Aspects of geostrophic flow and inertial oscillation in the ocean and atmosphere are analogous to the motion of a rapidly spinning gyroscope.

The dynamics of the rotating shallow-water (RSW) system are fascinating and counterintuitive. Small-amplitude motion consists of a mixture of two dynamical modes. One is called geostrophic flow, which has the lowest frequency. The other is called the inertia–gravity wave, which can take any frequency above a certain threshold. Inertia–gravity waves at the threshold frequency are called inertial oscillations and are controlled by rotation. Geostrophic flow and inertial oscillations are ubiquitous in more complex fluids, such as Earth's ocean and atmosphere. Thinking about these dynamical modes can be a challenge because the influence of rotation is unfamiliar to many people. Building intuition about geostrophic flow and inertia–gravity waves in the RSW system is important, however, because it provides a solid foundation for understanding more complex motions in real environmental fluids.

This article concerns a simple tangible device that has analogous dynamics to the RSW system: the gyroscope, or spinning top. Rotating rapidly, the gyroscope seems to defy gravity by not falling over, as long as it possesses sufficient spin. The dynamical analogy between a rotating fluid and a gyroscope is an interesting and novel finding that we explain here.^{1} The gyroscope may be valuable in the classroom as it may be used in demonstrations to illustrate RSW concepts that underpin large-scale motions in the ocean and atmosphere. The gyroscope is less useful as a tool to explain the basic physics of RSW flow. The physics of gyroscopic motion is analogous to particular aspects of RSW flow but has a distinct character of its own. Instructors seeking to explain the reasons for geostrophic flow and inertial oscillations in rotating fluids should focus on explaining the Coriolis effect. The Coriolis effect can be derived mathematically by a simple coordinate transformation into Earth's rotating frame of reference, and that is typically presented in the classroom.

This approach is unsatisfying to many people, however (see, e.g., Stommel and Moore 1989). Simple physical arguments to explain the Coriolis effect consider motion in the nonrotating frame. In this vein, Persson (1998) provides an historical account and many useful references. Phillips's (2000) explanation proceeds from dynamical principles familiar from elementary mechanics and also addresses geostrophic flow and inertial motion. Durran (1993) analyzes inertial motion in detail and clarifies the origin of the RSW equations by considering a physical model of a frictionless particle moving in a parabolic dish (see also Durran and Domonkos 1996). Among textbook accounts, Marshall and Plumb (2008) is comprehensive and accessible because it builds intuition via demonstrations with rotating tables. The present analogy between RSW dynamics and gyroscopes is not intended to replace these explanations of the basic physics behind the Coriolis effect.

We begin by reviewing geostrophic flow and inertial oscillations in the ocean and atmosphere. Examples of these modes are also presented in a rotating tank of water. Next the gyroscope is introduced and explained. Examples of gyroscopic motion are shown. We explain the analogies between the gyroscope and RSW dynamics in the following section. Finally, the limits to the analogy are discussed: it applies to specific aspects of geostrophic and inertial motion because important fundamental differences exist between each system. Table 1 summarizes this comparison and provides linkage to the dynamics of the oceans and atmosphere relevant to higher education. Four movies that demonstrate these principles and analogies are provided in the online auxiliary material (http://dx.doi.org/10.1175/BAMS-D-12-00023.3). Mathematical details in the main text are kept to a minimum, but they appear in an online supplement (http://dx.doi.org/10.1175/BAMS-D-12-00023.2) for clarity, as a resource template for educators, and for further study.

Similarities and differences between ocean/atmosphere rotating fluid dynamics and the gyroscope. The specific mathematical link is given by Eq. (7) and concerns the linear, longwave limit of rotating shallow-water flow (see the supplement for mathematical details). Ocean/atmosphere fluid dynamics are more complicated than gyroscope dynamics. Cross references to the main text and the online resources are indicated.

