Past exam of the mathematics course of the University of Cambridge / 2015 / iii / Paper 79 (Fluid Dynamics of Climate)

Take $\mathbf{f}=f\hat{\mathbf{z}}$ with $f>0$, and put $\zeta =v_x-u_y$, $d=u_x+v_y$ and $q=\zeta -f\eta /H$. Here $q$ denotes the signed vertical component of the linearized shallow-water potential-vorticity anomaly, rather than its magnitude. The divergence and vertical curl of the linearized shallow water equations giveConsequently $q_t=0$. Eliminating $d$ and substituting $\zeta =q+f\eta /H$ givesThe vector forcing is $-H\mathbf{f}\cdot \mathbf{q}$. This is potential-vorticity conservation in its linear, f-plane form: the conserved anomaly forces a stationary balanced part, while the homogeneous equation supports inertial-gravity waves.

For the initial strip, differentiating the discontinuous velocity in the sense of distributions givesThese are two oppositely signed vortex sheets. In the final geostrophic balance, $v_f=0$ and $u_f=-g{\eta }_f^{\prime }/f$. The height therefore solvesThis is the Rossby deformation radius. Decay at infinity fixes the Green function to $G(y)=-\lambda e^{-|y|/\lambda }/2$. Thus the geostrophic adjustment of a finite-width current has the particularly useful representationExpanding the exponentials inside the strip gives ${\eta }_f/H=-(U/c)e^{-a/\lambda }\sinh (y/\lambda )$; above the strip it gives $-(U/c)e^{-y/\lambda }\sinh (a/\lambda )$, and below it gives $(U/c)e^{y/\lambda }\sinh (a/\lambda )$. The height is continuous, odd and exponentially localized. Its derivative has the jumps required by the two Dirac delta functions.

Differentiating the height, rather than assuming a uniform final current, gives the complete velocity:The one-sided velocity jump is $U$ at each edge, with opposite orientations. The value exactly on an idealized vortex sheet is immaterial. Inside the strip the current remains in the original direction; outside it a return current develops.

Let $r=a/\lambda$ and $s=y/\lambda$. For $r\ll 1$, $u_f/U=1-r+O(r^2)$ over the narrow strip, while the outside return flow is approximately $-r{e}^{-|s|}$. The height varies almost linearly across the strip, $c{\eta }_f/(HU)\simeq -s$, and has extrema of magnitude $(1-e^{-2r})/2\simeq r$ at its edges. For $r\gg 1$, the central current is exponentially small: $u_f(0)/U=e^{-r}$. Each edge supports a layer of width $\lambda$, with $u_f/U\simeq \frac{1}{2}{e}^{-(a-|y|)/\lambda }$ on its inner side and $-\frac{1}{2}{e}^{-(|y|-a)/\lambda }$ outside. The corresponding height extrema approach $\pm HU/(2c)$, with almost zero height deep inside and far outside. Both requested profiles are drawn from the exact functions below; the dashed jumps represent one-sided velocity limits.

For balanced shallow-water energy as a signed potential-vorticity pairing, divide out the constant density and work per unit distance in $x$. The kinetic energy plus surface potential energy isMultiplication of the stationary height equation by ${\eta }_f$ and integration by parts, with vanishing end terms, yieldsThe same expression holds with $dA$ for a finite two-dimensional domain or a finite periodic length in $x$. The infinite strip's total energy is infinite, so both energies and their ratio are understood per unit $x$-length.

There is a sign defect in the printed energy expression: $|\mathbf{q}|$ must be replaced by the signed component $\hat{\mathbf{f}}\cdot \mathbf{q}$. Indeed, $|q_0|=|U|[\delta (y-a)+\delta (y+a)]$, so pairing it with the odd ${\eta }_f$ gives zero, although the balanced state has positive energy. The signed pairing givesSince $E_i=H{U}^{2}a$, the energy retention in finite-width geostrophic adjustment isFor $r\ll 1$ this ratio is $1-r+\frac{2}{3}{r}^2+O(r^3)$: nearly all energy remains in the narrow balanced current. For $r\gg 1$ it is approximately $1/(2r)$: only the edge regions retain balanced energy. Conservation of energy still holds for the inviscid evolution. The missing balanced energy is carried away by inertia-gravity waves; “final state” means the local balanced limit after those waves leave, rather than a loss of the total energy over the entire infinite domain.

