the north) after time $t$ at speed $v$ . It is just "constant acceleration $fv$ for a time $t$ " — the Coriolis acceleration barely rotates during a short flight, so it acts like a steady sideways push. This one formula answers every ballistics question below.

Which formula answers which question

you want	use	notes
the turning rate of any horizontal velocity	$f = 2\Omega\sin\varphi$ , eq. (4)	zero at equator, sign flips south
the equations of motion on the level plane	eq. (3)	add pressure/friction as needed
the free trajectory	inertial circle, eqs. (5)–(6)	$R = v/f$ , period $2\pi/f$ , speed-independent
sideways drift of a short flight	$\Delta s = \tfrac12 f v t^2$ , eq. (7)	valid while $ft \ll 1$
wind speed from a pressure map	$v_g = \|\partial p/\partial n\| / (\rho f)$ , eq. (9)	the geostrophic balance
does rotation matter at all?	$\mathrm{Ro} = U/fL$ , eq. (10)	rotation rules iff $\mathrm{Ro} \lesssim 1$

Forced motion: the geostrophic balance

Now put a steady real force into equation (3) and wait. A pressure gradient pushes air from high pressure toward low; the Coriolis force deflects the resulting motion rightward; the deflection continues until — remarkably — the air flows perpendicular to the push, with the Coriolis force exactly cancelling the pressure force. Nothing then accelerates, and the pattern can persist for days. This is the geostrophic balance, the single most important force balance on any weather map.

Figure

Geostrophic balance — the wind blows along the isobars, not down the pressure hill

Go deeper · Derivation C · 2 minThe geostrophic wind from a pressure map

Derivation CThe geostrophic wind from a pressure map

The goal is equation (9), the geostrophic wind speed — the formula by which a forecaster reads wind velocity straight off the isobar spacing of a surface chart. The strategy is a scale analysis followed by a two-line balance: we first show that for synoptic-scale flow the acceleration term in (3) is an order of magnitude smaller than the Coriolis and pressure terms (that is the statement $\mathrm{Ro}\ll1$ ), then set it to zero and solve. Watch the direction conclusion — it is more surprising than the magnitude: air does not flow from high to low pressure; it flows along the isobars, endlessly sideways to the push.

Step 1. Estimate the sizes of the three terms in (3) for a mid-latitude weather system: speed $U \sim 10\ \mathrm{m/s}$ , length scale $L \sim 1000\ \mathrm{km}$ , $f \sim 10^{-4}\ \mathrm{s^{-1}}$ . The acceleration term is $\mathrm{d}v/\mathrm{d}t \sim U^2/L \sim 10^{-4}\ \mathrm{m/s^2}$ ; the Coriolis term is $fU \sim 10^{-3}\ \mathrm{m/s^2}$ . Their ratio is the Rossby number:

\mathrm{Ro} = \frac{U}{f L} \sim 0.1

Why this justifies the balance: with $\mathrm{Ro}\sim0.1$ the inertia of the air is a 10% correction; to leading order the only players left standing are the Coriolis force and whatever real force is applied — here the pressure gradient.

Step 2. Set the acceleration in (3) to zero and keep the pressure-gradient force per unit mass, $-\nabla_h p/\rho$ :

0 = -\frac{1}{\rho}\nabla_h p - f\,\hat{\mathbf{z}}\times\mathbf{v}_g \qquad (8)

where: $\mathbf{v}_g$ — the geostrophic velocity, defined as the velocity that makes this balance exact. Cross both sides with $\hat{\mathbf{z}}$ to solve: $\mathbf{v}_g = (1/\rho f)\,\hat{\mathbf{z}}\times\nabla_h p$ — the velocity is perpendicular to the gradient, i.e. along the isobars, with low pressure on the left when $f>0$ .