## GEOSTROPHIC FLOW AND INERTIAL OSCILLATION IN ROTATING FLUIDS.

The extratropical ocean and atmosphere almost always exhibit geostrophic flow and inertial oscillation at large scales. For example, Fig. 1 shows satellite images of geostrophic winds and currents, both near Iceland. In **F**ig. 1a, the air swirls cyclonically (counterclockwise) around a low pressure center in the middle of the frame. The force due to the pressure gradient is directed toward this low pressure center, but the wind circles around it, nearly at right angles, and must complete several loops before reaching the center. In Fig. 1b, the surface ocean currents are seen through their action on green phytoplankton streaks. A counterclockwise vortex (an oceanic cyclone) is visible in the middle of the image as a spiral streak, similar to the atmospheric cyclone. Again, there is a low pressure center at the core of this vortex. The scales of motion are different in the two cases: the cyclone diameter is about 750 km in the atmosphere and about 40 km in the ocean. The time scales are different as well: a day or two for the atmosphere^{2} and a week or two for the ocean. Nevertheless, the fundamental dynamics in each case are the same. The flow is mainly geostrophic and dominated by Coriolis forces due to Earth's rotation.

*V,*length scale

*L*, and Coriolis parameter

*f*, equal to twice the rate of rotation about the vertical axis. In the periphery of the cyclones show in Fig. 1,

*Ro*is small and the flow is geostrophic. Toward the centers of the cyclones the Rossby number is larger. There, some of the inward pointing pressure-gradient force provides the centripetal force required for the fluid to flow around the low pressure center. When centripetal and Coriolis forces are similar in magnitude, the Rossby number is near one (called gradient wind balance). Even closer to the cyclone center, the centripetal force dominates the Coriolis force and the Rossby number is large (called cyclostrophic balance).

Figure 2 is a good example of (near-) inertial oscillation in the North Pacific Ocean. A storm excites inertial oscillations over an area 350 km on a side (van Meurs 1998). The surface water moves in clockwise circles for 3 weeks, as shown by the tracks of satellite-tracked surface drifters. The drifter tracks are coherent and in phase over most of the area shown (the phase varies only by about a quarter period in total). That is, the drifters all move north together, then all turn east, then all turn south, and so on. The period of the oscillation is about 19 h. The slow drift of the trajectories is due to geostrophic eddies like those in Fig. 1b. This low-frequency geostrophic motion is much more variable in space (nearby cycloids move in different directions). The in-phase circular movement is an inertial oscillation, however, which is distinctly different. In particular, the inertial oscillation is more predictable than the geostrophic eddies because it involves mainly linear dynamics. The motion of the geostrophic eddies involves nonlinear dynamics that can cause chaotic effects that make long-term predictions very hard. The coexistence of inertial oscillations and geostrophic eddies is an interesting property of ocean and atmosphere fluid dynamics.

Inertial oscillations are also present in the atmosphere. For example, the nocturnal jet is a low-level diurnal flow that has been explained as an inertial oscillation (Blackadar 1957). During the day the atmospheric boundary layer is actively mixed by convection owing to solar heating at the ground. At dusk, the convection intensity falls and the related frictional forces disappear. The balance of forces changes and air parcels experience an unbalanced Coriolis force that causes the flow to veer anticyclonically in an inertial oscillation. Evidence of inertial oscillations of this kind is not widespread, however, and observed inertial oscillations appear to be more commonly associated with fronts rather than sunset (Lundquist 2003).

Geostrophic f low and inertial oscillations are also easily shown using a rotating laboratory turntable. Figure 3 shows such an example of geostrophic flow using a portable apparatus that is designed for classroom demonstrations (the “Weather in a Tank” apparatus; Illari et al. 2009).^{3} A dipole eddy is visible. Flow in the lower (upper) vortex is counterclockwise (clockwise). The current is associated with a depression in the water surface in the lower vortex (low pressure) and an elevation in the upper one (high pressure). In subsequent frames the black dot, a floating piece of paper, moves between the two eddies.

Geostrophic flow is the dynamical balance between a pressure-gradient force and a Coriolis force. In Fig. 3, the pressure-gradient force acts from the high pressure center to the low pressure center. The Coriolis force is a fictitious deflective force that arises because of the rotation of the reference frame. It is proportional to flow speed and directed at right angles to both the rotation axis and the current direction. In Fig. 3, it opposes the pressure-gradient force as shown. The Coriolis force is associated with the geostrophic current at right angles visualized by the black dot and the colored dye.

