Course notes · OCN 620
9 Effects of Rotation: Coriolis and Centripetal Accelerations
9.1 Motivation
The importance of the earth’s rotation in the dynamics of ocean and atmospheric circulation can hardly be overstated. Effects of rotation are dominant on time scales of a day or more and space scales of order 10 km or more. In our daily experience, however, effects of the earth’s rotation are generally imperceptible. Our feel or intuition about the dynamics of rotating systems comes mainly from experiences with systems much smaller than the earth—such as children’s merry-go-rounds and small fluid vortices such as those that form around the bathtub drain. We may also have marvelled at how rotation transforms the dynamics of a top or gyroscope, allowing it to seem to defy gravity.
Effects of rotation can be mystifying and thoroughly unintuitive. Attempts at simple explanations are frequently incorrect and/or misleading—so don’t believe everything you read.
9.2 Inertial versus non-inertial frames of reference
Newton’s Second Law—which we use in the form $\vec{a} = \frac{\vec{F}}{m}$—is valid only for an inertial frame of reference, that is, a frame moving steadily or not at all. Any frame of reference that is accelerating is a non-inertial frame. Operationally, we define an inertial frame as one in which the Second Law holds.
Now, consider a magic elevator with a constant vertical acceleration, $a_e\hat{k}$. Let $\vec{a}_r$ be the acceleration of an object relative to the elevator. Then the full acceleration of the object is $\vec{a} = \vec{a}_r + a_e\hat{k}$ so we can write the Second Law as
$$ \begin{equation} \vec{a}_r + a_e\hat{k} = \frac{\vec{F}}{m} \end{equation} $$or
$$ \begin{equation} \vec{a}_r = \frac{\vec{F}}{m} - a_e\hat{k}. \end{equation} $$Define an “elevator force” as $\vec{F}_e = -ma_e\hat{k}$. Then the Second Law works again in the non-inertial frame—the acceleration relative to the elevator equals the sum of the forces divided by the mass:
$$ \begin{equation} \vec{a}_r = \frac{\vec{F} + \vec{F}_e}{m}. \label{eq:elev} \end{equation} $$(Recall that in the Second Law, the “force” term really represents the vector sum of all the forces, however many there may be—we just don’t bother to write out the summation explicitly.)
Conclusion: motion in a non-inertial frame of reference can be described or calculated with Newton’s Second Law by defining a fictitious force equal to minus the acceleration of the frame times the mass of the object. Looking at it another way, there is a fictitious force per unit mass equal to minus the acceleration of the reference frame. (This thought experiment was used by Einstein to show the local equivalence between gravity and the acceleration of a reference frame, part of his General Theory of Relativity—but that is another story.)
It seems all we have done is to move the acceleration of the reference frame from the right to the left side of the Second Law. Who cares? Anyone who must do physics experiments in the elevator may well care. If you are in the elevator, it is the most natural frame of reference for measuring positions, velocities, and accelerations—of a bouncing ball, for example. It is nice to be able to reconcile your measurements with the Second Law—and you can do it by defining the “elevator force”.
9.3 Statics in a rotating system: Centrifugal Force
We are not in a constantly accelerating elevator, but we are on a constantly rotating earth. Recall that a rock being swung at 1 revolution per second on a string of length 1 m is accelerating radially inwards at $(2\pi)^2~\mathrm{m~s^{-2}}$. This inward or centripetal acceleration of the rock is caused by the force of the string pulling on the rock. Now, take the point of view of a cockroach hanging on to the rock for dear life. (Ignore the wind she experiences.) She feels herself being pulled off the rock, radially outward. Good physicist that she is, she concludes there is an outward force—a centrifugal force—acting on the rock and on herself.
More generally, consider a body at rest in a reference frame rotating with angular velocity $\vec{\Omega}$. Let $\vec{R}$ be the position vector of the body relative to some origin on the axis of rotation. If $\phi$ is the angle between $\vec{\Omega}$ and $\vec{R}$, then the distance from the body to the axis of rotation is $r = R\sin\phi$. (Here and elsewhere, omitting the vector symbol means we are taking the magnitude of the vector. $R = |\vec{R}|$, etc.) The velocity of the body in a non-rotating frame has magnitude $\Omega r$ and azimuthal direction. Hence the velocity vector in the inertial system is just $\vec{v}_i = \vec{\Omega} \times \vec{R}$. (Here and in the following, we will use subscript $i$ to indicate a vector in the inertial system and $r$ a vector in the rotating system.)