Step 3. Take magnitudes across the isobars ( $n$ = direction of steepest pressure change):

v_g = \frac{1}{\rho f}\left|\frac{\partial p}{\partial n}\right| \qquad (9)

where: $\rho$ — air density $\approx 1.2\ \mathrm{kg/m^3}$ near the surface. Closer isobars mean stronger wind, inversely weighted by latitude through $f$ — the same pressure map implies fiercer winds nearer the equator, until the balance itself fails as $f\to0$ .

Step 4. Put in a real map's numbers: a healthy mid-latitude gradient of $3\ \mathrm{hPa}$ per $100\ \mathrm{km}$ is $\partial p/\partial n = 3\times10^{-3}\ \mathrm{Pa/m}$ . With $f = 10^{-4}\ \mathrm{s^{-1}}$ :

v_g = \frac{3\times10^{-3}}{1.2 \times 10^{-4}} = 25\ \mathrm{m/s}

Why this is believable: 25 m/s ≈ 90 km/h is a typical free-atmosphere wind above the friction layer — surface winds are slower and cross the isobars slightly because friction breaks the pure balance. *Verified: direction convention (low on the left, N. Hemisphere) reproduces Buys Ballot's law and observed counterclockwise flow around cyclones; magnitudes match Holton's dynamic-meteorology worked values.*

The Rossby number introduced in Step 1 deserves promotion to a law of its own, because it is the single dial that decides every "does Coriolis matter?" argument:

\mathrm{Ro} = \frac{U}{f\,L} \qquad (10)

where: $U$ — characteristic speed, $L$ — characteristic size of the flow. $\mathrm{Ro} \gg 1$ : inertia wins, rotation is a perturbation (balls, rivers on small scales, tornadoes — a tornado's spin comes from stretching pre-existing synoptic vorticity, not from local Coriolis force). $\mathrm{Ro} \lesssim 1$ : rotation organizes the flow (hurricanes, $\mathrm{Ro}\sim1$ ; weather systems, $\mathrm{Ro}\sim0.1$ ; ocean gyres, $\mathrm{Ro}\sim0.001$ ).

Figure

The Rossby ladder — who obeys the Coriolis force and who ignores it

Two consequences of the balance run the planet's plumbing. Cyclones and anticyclones: around a northern low, the inward pressure force is balanced by outward-deflecting Coriolis force only if air circulates counterclockwise; around a high, clockwise — and both flip south of the equator. Because $f\to0$ at the equator, the balance is impossible there: tropical cyclones cannot organize within about $5°$ of the equator and never cross it. Ocean gyres: wind stress plus the latitude dependence of $f$ — the beta effect, $\beta = \mathrm{d}f/\mathrm{d}y = 2\Omega\cos\varphi/R_E \approx 1.6\times10^{-11}\ \mathrm{m^{-1}s^{-1}}$ at $45°$ — squeezes the return flow of each basin into a narrow, fast western boundary current. The Gulf Stream and the Kuroshio owe their existence, position, and violence to the fact that $f$ grows with latitude. One more horizontal-plane effect, humbler but venerable: Baer's law. A river flowing at $1\ \mathrm{m/s}$ feels a rightward Coriolis acceleration $fv \approx 10^{-4}\ \mathrm{m/s^2}$ — about $10^{-5}\,g$ . Tiny, but it tilts the water surface by an angle $fv/g \sim 10^{-5}$ , piling water perhaps a centimetre higher on the right bank of a kilometre-wide river and biasing the secondary circulation that erodes banks. Nineteenth-century geographers (Baer, studying great Siberian rivers) claimed north-flowing rivers cut their right banks steeper; the modern verdict is that the effect is real but usually subdominant to meandering dynamics — a good lesson in honest scale analysis.