In Fig. 3, the pressure gradient is due to changes in the water level η in the tank (similarly, the pressure gradient in Fig. 1b is also due to changes in sea level). The geostrophic current speed equals −*M*/*f*, where *M = −g∂*η/*∂s* for distance *s* measured from the lower eddy to the upper one (*g* is gravitational acceleration). The change in water level between the eddies is only about 40 *μ*m, visible with special optical altimetry equipment (Rhines et al. 2007) but not with the naked eye (in Fig. 1b the sea level differences are 10–20 cm). The counterintuitive hallmark of geostrophic fluids is to support these pressure gradients over long periods of time.

Figure 4 shows an example of inertial oscillation using the laboratory turntable.^{4} Initially there is no flow. The experimenter blows over the water surface for a second or two, as indicated. The particles then execute inertial oscillations with a period of 1.5 s for the tank rotation rate of 20 rpm. Notice how the motion is spatially uniform and in phase, so that peaks in the *y** position in the lower panel are synchronized, similar to Fig. 2. There is no horizontal variation in the current under inertial oscillation, unlike geostrophic flow.

Inertial oscillation is a temporally oscillating flow of a rotating fluid in the absence of pressure-gradient forces. Particles in this motion trace horizontal circular paths. Inertial oscillation occurs from the continuous deflection of moving fluid by the associated Coriolis force, which is perpendicular to the velocity. The period of inertial oscillations equals 2π/*f*, or *T*/2 here for turntable rotation period *T*, consistent with the observed period of 1.5 s. Notice that the period of inertial oscillations decreases as the rotation period decreases. Likewise, the frequency of inertial oscillation is equal to the Coriolis parameter, which is proportional to the rotation rate.

## GYROSCOPE NUTATION AND PRECESSION.

Geostrophic flow and inertial oscillation, the two fundamental modes of RSW dynamics, are analogous to the motion of a gyroscope. Before describing this connection in detail, the fundamental modes of a gyroscope are explained.

The spinning gyroscope behaves differently to the resting gyroscope. A freestanding gyroscope with no spin simply topples over in a familiar way. A rapidly spinning gyroscope does not. There are two unexpected behaviors, which are called precession and nutation (as shown in Fig. 6).^{5} Precession is the progressive movement of the gyroscope spindle under torque: the spindle traces out a cone (left panel of Fig. 6). Two common configurations provide the torque. The simplest option rests the gyroscope spindle on a flat surface (Fig. 5, bottom; Fig. 6). The gravitational force acting through the center of mass of the freestanding gyroscope and the normal reaction at the surface then furnish the torque. For a gimballed gyroscope, torque is applied by pushing on the spindle: for example, with a finger or a weight (Fig. 5, top; Fig. 7). Despite these differences in configuration, the gyroscope's response to a gravitational torque is essentially the same: the free tip of the gyroscope traces out a horizontal circle (Fig. 6). This means that the spindle traces out a vertical cone with apex at the fixed point. With no spin, the gyroscope topples over with the free tip moving in a vertical plane. This is the most unexpected property of a spinning gyroscope: it responds to a push by moving at right angles.^{6}

Nutation—from “nodding” in Latin—is an oscillatory wobble of the gyroscope's spindle (middle panel of Fig. 6). Nutation is not caused by a torque: hence the alternative name torque-free precession. It arises when the total angular momentum of the gyroscope does not equal the angular momentum about the gyroscope's spindle. In other words, the gyroscope is simultaneously spinning about the spindle axis and also another axis. Both freestanding and gimballed gyroscopes nutate. Nutation is seen most easily in the gimballed gyroscope because then no gravitational torque applies. There is no precession and knocking the spindle causes pure nutation. The nutation shown in Fig. 6, as well as in the online movie, shows a pure nutation.