On a sphere of radius $R$ rotating at angular speed $\Omega$, the Coriolis parameter is $f=2\Omega \sin \varphi$. Near latitude ${\varphi }_0$, write $y=R(\varphi -{\varphi }_0)$, use locally eastward and northward Cartesian coordinates, and expandThe beta plane retains the first northward variation of planetary vorticity while neglecting higher latitude dependence and metric curvature. A midlatitude local calculation assumes $L/R\ll 1$ and $|\beta |L/|f_0|\ll 1$; near the equator the distinct equatorial beta plane has $f_0=0$, so the midlatitude low-frequency reduction used below does not apply.

For a homogeneous shallow layer, the hydrostatic approximation gives $p=p_{\mathrm{atm}}+\rho g(\eta -z)$ and hence $p_x/\rho =g{\eta }_x$, $p_y/\rho =g{\eta }_y$, independently of depth. The approximation follows from the small aspect ratio $H/L$ and neglect of vertical acceleration. Differentiate the linear horizontal momentum equations in $z$. Their depth derivatives obeyStarting from rest gives $u_z=v_z=0$ initially, and the unique solution remains zero. This establishes depth independence of hydrostatic shallow-water flow; an arbitrary pre-existing shear would not be removed merely by taking the hydrostatic approximation. Integrating the continuity equation between the rigid bottom and the moving surface gives ${\eta }_t+H(u_x+v_y)=0$ at linear order. With the depth-integrated shallow-water transports $U=Hu$, $V=Hv$, one obtains

It is useful to make the coefficient approximation in the height reduction explicit. Put ${\mathcal{F}}_y={\partial }_t^2+f(y{)}^2$. Differentiating the two momentum equations in time and eliminating the other transport givesSince ${\partial }_y({\mathcal{F}}_{y}V)={\mathcal{F}}_y{V}_y+2f\beta V$, taking their horizontal divergence and using the continuity equation givesApplying ${\mathcal{F}}_y$ proves the exact variable-Coriolis shallow-water height equationAt the reference latitude, or after the usual local freezing of undifferentiated $f$ factors to $f_0$, this gives the height relation written with ${\mathcal{F}}_0={\partial }_t^2+f_0^2$. With $f=f_0+\beta y$ over a finite region, that constant-coefficient version is a local approximation, not an exact identity. Keeping ${\mathcal{F}}_y$ as above avoids silently commuting a variable Coriolis parameter through a spatial derivative.

For the slow Rossby wave branch, take $|\omega |/|f_0|\ll 1$, approximate ${\mathcal{F}}_0$ by $f_0^2$, and discard the two time-derivative terms on the right compared with $-f_0^2{\eta }_x$. This is the regular midlatitude long-time ordering, with nondegenerate zonal variation. If $a=f_0/c$, the result isThe printed low-frequency equation has the opposite right-hand sign. The minus sign follows directly from the preceding eliminated equation and is also the sign needed for the printed isofrequency-circle centre. An independent check uses geostrophic balance: $\psi =g\eta /f_0$, $(u,v)=(-{\psi }_y,{\psi }_x)$, and the shallow-water quasi-geostrophic potential vorticity is ${\nabla }_h^2\psi -a^2\psi +\beta y$. Linear potential-vorticity conservation gives $({\nabla }_h^2\psi -a^2\psi {)}_t+\beta {\psi }_x=0$.

For $\eta =A{e}^{i(kx+ly-\omega t)}$, the shallow-water Rossby-wave dispersion relation isFor fixed $l$ and $\beta >0$, it is odd in $k$, zero at $k=0$, negative for $k>0$, and tends to zero from below as $k\to +\infty$. Its minimum occurs at $k=\sqrt{l^2+a^2}$ with ${\omega }_{\min }=-\beta /[2\sqrt{l^2+a^2}]$. Thus both long and short waves have small frequency. The zonal phase velocity is westward, $\omega /k<0$, whereas the zonal group velocity isIt changes sign at the frequency minimum.

At fixed nonzero frequency, completing the square yields the Rossby-wave isofrequency circleThe radius is real only if $|\omega |\le |\beta |/(2|a|)$. For $\beta >0$, the branch with $k>0$ has $\omega <0$ and a circle centred on the positive $k$-axis. If instead the printed plus-sign wave equation is taken literally with this same Fourier convention, its dispersion is $\omega =+\beta k/(k^2+l^2+a^2)$ and its circle is centred at $(+\beta /(2\omega ),0)$. These two conventions cannot be mixed.