Limiting cases

Case	Result	Meaning
equator, $\varphi = 0$	$f = 0$	no horizontal deflection; geostrophic balance impossible; hurricanes cannot form or cross
pole, $\varphi = \pm90°$	$f = \pm2\Omega$	maximal deflection; inertial period $T_f = \pi/\Omega = 11.97$ h, half a sidereal day
Southern Hemisphere	$f < 0$	every handedness flips: deflection leftward, inertial circles counterclockwise, lows clockwise
short times, $ft \ll 1$	$\Delta s \approx \tfrac12 f v t^2$	the circle degenerates to a parabolic sliver — the ballistics regime
$\mathrm{Ro} \gg 1$	inertia dominates	Coriolis is a small perturbation: balls, tornado cores, lab-scale flows
$\mathrm{Ro} \ll 1$	geostrophic balance	flow along isobars; weather maps become velocity maps

Common trapThe Coriolis force curves everything — so my football swerved because of it

Almost never. The horizontal Coriolis acceleration is $fv \sim 10^{-4}\,v$ : for a 30 m/s pass airborne 2 s the drift is $\tfrac12 f v t^2 \approx 6$ mm — three orders below aerodynamic swerve from spin. Before invoking Coriolis, compute $\mathrm{Ro} = U/fL$ (or just $\tfrac12 fvt^2$ ): rotation earns a vote only when the answer says so, which in daily life it almost never does (draining sinks included — the parent article does that arithmetic). Two subtler traps: do not use $2\Omega$ where $f = 2\Omega\sin\varphi$ belongs — the $\sin\varphi$ is the whole geography of the effect; and do not confuse horizontal deflection of horizontal motion (this article, set by $\Omega\sin\varphi$ ) with the eastward deflection of falling bodies (set by $\Omega\cos\varphi$ — maximal at the equator, where the horizontal effect is zero). They are perpendicular projections of the same cross product, and swapping their trig factors is the classic exam error.

Worked examples

1 · A drifting buoy after a storm (confidence builder)

A storm leaves a surface buoy at latitude $57°\mathrm{N}$ (central Baltic) drifting at $v = 0.25\ \mathrm{m/s}$ with no further forcing. Predict the radius, period, and sense of its motion.

Step 1. Coriolis parameter: $f = 2\Omega\sin\varphi = 2 \times 7.292\times10^{-5} \times \sin 57° = 2 \times 7.292\times10^{-5} \times 0.839 = 1.223\times10^{-4}\ \mathrm{s^{-1}}$ .

Step 2. Inertial-circle radius: $R = v/f = 0.25 / 1.223\times10^{-4} = \boxed{2.0\ \mathrm{km}}$ .

Step 3. Period: $T_f = 2\pi/f = 6.283 / 1.223\times10^{-4} = 5.14\times10^{4}\ \mathrm{s} = \boxed{14.3\ \mathrm{h}}$ , clockwise viewed from above (Northern Hemisphere).

Step 4. Sanity check against the record: the classic 1969 Baltic drifter data show loops of a few kilometres with a period near 14 hours at this latitude — the textbook formula is the observation. Note what the period does not depend on: the speed. A faster buoy draws a bigger circle in the same 14.3 hours.

2 · Naval gunnery at mid-latitude (exam level)

A naval gun at latitude $50°\mathrm{N}$ fires a shell at $v = 800\ \mathrm{m/s}$ over a range of $30\ \mathrm{km}$ (flight time $t \approx 37.5\ \mathrm{s}$ , treating the trajectory as flat). Find the Coriolis miss distance and its direction. Then state what changes at $50°\mathrm{S}$ .

Step 1. Coriolis parameter: $f = 2 \times 7.292\times10^{-5} \times \sin 50° = 2 \times 7.292\times10^{-5} \times 0.766 = 1.117\times10^{-4}\ \mathrm{s^{-1}}$ .

Step 2. Short-flight check: $ft = 1.117\times10^{-4} \times 37.5 = 4.2\times10^{-3} \ll 1$ — the parabolic-sliver formula (7) applies.