An American football is a familiar object that is often seen nutating. When thrown correctly, the ball only spins about its axis of symmetry. Then no wobbling of the ball occurs (no nutation). Thrown incorrectly, the axis of symmetry traces out a cone with respect to the center of mass and the ball wobbles rapidly in the air. In this case, the total angular momentum vector contains a component that is not along the symmetry axis.^{7}

Usually, gyroscopes exhibit both precession and nutation together. In the case of a freestanding gyroscope, the spindle tip loops in a cycloid (right panel of Fig. 6). The case of a gimballed gyroscope under a gravitational torque from a weight is shown in Fig. 7.^{8} Both nutation and precession are observed. Nutation is the high-frequency circular motion of the gyroscope tip, and precession is the low-frequency sweep of the tip in the horizontal plane. Taken together, the tip motion is cycloidal. Friction in the gimbal bearings causes the nutation to slowly die away as seen in the bottom panel of Fig. 7. The precession lasts much longer, although friction eventually causes the gyroscope tip to point down as it would in the absence of spin. The force of the weight in Fig. 7 is directed to move the tip downward (the torque is horizontal), but the gyroscope responds by precessing at right angles in the horizontal direction (the precession vector is vertical).

*mg*is the weight driving the precession, applied at a horizontal distance

*l*from the gyroscope's fixed point (so

*mgl*is the torque);

*I*

_{3}is the moment of inertia about the spindle,

*I*is the moment of inertia perpendicular to the spindle, and

*ω*

_{3}is the angular velocity about the spindle.

Second, Eq. (2) shows that the precession frequency (angular velocity) is proportional to the applied force *mg*. This fact is another surprising property of the precessing gyroscope. Simple Newtonian dynamics says that an applied force is proportional to acceleration of a body in the direction of the force. The toppling gyroscope with no spin behaves like this. The spinning gyroscope seems totally different because the force is proportional to the speed of the body, an Aristotelian idea. The naive intuition based on Newtonian dynamics (ignoring the flywheel spin) is therefore qualitatively wrong!

Third, a related question is why a spinning gyroscope does not topple over. An energy analysis illuminates this issue (see the supplement for details). The gravitational potential energy is a relatively small contribution to the total gyroscope energy when it spins fast. A much larger effective potential energy exists that depends on the spin. This potential energy is quadratic in the topple angle and forms a potential well (Fig. S2 in the supplement). As the gyroscope topples over it gains kinetic energy. This kinetic energy is traded with the effective potential energy. From this point of view, nutation can be viewed as oscillations inside the potential energy well. The nutating gyroscope is trapped inside the potential well and cannot topple. The well becomes narrower as the gyroscope spins faster so that the nutation amplitude decreases with increasing spin.

### GYROSCOPES AND TOPS

The gyroscope, or top, is a simple device that takes many forms. The key common elements are a heavy flywheel, which can turn rapidly on a spindle (axis of symmetry; Fig. 5). There is a mechanism to give the flywheel a large angular velocity. A toy top is spun by finger and thumb, for instance, or a thread is wrapped round the spindle and pulled away. Tops spin while the spindle is supported by a solid object like a table. Gyroscopes also have bearings and a cage to support the flywheel and spindle. The bottom panel of Fig. 5 shows examples of two freestanding gyroscopes and a top.

More sophisticated gyroscopes have a mechanism called a gimbal that allows the spindle axis to move freely. The spindle can point in any direction, while the center of mass remains fixed. Industrial gyroscopes, used in inertial navigation instruments in aircraft and ships, for instance, operate this way. They also have a motor to keep the gyroscope spinning. The Massachusetts Institute of Technology and the AC Spark Plug Division of General Motors (MITAC) gyroscope in the top-left panel of Fig. 5 is an example. The brass lecture gyroscope in the topright panel of Fig. 5 uses a rubber wheel on an electric twist drill to spin it. The flywheel can reach speeds greater than 7000 rpm and will spin for more than 20 min.

Gyroscopes and tops are engaging devices that have fascinated curious children for generations. Spinning gyroscopes behave in a counterintuitive way that is quite different from their behavior with no spin.

## ROTATING FLUIDS ARE ANALOGOUS TO GYROSCOPES.

Having explained the fundamental modes of gyroscope motion, the link to rotating fluids can be stated (see also Table 1).

Geostrophy is analogous to precession. The analogy can be seen several ways. In both cases, the system responds to a steady force by moving at right angles and not in the direction of the force. The speed of motion, not the acceleration, is proportional to the force. This means that the motion stops when the force disappears. The speed is inversely proportional to the rotation rate [think of the geostrophic formula and the precession Eq. (2)]. The fluid analog of the nondimensional gyroscope number *ϵ* is the Rossby number, which is explained in the previous section on geostrophic flow and inertial oscillation in rotating fluids. *S*mall *ϵ* and small Rossby number mean that rotation dominates the dynamics (the Rossby number is about 10^{−3} for the rotating tank in Figs. 3 and 4).