For reflection of a Rossby wave at a meridional wall, the stationary wall preserves frequency, and its translation invariance in $y$ preserves the tangential wavenumber. Thus ${\omega }_r={\omega }_i$, $l_r=l_i$. Both $k$ values solveTheir sum and product giveAn incident wave in $x>0$ travels toward the wall in group velocity, not necessarily in phase velocity. For $\beta >0$, $0<k_i<\sqrt{l_i^2+a^2}$ has $c_{gx,i}<0$, while its partner $k_r>\sqrt{l_i^2+a^2}$ has $c_{gx,r}>0$. The double-root case has zero normal group velocity and does not describe a wave packet incident on the wall.

At leading quasi-geostrophic approximation, the impermeability condition is $u=-g{\eta }_y/f_0=0$. For $l_i\ne 0$, the boundary condition givesThe reflected height therefore has equal amplitude and a phase change of $\pi$. Equality here concerns the height or quasi-geostrophic streamfunction amplitudes in the reduced model. At the degenerate $l_i=0$, geostrophic no-normal-flow is automatically satisfied and alone does not determine their ratio; the usual homogeneous wall-streamfunction condition supplies $A_r=-A_i$ if imposed. Retaining the small ageostrophic transport changes the boundary condition towhich tends to $-1$ in the regular low-frequency ordering but is not generally of unit modulus. In particular, $l_i=0$ gives $A_r/A_i=-k_i/k_r$ in that more complete boundary relation. Thus the printed equal-amplitude assertion requires the nondegenerate leading quasi-geostrophic approximation, or an explicit homogeneous wall condition; it is not a general exact shallow-water reflection law.

Use the quasi-geostrophic streamfunction convention $u_n=-{\psi }_{n,y}$, $v_n={\psi }_{n,x}$, and write $F=k_I^2$, $f=f_0+\beta y$. In a steady, large-scale basin interior, neglect the material derivatives of relative vorticity and interfacial stretching compared with advection of planetary vorticity. With small Rossby number, weak nonlinear eddy terms, no significant interior friction or topographic forcing, and the specified forcing confined to layer 1, Sverdrup balance isFor equal depths $h$, the depth-integrated meridional transport is $h({\bar{v}}_1+{\bar{v}}_2)=hW/\beta$. Here $W$ is the normalized potential vorticity source appearing in the evolution equation. If the dimensional wind stress curl is used, its usual layer forcing is $W=({\nabla }_h\times \mathbf{\tau }{)}_z/({\rho }_{0}h)$, so $\beta h({\bar{v}}_1+{\bar{v}}_2)=({\nabla }_h\times \mathbf{\tau }{)}_z/{\rho }_0$.

For $\beta >0$ and $W<0$, the upper-layer interior transport is southward. In a closed subtropical basin, negative wind stress curl also corresponds to downwelling Ekman pumping for $f_0>0$ and an anticyclonic gyre. A northward return transport is needed to close the circulation; its narrow western boundary current requires processes outside the frictionless Sverdrup balance. The local interior equations by themselves do not specify the detailed boundary-current structure.

Expanding the layer equation shows the physical budget:The first term is the change of planetary vorticity as a fluid parcel moves north or south. The second is the change of relative vorticity. The last is vortex stretching in layered quasi-geostrophic flow: displacement of the interface changes layer thickness and thus the stretching contribution to potential vorticity. Its opposite signs in the two layer definitions express their thickness changes in opposite directions. The wind stress curl supplies or removes upper-layer potential vorticity. Each $D_n/Dt$ follows its own layer velocity, rather than a common velocity for both layers.

For uniform $W$, choose the local two-layer Sverdrup interiorUnforced background zonal currents and arbitrary additive interface offsets have been set to zero. This choice is also an exact uniform-flow solution of the stated forced equations, not just a leading balance: ${\bar{q}}_1=f_0+\beta y-FVx$, ${\bar{q}}_2=f_0+\beta y+FVx$, and ${\bar{v}}_1{\bar{q}}_{1,y}=V\beta =W$. The local interface slope can be nonzero even though its stretching contribution is constant along each basic-state trajectory. A streamfunction linear in $x$ is an interior-patch description, not a complete globally bounded basin solution.

The potential-vorticity gradients in a meridional two-layer current are important here: besides ${\bar{q}}_{n,y}=\beta$, one has ${\bar{q}}_{1,x}=-FV$ and ${\bar{q}}_{2,x}=FV$. They must be retained when linearizing, even though the basic relative vorticity vanishes.