Step 3. Miss distance: $\Delta s = \tfrac12 f v t^2 = 0.5 \times 1.117\times10^{-4} \times 800 \times (37.5)^2 = 0.5 \times 1.117\times10^{-4} \times 800 \times 1406 = \boxed{63\ \mathrm{m}}$ , to the right of the aim line — regardless of firing azimuth, because $-f\hat{\mathbf{z}}\times\mathbf{v}$ turns every horizontal velocity by the same rate. That is 0.2% of the range: irrelevant at 3 km, a full ship-width miss at 30 km, which is why gunnery tables have carried latitude columns since the era of dreadnoughts.

Step 4. Southern Hemisphere: $\sin(-50°) < 0$ flips $f$ , so the same shot deflects $63\ \mathrm{m}$ to the left. Naval lore holds that at the Battle of the Falklands (1914, $52°\mathrm{S}$ ) British salvos initially straddled wide because corrections tabulated for northern latitudes were applied with the wrong sign — doubling the error to over 100 m. The documentary record is thin, but the physics of the story is exactly right, which is why every mechanics text retells it.

Step 5. Perspective: the same $f$ acting on a 10 m/s wind for a day produces not a 63 m nudge but a complete 100-km circle — equation (7) versus equation (6) is nothing but short time versus long time under one relentless sideways push.

Origins · 1 minGaspard-Gustave de Coriolis and how this was discovered

Gaspard-Gustave de Coriolis

1792–1843 · École Polytechnique, Paris

Coriolis derived his celebrated theorem for machines, not planets: his 1835 memoir on the equations of relative motion in rotating systems was written for waterwheels and turbines, whose channels carry moving parts within a spinning frame. He never applied the deflection to the atmosphere — that step waited a generation, for William Ferrel (1856), who put the rotating-frame equation on the weather map and derived the balance now called geostrophic, and for Buys Ballot (1857), whose sailors' rule — wind at your back, low pressure on your left — is equation (8) spoken aloud. The scalar $f = 2\Omega\sin\varphi$ that meteorologists now simply call "the Coriolis parameter" is thus a three-author quantity: Coriolis supplied the 2, the geographers supplied the $\sin\varphi$ , and the planet supplied the $\Omega$ . Coriolis himself, modest and chronically ill, spent his last years as director of studies at the École Polytechnique, better known to his contemporaries for defining mechanical work and for his mathematical theory of billiards.

History

1735

— George Hadley explains the trade winds by the differing eastward speeds of air at different latitudes — deflection of meridional flow, half the full story.

1835

— Gaspard-Gustave de Coriolis publishes the complete acceleration transformation for rotating machines, factor of 2 included.

1856

— William Ferrel applies the rotating-frame equations to the atmosphere and derives the geostrophic relation between wind and pressure.

1857

— Christophorus Buys Ballot states the empirical sailors' law: back to the wind, low pressure on the left (Northern Hemisphere).

1905

— Vagn Walfrid Ekman, prompted by Nansen's observation that Arctic ice drifts to the right of the wind, derives the Ekman spiral from the balance of Coriolis and frictional forces.

Going further

References. Morin, Introduction to Classical Mechanics, Ch. 10 (Accelerated Frames) · Goldstein, Classical Mechanics, 3rd ed., Sec. 4.10 · Landau & Lifshitz, Mechanics, Vol. 1, Sec. 39 · Taylor, Classical Mechanics, Ch. 9 · Marion & Thornton, Classical Dynamics, Ch. 10

Provenance. Coriolis parameter $f=2\Omega\sin\varphi$ and the rotating-frame equation follow Goldstein Sec. 4.10, Morin Ch. 10, and Marion-Thornton Ch. 10. Earth's sidereal rate $\Omega=7.292\times10^{-5}\,\mathrm{rad/s}$ is standard (one rotation per sidereal day, 86164 s). Verifier should recompute: (1) $f$ at $\varphi=30^\circ,45^\circ,90^\circ$ ; (2) inertial radius $R=v/f$ and period $T_f=2\pi/f$ in the worked examples; (3) the geostrophic speed $25\,\mathrm{m/s}$ from the quoted gradient; (4) projectile deflection $\tfrac12 f v t^2\approx36\,\mathrm{m}$ . Foucault precession rate $\Omega\sin\varphi$ (factor-of-2 distinction from $f$ ) per Landau-Lifshitz Sec. 39 and Taylor Ch. 9.