Inertial oscillation is analogous to nutation. In both cases the frequency is proportional to the rotation rate [think of the inertial oscillation formula and the nutation Eq. (3)]. Similarly, when the amplitude of oscillation is measured by the radius of the trajectory, the amplitude is inversely proportional to the rotation rate. In other words, faster spin means faster smaller inertial circles and nutations. Moreover, the spectral gap between the high- and low-frequency modes grows with rotation rate. For the rotating fluid the ratio of inertial frequency to the inverse geostrophic advection time scale grows as *Ro*^{−1} (where *V*/*L* defines the inverse advection time scale; if *L* denotes the length of a periodic domain, *V*/*L* is the frequency at which a floating particle returns to the same spot). For the gyroscope, the ratio of nutation frequency to precession frequency grows as *ϵ*^{−2}.

*f*plane (

*f*constant) as

*u*,

*υ*) are the velocity components in the direction of the forcing (−

*M*) and 90° to the right, respectively.

^{9}The small-amplitude gyroscope dynamics for fast spin are

*θ*,

*ϕ*) are angles measuring the position of the tip of the spindle on the sphere it circumscribes (

*θ*is colatitude and

*ϕ*is longitude on this sphere; think of Fig. 6 and see Fig. S1). With the spindle horizontal as in Fig. 7 (

*θ*

_{0}= 90°) the two systems are related under the exchange

*f*and

*ω*means that inertial oscillation and nutation have opposite senses. That is, inertial oscillation is clockwise in the Northern Hemisphere (Figs. 2, 4), whereas nutation is counterclockwise (Figs. 6, 7).

_{n}When the spindle points up (*θ*_{0} < 90°) there are some minor geometric differences in the equations, but the physical analogy remains. In fact, the physical analogy remains under more general conditions too. Both the assumptions of fast spin and smallamplitude oscillations can be relaxed. Then, the gyroscope dynamics coincide with the longwave limit of RSW motion on the sphere. There is a slight difference in a geometrical prefactor in the RSW Coriolis term (see the supplement). This prefactor difference means that the gyroscope exhibits no analogy to the beta effect, which is due to variations in *f* with latitude.

The analogy extends beyond geostrophy and inertial oscillation to related phenomena. Geostrophic adjustment can be illustrated with a gyroscope, for example. Geostrophic adjustment concerns the transient RSW response to an impulse or to an arbitrary initial condition. In general, geostrophic and inertial oscillations are excited, plus some inertia–gravity waves. After several rotation periods the inertial oscillations decay because of weak dissipative processes (which are ignored in the equations above; the inertia–gravity waves propagate into the far field and/or get damped). Geostrophic f low then persists on its own. This process can be seen in Fig. 4 and the inertial oscillation movie. Physically analogous adjustment occurs in the gyroscope. Arbitrary initial conditions excite both nutation and precession. After several rotation periods the nutation decays, however, because of friction (Fig. 7). Releasing the gyroscope from rest is the easiest way to show this behavior.

There are further theoretical analogies too. For example, geostrophy is the minimum energy state in linear undamped RSW flow. Likewise, precession is the minimum energy state for the gyroscope. This means that pure precession is the equivalent to balanced flow in rotating fluids, which is an important theoretical concept. Moreover, gyroscope dynamics and the (longwave limit) RSW equations can be derived from closely related Lagrangian functions using a variational method (supplement). This illuminates the basis of RSW dynamics, as well as the Coriolis force, in a nontraditional way. Finally, conserved energy and angular momenta play leading roles in the theory governing the dynamics of each system.