Let $K^2=k^2+l^2$ and $A=K^2+F$. For a normal mode the two-layer quasi-geostrophic potential vorticity amplitudes areBecause the basic velocities are northward $(V,0)$ in the two layers, their advection frequencies are $\sigma -lV$ and $\sigma$. The uniform forcing has no perturbation, $W^{\prime }=0$. The linear equations are thereforeIn particular, a perturbation's zonal velocity advects the basic interface-induced zonal potential-vorticity gradient. Substitution of the plane wave givesA nonzero disturbance exists exactly when the determinant vanishes. The requested relation for linear stability of a meridional two-layer current isEquivalently, with $\mathcal{B}=K^2(K^2+2F)$,Its discriminant isFor $K>0$, exponential baroclinic instability occurs when this discriminant is negative; otherwise the two frequencies are real. This also displays the stabilizing contribution of the planetary vorticity gradient when $k\ne 0$. As a check, $V=0$ gives the uncoupled barotropic mode and baroclinic mode frequencies $-\beta k/K^2$ and $-\beta k/(K^2+2F)$.

For $k=0$ and $l\ne 0$, divide the previous determinant equation by $l^2(l^2+2F)$ to obtainThusWith the chosen $e^{-i\sigma t}$ convention, a positive imaginary part of $\sigma$ is growth. For a nonzero basic shear $V$, the necessary condition for meridional-wave instability of a two-layer Sverdrup flow isFor these nondegenerate modes it is also sufficient. The growth rate isModes with $l^2>2k_I^2$ have real frequencies. Equality is the zero-growth coalescence of the two roots, while $l=0=k$ is a spatially uniform degeneracy and is not a growing finite-wavelength disturbance.

For $k=0$, the perturbation has $v^{\prime }={\psi }_x^{\prime }=0$, so it does not directly advect the northward planetary vorticity gradient. Consequently $\beta$ has no explicit restoring contribution to the fixed-$V$ instability criterion. It still sets the Sverdrup balance velocity: at fixed wind forcing, $V=W/\beta$, andIncreasing $|\beta |$ at fixed $W$ reduces the shear and the growth rate, without creating a finite-$\beta$ wavelength cutoff for these purely meridional modes. The limit $\beta \to 0$ at fixed $W$ is not a valid finite Sverdrup balance; its divergent velocity violates the weak-flow assumptions. Growth draws on the interfacial displacement and vertical shear, rather than on a time-dependent wind forcing, whose perturbation was set to zero.

Use $(\bar{u},\bar{v})=(-{\bar{\psi }}_y,{\bar{\psi }}_x)$ and the hydrostatic thermal-wind balance convention $\bar{b}=f_0{\bar{\psi }}_z$. Because the basic state is independent of $x$, $\bar{v}=0$. The thermal-wind relation isIntegration from the resting lower boundary givesA compatible quasi-geostrophic streamfunction iswhere $C$ is a spatial constant. Its horizontal Laplacian is $z{A}^{\prime \prime }(y)/f_0$ and its second vertical derivative is $B^{\prime }(z)/f_0$. Substituting into the given three-dimensional quasi-geostrophic potential vorticity yieldsIn particular, the horizontal-curvature term has a plus sign: $\bar{\zeta }=-{\bar{u}}_y=z{A}^{\prime \prime }/f_0$.

The reference-buoyancy convention in quasi-geostrophic potential vorticity should be stated. If $\bar{b}$ includes the full reference profile $N^{2}z$, the formula above includes its constant contribution $f_0$. If instead the quasi-geostrophic streamfunction represents pressure relative to that reference hydrostatic state, use $\bar{b}-N^{2}z=f_0{\bar{\psi }}_z$. Then the quoted $\bar{q}$ is reduced by the constant $f_0$. Both conventions give exactly the same velocity and potential-vorticity gradients, and therefore the same subsequent dynamics. The coefficient $N$ in the stated linear model is a prescribed reference buoyancy frequency; if it is also required to equal the full physical buoyancy frequency everywhere, then $B^{\prime }(z)=N^2$, rather than an arbitrary function. The formal expression above retains the requested arbitrary $B$ with the stated constant coefficient.

Put $S=M^2/f_0$. The thermal wind is $\bar{U}(z)=-Sz$, and the basic potential vorticity is spatially constant: $\bar{q}=2f_0$ when the full reference buoyancy is included, or $f_0$ when that reference contribution has been subtracted. Either way ${\nabla }_h\bar{q}=0$.