There is nothing in this argument that depends on the vector being a position vector. For any vector $\vec{q}$ that is constant in a rotating system, its time derivative in the inertial system will be
$$ \begin{equation} \left.\frac{d\vec{q}}{dt}\right|_i = \vec{\Omega} \times \vec{q}. \end{equation} $$We can use this rule to find the acceleration in the inertial system:
$$ \begin{align} \vec{a}_i = \left.\frac{d\vec{v}_i}{dt}\right|_i &= \vec{\Omega} \times \vec{v}_i \\ &= \vec{\Omega} \times (\vec{\Omega} \times \vec{R}). \label{eq:centrip} \end{align} $$This expression $\eqref{eq:centrip}$ is the centripetal acceleration. Its magnitude is $\Omega^2 r$. Note that the parentheses are necessary; the vector product is not associative.
The static case (no motion) of Newton’s Second Law in a rotating frame of reference is then
$$ \begin{equation} \vec{a}_r = 0 = \frac{\vec{F}}{m} - \vec{\Omega} \times (\vec{\Omega} \times \vec{R}) \end{equation} $$where the last term is the centrifugal force per unit mass. Like the “elevator force per unit mass” in $\eqref{eq:elev}$, $\frac{\vec{F}_e}{m}$, it is a fictitious force.
9.4 Dynamics in a rotating system: Coriolis Force
It is easy to generalize the argument above by relaxing the restriction that $\vec{q}$ be constant in the rotating system; just add its rate of change in the rotating system to its rate of change caused by the rotation:
$$ \begin{equation} \left.\frac{d\vec{q}}{dt}\right|_i = \left.\frac{d\vec{q}}{dt}\right|_r + \vec{\Omega} \times \vec{q}. \label{eq:transform} \end{equation} $$Applying $\eqref{eq:transform}$ to the position vector $\vec{R}$, we get the velocity in the inertial system:
$$ \begin{align} \vec{v}_i = \left.\frac{d\vec{R}}{dt}\right|_i &= \left.\frac{d\vec{R}}{dt}\right|_r + \vec{\Omega} \times \vec{R} \\ &= \vec{v}_r + \vec{\Omega} \times \vec{R}. \end{align} $$Now apply $\eqref{eq:transform}$ to the velocity, $\vec{v}_i$, to get the acceleration:
$$ \begin{align} \vec{a}_i = \left.\frac{d\vec{v}_i}{dt}\right|_i &= \left.\frac{d\vec{v}_i}{dt}\right|_r + \vec{\Omega} \times \vec{v}_i. \\ &= \left.\frac{d\vec{v}_r}{dt}\right|_r + \left.\frac{d(\vec{\Omega} \times \vec{R})}{dt}\right|_r + \vec{\Omega} \times \vec{v}_r + \vec{\Omega} \times (\vec{\Omega} \times \vec{R}) \\ &= \underbrace{\vec{a}_r}_{\text{local accel.}} + \underbrace{2\vec{\Omega} \times \vec{v}_r}_{\text{Coriolis}} + \underbrace{\vec{\Omega} \times (\vec{\Omega} \times \vec{R})}_{\text{centripetal}} \label{eq:full} \end{align} $$We have used the product rule with $\frac{d\vec{\Omega}}{dt} = 0$ and $\left.\frac{d\vec{R}}{dt}\right|_r = \vec{v}_r$.
In $\eqref{eq:full}$ we find the acceleration as measured in the rotating frame, the centripetal acceleration as in $\eqref{eq:centrip}$, and an entirely new term: the Coriolis acceleration. Note that the centripetal acceleration is a function of position but not velocity; while the Coriolis acceleration is a function of velocity but not position.
9.5 Hand-waving explanation of Coriolis force, given centrifugal force
Imagine yourself on a merry-go-round that is rotating counterclockwise (positive rotation; the direction of rotation of the earth viewed from above the north pole). To begin, you are stationary relative to the merry-go-round. In an inertial frame of reference, you are accelerating radially inward; in your frame of reference you are stationary, and you experience a centrifugal force, radially outward. To maintain your position, you hang on, pulling yourself inward. The centrifugal force is proportional to your distance from the center times the square of the rotation rate.