Builds on: Coriolis force · Time derivatives in rotating frames · Angular velocity vector of a rotating frame · Effective gravity on the rotating Earth · Non-inertial reference frames. Leads to: Centrifugal force · Euler force · Effective gravity on the rotating Earth.

Why is the deflection always 90° to the right, no matter which way I move?

Because the horizontal Coriolis acceleration is $-f\,\hat{\mathbf{z}}\times\mathbf{v}_h$ , and the operation " $\hat{\mathbf{z}}\times$ " rotates any horizontal vector by exactly 90° — it has no preferred compass direction of its own. The only directional information in the problem is the local vertical $\hat{\mathbf{z}}$ and the sign of $f$ . Move east, get pushed south; move south, get pushed west; the push tracks your velocity like a rudder fixed 90° to starboard. That isotropy is why one scalar $f$ suffices for the whole horizontal plane, and why the free trajectory is a perfect circle rather than some direction-dependent oval. The handedness itself comes from $\sin\varphi$ : south of the equator $f<0$ and the rudder is fixed to port instead.

The Coriolis acceleration is only about 0.001 m/s² — how can something that feeble run the entire atmosphere?

Because it never stops and never cancels. A pressure gradient eventually exhausts itself by filling the low; friction dies with the motion; but the Coriolis force is proportional to velocity and simply persists, accumulating turning at rate $f$ forever. Over one hour a $10^{-3}\ \mathrm{m/s^2}$ push deflects a 10 m/s wind by $\sim20°$ ; over a day it wins completely, forcing the flow into geostrophic balance. The honest comparison is not force against force but timescale against timescale: whenever a flow lives longer than $1/f \approx 3$ hours and stretches wider than $U/f$ , rotation has had time to organize it. That is the content of the Rossby number — weather systems sit at $\mathrm{Ro}\sim0.1$ precisely because they are slow and vast, not because the Coriolis force is strong.

Inertial circles run clockwise in the north, but cyclones spin counterclockwise — same hemisphere, opposite senses. Which one is wrong?

Neither — they are different force balances. An inertial circle is free motion: the Coriolis force is the only player, it points from the moving parcel toward the circle's centre, and the resulting orbit is anticyclonic (clockwise, N). A cyclone is forced motion: a real pressure-gradient force points inward toward the low, and it is stronger than needed for the curvature, with the Coriolis force pointing outward to balance it — that requires the air to circulate counterclockwise so its Coriolis push (90° right of velocity) aims away from the centre. Remember it this way: Coriolis-as-centripetal gives anticyclonic loops; Coriolis-as-centrifugal-balancer gives cyclonic flow around lows. Spotting which role the Coriolis force plays is the first question to ask of any rotating-flow diagram.

Do snipers really correct for the Coriolis force, or is that movie physics?

At extreme range, they really do. Take a 850 m/s round flying 1000 m: flight time $t \approx 1.2\ \mathrm{s}$ , and at $45°$ latitude the horizontal drift is $\tfrac12 f v t^2 = 0.5 \times 1.03\times10^{-4} \times 850 \times 1.44 \approx 6\ \mathrm{cm}$ — comparable to the vital-zone margin at that distance, so long-range shooting calculators include it (along with the vertical Eötvös correction, which depends on firing azimuth). At 300 m the same arithmetic gives millimetres, buried beneath wind and breathing. The rule of thumb is the one this article keeps repeating: compute $\tfrac12 fvt^2$ before invoking the rotating planet. Movies get the existence right and the magnitude wrong in both directions.