For a smooth normal mode, the linear quasi-geostrophic potential-vorticity equation becomesTake $k\ne 0$ and $K>0$, as required for a nontrivial wave with the printed phase-speed parameterization. Except possibly at an isolated critical level of a shear-flow wave, this implies $\hat{q}=0$; regularity then extends that condition through the isolated level. HenceDecay at infinity excludes the growing exponential, giving the decaying vertical structure of a semi-infinite Eady edge waveThe amplitude $C$ is arbitrary. The semi-infinite Eady model supports a wave trapped at its lower boundary with penetration depth $m^{-1}$. The zero-horizontal-wavenumber case has no nonzero decaying solution of this homogeneous vertical equation. Singular neutral interior potential vorticity sheets belong to a different continuous-spectrum class; they are not the smooth decaying edge-wave eigenfunction requested here.

The PDF contains the request for this vertical structure and uses the phase factor $e^{i(kx+ly-kct)}$. The TeX aid omits that request and also inserts a duplicate factor without $c$. The solution uses the original PDF form throughout.

The interior potential-vorticity equation alone does not determine the frequency. Use the rigid-boundary buoyancy condition for quasi-geostrophic waves. Linear adiabatic buoyancy conservation about the basic state givesAt $z=0$, the impermeability condition sets $w^{\prime }=0$ and the basic velocity is zero. ThereforeSubstitution of ${\hat{\psi }}^{\prime }(0)=-m\hat{\psi }(0)$ yields the dispersion relation of a semi-infinite Eady edge waveFor $f_0>0$, this is $c=-M^2/[N\sqrt{k^2+l^2}]$. The sign is fixed by both the thermal-wind relation and the lower-boundary buoyancy budget.

The steering height of a semi-infinite Eady edge wave is obtained by cancelling the phase-speed term against the basic advection in the wave-following frame:Thus the phase velocity equals the basic thermal-wind velocity at one penetration depth. For nonzero $M$ they have the same sign in the laboratory frame. If “exactly opposes the thermal wind” is read literally as $c=-\bar{U}(z)$, it instead gives $z=-1/m$, outside the fluid: there is no positive physical height with opposite laboratory velocities. The positive height above is the physically meaningful critical level, where the $-c$ and $\bar{U}$ contributions to the wave-frame material derivative oppose and cancel. This distinction resolves the ambiguous printed wording without reversing the derived wave speed. If $M=0$, the shear and $c$ vanish, so there is no distinguished steering height.

For an exponentially growing normal mode with $k\ne 0$, $\mathrm{Im}(kc)>0$. Such a complex $c$ cannot equal the real basic velocity anywhere, so the interior equation forces $\hat{q}=0$ at every height. Decay then fixes the unique vertical structure $C{e}^{-mz}$, and the lower-boundary condition fixes $c=-S/m$, which is real. This contradiction proves the absence of exponential instability in the semi-infinite Eady model:For $k=0$, there is no advective frequency in the linear interior equation, so an exponentially growing mode would again have $q^{\prime }=0$. The wall buoyancy equation then forces ${\psi }_z^{\prime }(0)=0$, which the nonzero decaying exponential cannot satisfy. Thus these modes do not provide a missing growing branch. This conclusion rules out exponential eigenmode growth, not every possible transient response or singular neutral potential vorticity disturbance.

With a second rigid boundary at $z=D$, the same uniform-shear flow is the Eady model. Both boundaries carry buoyancy perturbations and support Boundary Rossby waves. Their intrinsic propagation directions oppose one another relative to their respective local basic flows. At suitable wavelengths their interacting fields can phase-lock and release basic-state potential energy, producing baroclinic instability. The single-boundary system lacks that second interacting boundary wave.

For a direct comparison, write $\Lambda ={\bar{U}}_z=-S$ and $m=NK/|f_0|$. The vertical solution between the two boundaries is $A\cosh (mz)+B\sinh (mz)$, and their buoyancy conditions areEliminating $A,B$ gives the finite-depth Eady dispersion relationThe bracket is negative for $0<mD<2.399\dots$, so those modes with $k\ne 0$ have a growing and a decaying member. At short wavelengths the two edge waves interact weakly and the frequencies are real. In the semi-infinite limit, the lower-wave root tends to $c=\Lambda /m=-S/m$, recovering the neutral result above.