Now, start walking around the periphery of the merry-go-round, in the direction of rotation. In the inertial frame of reference, you are now rotating faster than before, hence accelerating inward toward the center faster than before. In the frame of reference of your body, you are still stationary, but you feel a larger centrifugal force, corresponding to your greater rotation rate in the inertial system. In the frame of reference of the merry-go-round, the centrifugal force stays the same—by definition, it is related to the rotation rate of the frame of reference only. What you identify as the increase in centrifugal force in your body reference frame is predominantly identified as the Coriolis force in the merry-go-round reference frame. It is radially outward when you walk in the direction of rotation–hence it is to the right. If you walked opposite to the direction of rotation, the centrifugal force in your body frame of reference would decrease, or in the merry-go-round frame there would be a Coriolis force radially inward, opposite to the (constant) centrifugal force. Hence, again the Coriolis force is to the right of your velocity vector relative to the merry-go-round.
Suppose you walk radially inward. Your azimuthal velocity (that is, in the direction of rotation) is proportional to your distance from the center. Hence, to walk radially inward, your azimuthal velocity must be decreased—you must accelerate (in the inertial reference frame) in a direction opposite to the direction of rotation. This acceleration to your left must be caused by a force exerted on you by the merry-go-round to your left. What do you feel? You feel yourself being shoved to the right; the leftward push on you by the merry-go-round (as you hang on to the bars, or lean to the left) is simply balancing the force you feel to the right, so that you can go straight in toward the center without deviating to the right in the merry-go-round frame of reference. Again, this force to the right of your velocity in the merry-go-round frame is the Coriolis force. (Don’t forget: all this time you are still feeling the centrifugal force in addition to the Coriolis force. As you walk radially inward, the centrifugal force is decreasing with distance from the center, so you no longer have to pull yourself in so hard, or lean so far inward.)
Now, move radially outward. Your acceleration in the inertial frame must be in the direction of rotation, which is to the left of your velocity; so the Coriolis force is still to the right of your velocity.
If you are comfortable with the idea of conservation of angular momentum—a skater spins faster by pulling arms and legs closer to the body’s axis—you may prefer to think about the Coriolis force associated with radial motion in those terms. Imagine a rock swinging around on a string threaded through the hole in a spool held above your head, so you can shorten the radius of the rock’s circle by pulling down on the string. If you do that, the rock will rotate faster to conserve angular momentum. In the frame of reference with the original rotation rate it will accelerate in the direction of rotation as the radius is decreased. In that frame of reference, the acceleration can be attributed to a force in the direction of rotation: the Coriolis force.
9.6 Centrifugal and Coriolis forces as seen in oceanography and meteorology
Recall that the centrifugal force per unit mass is a function of position only (given a constant rotation rate for the earth—an excellent approximation). The same is true of the gravitational force; gravitation and the centrifugal force feel the same to us, so we simply combine them and call it all gravity. The direction a plumb bob hangs is really determined by the vector sum of the gravitational force and the centrifugal force. A level surface is defined as one normal to this combined force.
The Coriolis force per unit mass is a function of velocity only, and it is normal to both the earth’s rotation axis and the velocity vector. For example, a current moving to the east is subject to a Coriolis force radially outward, away from the axis. If the current is at 45°N, then this Coriolis force vector will have equal components upward and to the south. It turns out that the vertical component (upward in this example) is almost never of any dynamical consequence, and so it is almost always ignored at the outset. Although it may be as large as, or larger than the horizontal component, it is minuscule relative to the other vertical forces, and its horizontal gradients are minuscule relative to the horizontal gradients of the other vertical forces—it just does not do anything interesting. So, in oceanography and meteorology, what you see presented as the “Coriolis force” is really not the whole force, but just its horizontal component.
And what you see is not even the entire horizontal component. For motions of sufficiently long time scale and large spatial scale that the Coriolis force is important, the vertical component of velocity is very small compared to the horizontal component. Therefore the part of the Coriolis force caused by the vertical velocity component is also small, and is customarily ignored at the outset. It does not even appear in the usual forms of the momentum balance equations.
Exercise: Starting from $\eqref{eq:full}$, derive the expression for the part of the Coriolis force used in oceanography.
Notes by Eric Firing · Last updated 2014/09/22