## DISCUSSION OF THE ANALOG Y AND DIFFERENCES.

The RSW/gyroscope analogy is not perfect. Clear and distinct differences exist between the gyroscope and the longwave limit of the RSW system because of the different geometrical prefactor in the Coriolis term mentioned above. Lack of a beta effect in the gyroscope means that there is no westward drift for nutation as there is for inertial oscillation. Also, the RSW system is more complicated than the gyroscope because gravity (and inertia–gravity) waves also exist. They arise because of spatial variations in the fluid. There are no equivalent waves or spatial variations for the gyroscope.^{10}

There are other limits to the analogy and some misconceptions to be avoided. First, a useful mental picture of fluid motion is to think about infinitesimal moving fluid parcels. The vorticity of each parcel is analogous to the angular velocity vector of a small solid mass. As the parcels tilt and stretch, the vorticity also tilts and stretches as if the vorticity vector itself was attached to neighboring f luid parcels (Vallis 2006). Stretching of neighboring parcels (and hence the vortex line connecting them) amplifies the vorticity. This mental picture is unrelated to the RSW/gyroscope analogy: vortex lines in a fluid are not like little gyroscopes. The gyroscope analogy applies to RSW flow in which fluid parcels do not tilt or stretch.

The fundamental physics of RSW f low and gyroscopes are not the same. Specifically, the role of rotation in each system is different. In RSW flow (and the real ocean and atmosphere) Coriolis forces arise because we measure distance in a coordinate system that moves with the rotating Earth. For the gyroscope, the coordinate system is fixed to nonrotating space (see Fig. S1). This basic difference means that an inertial oscillation, unlike nutation, need not have any angular momentum measured in a nonrotating frame. For example, an ideal parcel pushed southward from Earth's North Pole will undergo inertial oscillations as viewed from Earth's surface. However, from the perspective of an observer in a fixed frame in space, the parcel moves back and forth in a straight line near the pole. Nutation of a gyroscope does not depend on the choice of a rotating reference frame in the same way. Instead, nutation only exists when the gyroscope spins about its axis and possesses angular momentum. Also the (*x*,*y*) coordinates for the RSW Eq. (5) are different to the (*θ*,*ϕ*) gyroscope angles. In the *f*-plane RSW system, one is free to rotate the horizontal axes: there is no physical distinction between moving east or moving north. The (*θ*,*ϕ*) gyroscope angles cannot be exchanged in this manner. Finally, *θ* (colatitude) and *ϕ* (longitude) are restricted to certain ranges because they are angles: the (*x*,*y*) coordinates may or may not be restricted, depending on the geometry of the RSW domain.

Geostrophy and inertial oscillation are analogs of gyroscopic precession and nutation in the following sense: there is a correspondence and partial similarity between the phenomena, although they are essentially different. Analogies of this type are common in oceanic and atmospheric science. For instance, the notion of “potential vorticity substance” (Haynes and McIntyre 1990) draws fruitful analogies between potential vorticity (PV) of a fluid and the concentration of a chemical species. Both quantities satisfy a tracer equation, although PV and chemical concentration are dissimilar in other ways. A second example concerns the Lorenz (1963) equations that were developed as a model for atmospheric convection. The Lorenz model applies to many disparate systems. It has been used as a simple model for El Niño–Southern Oscillation (Vallis 1988), a thermohaline loop oscillator (Malkus 1972), and lasers (Haken 1975). Finally, the idea of a “topographic beta effect” draws analogies between changes in the local Coriolis parameter with latitude and changes in fluid depth. Again, there are important fundamental differences between the two ideas but the analogy is valuable and widely exploited. The correspondence between geostrophy and precession, as well as inertial oscillation and nutation, is of the same type as these examples.

## CONCLUSIONS.

The RSW/gyroscope analogy is fascinating and thought provoking. It is also potentially useful in demonstrating the dynamics of rotating fluids in various educational settings. At the most basic level, one can simply show that a gyroscope will not fall over when it spins. This strange behavior is similar in oceanic and atmospheric weather systems because the Earth spins. They are strange for similar reasons: in both cases, rotation is a controlling effect. Students at virtually any level can also carefully push on a spinning gyroscope. The spindle slips sideways according to the force and inversely with the spin. The push of the finger is analogous to the pressure-gradient force in geostrophy and the opposing push of the gyroscope is analogous to the Coriolis force. This is a profound experience, even for a seasoned instructor, whether in an advanced dynamics course or when illustrating basic principles of motion for precollege-level students. The brass lecture gyroscope in Fig. 5 spinning 100 times per second is particularly memorable.