What happened to the other component, Ω cos φ? Does it just disappear for horizontal motion?

It stays in the equations but changes jobs. The expansion in Derivation A shows $\Omega\cos\varphi$ multiplying two terms: one takes vertical velocity into eastward acceleration (that is the eastward deflection of a dropped stone, maximal at the equator), the other takes eastward velocity into vertical acceleration — the Eötvös effect, which makes an eastbound ship measurably lighter and a westbound one heavier (a 20 m/s eastbound velocity changes apparent gravity by $2\Omega\cos\varphi\,v \approx 3\times10^{-3}\ \mathrm{m/s^2}$ , detectable by shipboard gravimeters and corrected in every marine gravity survey). Neither term produces horizontal deflection of horizontal motion, which is why the level-plane dynamics closes on $f = 2\Omega\sin\varphi$ alone. The Coriolis force article works both vertical-coupling effects in full.

Why is the inertial period called the half-pendulum-day — where does the factor of two between it and the Foucault pendulum come from?

The two clocks measure different things. A Foucault pendulum's swing plane precesses at $\Omega\sin\varphi$ — the plane is a headless line, and its physics (see the Foucault pendulum article) delivers precession at the frame rate itself. A free particle's velocity vector rotates at $f = 2\Omega\sin\varphi$ — the full Coriolis coupling, twice as fast. So the pendulum day is $T_{\mathrm{sid}}/\sin\varphi$ while the inertial circle closes in $T_{\mathrm{sid}}/(2\sin\varphi)$ : exactly half, hence the name. The factor of 2 is the same one born in the derivation of the Coriolis force (two equal kinematic contributions), surviving intact in the particle orbit but halved for the pendulum because the swing retraces its line each half-period. If a velocity spectrum peaks at period 14.3 h at latitude 57°, you are seeing $f$ , not $\Omega\sin\varphi$ .

Is Baer's law — rivers eroding their right banks — actually true, or is it bathtub folklore at geographic scale?

It sits honestly in between. The Coriolis push on a $1\ \mathrm{m/s}$ river is real and computable: $fv \approx 10^{-4}\ \mathrm{m/s^2}$ , tilting the water surface by $fv/g \sim 10^{-5}$ — about a centimetre of superelevation across a kilometre-wide channel, permanently biased toward the right bank (N). Unlike the bathtub (where residual swirl dwarfs Coriolis by four orders), a great river is wide, slow, and persistent, so the bias genuinely accumulates in the secondary circulation and sediment transport. But meander dynamics, geology, and floods push harder and less politely, so the Coriolis signature is a statistical tilt, not a law visible on every bend. Modern surveys of large north-flowing rivers (Baer's original evidence, the Volga and Siberian rivers) find right-bank steepening as a weak but detectable average. Verdict: real physics, modest effect — the perfect scale-analysis exam question.

If the wind blows along the isobars, how does a low-pressure system ever fill in?

Through the places where the geostrophic balance breaks. Near the surface, friction slows the wind, which weakens its Coriolis force, which lets the pressure gradient win slightly — so surface air spirals inward across the isobars at 10–30°, converging into the low. That convergence forces ascent, and rising moist air is precisely what powers clouds, rain, and the storm itself; aloft, compensating divergence exports the mass. Meanwhile the balance also fails wherever $\mathrm{Ro}$ grows — in tight curvature around the core, where the gradient-wind correction matters. So the geostrophic balance is best read as the skeleton of the flow: exact nowhere, controlling everywhere, with the interesting weather living in its small, systematic violations.

Coriolis effects on horizontal motion

Where this lives today

The physical picture

The horizontal equations

Free motion: inertial circles

Which formula answers which question

Forced motion: the geostrophic balance

Limiting cases

Worked examples

1 · A drifting buoy after a storm (confidence builder)

2 · Naval gunnery at mid-latitude (exam level)

Going further

Before this

Where this leads