% vim: tw=50 % 07/03/2023 10AM \begin{note*}[Non-examinable] Can show (Part II Waves) that a wave packet (a group of nearly monochromatic waves) moves with the \emph{group velocity} $c_g = \dfrac{q}{k}$. \end{note*} \subsubsection*{Limiting Cases} \begin{enumerate}[(1)] \item Deep-water limit. $kh \gg 1$, $\tanh kh \simeq 1$. So $w = \sqrt{gk}$, $c = \sqrt{g / k}$ and $c_g = \half c$ independent of $h$. For example Atlantic storm produces swell waves with period $15\mathsf{s}$, $\omega = \frac{2\pi}{15} = 0.4\mathsf{s^{-1}}$, $k^{-1} = \frac{5}{\omega^2} = 60\mathsf{m} \ll \text{ocean depth} = 4\mathsf{km}$ (average). $c = \frac{\omega}{k} = 25\mathsf{ms^{-1}} = 2000 \mathsf{km / day}$ so arrive before the storm. \item Shallow water limit: $kh \ll 1$, $\tanh kh \simeq kh$. So $\omega = \sqrt{gh} k$, $c = \sqrt{gh}$, $c_y = c$ independent of $k$. For example flood waves on a river, $h = 2\mathsf{m}$ $\implies$ $c = \sqrt{gh} = 4 \mathsf{ms^{-1}} = 16\mathsf{kmh^{-1}}$. Or tsunami, $h = 4\mathsf{km}$ (average) $\implies$ $c = \sqrt{gh} = 200 \mathsf{ms^{-1}}$ roughly the speed of a commercial plane. Wavelength $\lambda \simeq 500 \mathsf{km}$. (Boxing Day 2004). \end{enumerate} \subsubsection*{Velocities} $\zeta = \zeta_0 e^{i(kx - \omega t)}$, $\phi = -\frac{i\omega \zeta_0}{k} \frac{\cosh k(z + h)}{\sinh kh} e^{i(kx - \omega t)}$. \[ \bf{u} = \nabla \phi = (\cosh k(z + h), -i\sinh k(z + h)) \frac{\zeta_0 w}{\sinh kh} e^{i(kx - \omega t)} \] (real part understood). $u_x$ and $\zeta$ are in phase. \begin{center} \includegraphics[width=0.6\linewidth] {images/d0902784bcd311ed.png} \end{center} \subsubsection*{Deep water limit ($kh \gg 1$)} $\cosh, \sinh \simeq e^{+}$. Hence $\bf{u} = (1, -i) e^{k\zeta} \zeta_0 \omega e^{i(kx - \omega t)}$ circular motion, smaller at depth. \subsubsection*{Shallow water limit ($kh \ll 1$)} $\bf{u} \simeq \left( \frac{1}{kh}, -i \frac{z + h}{h} \right) \zeta_0 \omega e^{i(kx - \omega t)}$ ($\frac{1}{kh}$ large, so horizontal velocity $\gg$ vertical velocity). \subsubsection{Standing Waves in a Container} Consider water in a deep rectangular box $0 < x < a$, $0 < y < b$, $z < 0$. Look for linearised waves with surface at $z = \zeta(x, y, t)$, $|z| \ll a, b$, $|\nabla \zeta| \ll 1$. \begin{center} \includegraphics[width=0.6\linewidth] {images/0119c5eabcd411ed.png} \end{center} Want to solve $\nabla^2 \phi = 0$ in $z < 0$, $\pfrac{\phi}{n} = 0$ on side boundaries. $\nabla \phi \to 0$ as $z \to -\infty$. Also want on $z = 0$, $\pfrac{\zeta}{t} = \pfrac{\phi}{z}$ and $\rho \pfrac{\phi}{t} + \rho g \zeta$ independent of $x, y$. Try separation of variables $\phi(x, y, z, t) = A \cos \frac{m\pi x}{a} \cos \frac{n\pi y}{b} e^{kx} e^{-i\omega t}$. \myskip Where Laplace's equation $-\frac{m^2\pi^2}{a^2} - \frac{n^2\pi^2}{b^2} + k^2 = 0$. Kinematic boundary condition \[ \pfrac{\zeta}{t} = \pfrac{\phi}{z} \implies \zeta = \frac{iAk}{\omega} \cos \frac{m\pi x}{a} \cos \frac{n\pi y}{b} e^{-i\omega t} \] dynamic boundary condition \[ \pfrac{\phi}{t} + g\zeta \implies -i\omega A + g \frac{iAk}{\omega} = 0 \implies \omega^2 = gk \] as before for deep-water. Only difference is that $k$ is quantised by the side walls: \[ k_{mn} = \left( \frac{m^2\pi^2}{a^2} + \frac{n^2\pi^2}{b^2} \right)^{\half} \] These are standing waves / normal modes. $\zeta = f_{mn}(x, y) \cos(\omega_{mn} t)$. \subsubsection*{The Rayleigh Instability} Consider an upside down container \begin{center} \includegraphics[width=0.3\linewidth] {images/44edad70bcd611ed.png} \end{center} Obvious instability! Just let $g \to -g \implies \omega^2 = -gk$ $\implies$ $\omega = \pm i\sqrt{gk}$, $e^{-i\omega t} = e^{\pm \sqrt{gk} t}$.