To this end we have used gyroscope demonstrations in kindergartens and elementary schools and with high-school science teachers, undergraduate humanities students, graduate students in rotating fluid dynamics, and senior faculty in oceanic and atmospheric science. The anecdotal response has been positive, although we have no decisive evidence as to their pedagogical effectiveness. The mathematical analysis supports the demonstration but is often not needed.

## ACKNOWLEDGMENTS

We are supported by NSF Award DUE-0618483 to the “Weather in a Tank” team. Steve Wonnell, Bill Ruff, and Sathappan Ramesh of the Johns Hopkins Department of Physics and Astronomy kindly helped us film their MITAC gyroscope. Robert Nedbor-Gross helped make the movie of geostrophic flow. The *Aqua and Terra* satellite images in Fig. 1 are courtesy of NASA/GSFC MODIS Rapid Response. We thank Paul Croft for helpful comments.

## REFERENCES

Barker, E. F., 1960: Elementary analysis of the gyroscope.

,*Amer. J. Phys.***28**, 808–810.Blackadar, A. K., 1957: Boundary layer wind maxima and their significance for the growth of nocturnal inversions.

,*Bull. Amer. Meteor. Soc.***38**, 283–290.Brancazio, P. J., 1987: Rigid-body dynamics of a football.

,*Amer. J. Phys.***55**, 415–420.Butikov, E., 2006: Precession and nutation of a gyroscope.

,*Eur. J. Phys.***27**, 1071–1081.Durran, D. R., 1993: Is the Coriolis force really responsible for the inertial oscillation?

,*Bull. Amer. Meteor. Soc.***74**, 2179–2184.Durran, D. R., and S. K. Domonkos, 1996: An apparatus for demonstrating the inertial oscillation.

,*Bull. Amer. Meteor. Soc.***77**, 557–559.Eastman, P. C., 1975: Painless precession.

,*Amer. J. Phys.***43**, 365–366.Edwards, P. L., 1977: A physical explanation of the gyroscope effect.

,*Amer. J. Phys.***45**, 1194–1195.French, A. P., 1971:

. W. W. Norton & Company, 743 pp.*Newtonian Mechanics*Haken, H., 1975: Analogy between higher instabilities in fluids and lasers.

,*Phys. Lett.***53A**, 77–78.Haynes, P. H., and M. E. McIntyre, 1990: On the conservation and impermeability theorems for potential vorticity.

,*J. Atmos. Sci.***47**, 2021–2031.Haynes, P. H., C. J. Marks, M. E. McIntyre, T. G. Shepherd, and K. P. Shine, 1991: On the “downward control” of extratropical diabatic circulations by eddy-induced mean zonal forces.

,*J. Atmos. Sci.***48**, 651–678.Illari, L., and Coauthors, 2009: “Weather in a Tank”— Exploiting laboratory experiments in the teaching of meteorology, oceanography, and climate.

,*Bull. Amer. Meteor. Soc.***90**, 1619–1632.Lock, J. A., 1989: An alternative approach to the teaching of rotational dynamics.

,*Amer. J. Phys.***57**, 428–432.Lorenz, E. N., 1963: Deterministic nonperiodic flow.

,*J. Atmos. Sci.***20**, 130–141.Lundquist, J. K., 2003: Intermittent and elliptical inertial oscillations in the atmospheric boundary layer.

,*J. Atmos. Sci.***60**, 2661–2673.Malkus, W. V. R., 1972: Non-periodic convection at high and low Prandtl number.

,*Mem. Soc. Roy. Sci. Liege***6**, 125–128.Marshall, J. C., and R. A. Plumb, 2008:

. Elsevier, 319 pp.*Atmosphere, Ocean, and Climate Dynamics: An Introductory Text*Persson, A., 1998: How do we understand the Coriolis force?

,*Bull. Amer. Meteor. Soc.***79**, 1373–1385.Phillips, N. A., 2000: An explication of the Coriolis effect.

,*Bull. Amer. Meteor. Soc.***81**, 299–303.Rasmussen, E. A., and J. Turner, 2003:

. Cambridge University Press, 612 pp.*Polar Lows: Mesoscale Weather Systems in the Polar Regions*Rhines, P. B., E. G. Lindahl, and A. J. Mendez, 2007: Optical altimetry: A new method for observing rotating f luids with applications to Rossby and inertial waves on a polar beta-plane.

,*J. Fluid Mech.***572**, 389–412.Stommel, H. M., and D. W. Moore, 1989:

. Columbia University Press, 297 pp.*An Introduction to the Coriolis Force*Vallis, G. K., 1988: Conceptual model of El Niño and the Southern Oscillation.

,*J. Geophys. Res.***93**, 13 979–13 991.Vallis, G. K., 2006:

. Cambridge University Press, 745 pp.*Atmospheric and Oceanic Fluid Dynamics: Fundamentals and Large-Scale Circulation*van Meurs, P., 1998: Interactions between near-inertial mixed layer currents and the mesoscale: The importance of spatial variabilities in the vorticity field.

,*J. Phys. Oceanogr.***28**, 1363–1388.

^{}

^{1} “Gyroscopic pumping” is sometimes discussed in the atmospheric dynamics and astrophysics literature, but that refers to meridional overturning driven by azimuthal stresses (e.g., Haynes et al. 1991).

^{}

^{2} The atmospheric cyclone in Fig. 1a is a polar low, a small, short-lived depression [for another example and further details, see Fig. 1.16 in Rasmussen and Turner (2003)]. Many atmospheric cyclones have larger scales and a life cycle of several days to a few weeks, however.

^{}

^{3} See also the online movie geostrophic_movie.mov.

^{}

^{4} See also the online movie inertial_osc_movie.m4v.

^{}

^{5} See also the online movie gyroscope_schematic_movie.m4v.

^{}

^{6} Multiple physical explanations for precession exist. They include conservation of vector angular momentum (French 1971; Butikov 2006) and consideration of the forces on point masses comprising the gyroscope rotor (Edwards 1977; Eastman 1975; Barker 1960).

^{}

^{7} For physical explanations of nutation, see Butikov (2006), French (1971), and Lock (1989). For a discussion of American football dynamics, see Brancazio (1987).

^{}

^{8} See also the online movie gyroscope_movie.m4v.

^{}

^{9} Note that the longwave limit gives vanishing pressure-gradient force. For this reason, the forcing *M* of the longwave RSW equations is not a real pressure-gradient force but some other uniform body force. This nuance does not undermine the physical analogy between the two systems. The supplement gives a more precise and general statement of the relation.

^{}

^{10} Some gyroscopes have modified flywheels with variable moments of inertia and hence more complex behavior. The large gyroscope in the bottom panel of Fig. 5 has a hollow flywheel partially full of liquid, for example (a “hydro gyro”). The liquid causes interesting new oscillations associated with surface waves on the liquid/air interface inside the flywheel. These oscillations are not analogous to spatial variations in RSW flow, however.

## APPENDIX.

*Auxiliary Materials.* Five items are included as auxiliary materials. They are as follows:

Supplement to “Analogies of ocean/atmosphere rotating fluid dynamics with gyroscopes: Teaching opportunities,” which contains a technical discussion of gyroscope dynamics and mathematical details of the links to the RSW system.

geostrophic_movie.mov, a movie showing geostrophic flow in a rotating tank of water, as in Fig. 3. The movie consists of a sequence of still images taken by a camera mounted above the tank and rotating with the tank. There is an image about every 3 s (see the time indicated in the bottom left of the frame). The tank contains water 10 cm deep, rotating at 20 rpm. The frame size is 18.5 by 12.5 cm; the tank itself is square with side 40 cm. The water is gently stirred (off frame) to produce a turbulent flow that is in geostrophic balance, similar to the ocean eddies shown in Fig. 1. The red and blue colors are food dyes that move with the flow. The black dot is a floating piece of paper from a hole punch.

inertial_osc_movie.m4v, a movie of inertial oscillation in a rotating tank of water, as in Fig. 4. See the description of the geostrophic movie for more details.

gyroscope_schematic_movie.m4v, a movie illustrating the fundamental gyroscopic modes: precession (under a torque), nutation (torque-free precession), and both precession and nutation. See also Fig. 6.

gyroscope_movie.m4v, a slow-motion clip of the MITAC gyroscope undergoing nutation and precession, similar to the top panel of Fig. 7.