Fig. 1. Internal wave modes in Lake Biwa computed assuming two-layer stratification. Color and arrows show interface displacement and average velocity in upper layer, respectively.
Fig. 2. Rossby wave modes in Lake Biwa computed assuming two-layer stratification. Color and arrows show interface displacement and average velocity in upper layer, respectively. Although phases of these waves keep moving to east, they are standing waves from an energetics point of view.
Fig. 3. Results of a CWR-ELCOM simulation with extracted 4 internal wave modes and 2 Rossby wave modes. Color and arrows show interface displacement and average velocity in upper layer, respectively.
Fig. 4. Effect of wind direction on excitation of different internal wave modes. Vertical axis is normalized maximum energy input from spatially uniform winds. Wind direction is positive clockwise, and 0 when the wind is blowing from the north. V#H#: vertical and horizontal mode number. Reformatted from Shimizu et al. (2007, Limnol. Oceanogr.).
Fig. 5. Energy partitioning among different classes of modes, extracted from the results of a CWR-ELCOM simulation using two-layer approximation. (a) Total energy and (b) accumulated energy input from winds. Solid lines show accumulation for each class, and dashed, dotted, and dash-dot lines correspond to accumulation up to the 1st, 5th, and 10th internal wave modes, but up to 1st, 20th, and 50th Rossby wave modes. RW: Rossby wave, TW: topographic wave, SW: surface wave, and IW: internal wave. Reformatted from Shimizu et al. (2007, Limnol. Oceanogr.).
Modal analysis is a common method used to analyse vibration problems in physics and engineering. For two-layer-stratified lakes, the method of calculating basin-scale modes has been known for many years1),7). However, a lack of theoretical understanding prevented physical limnologists from taking full advantage of the method. My idea was to enable a wide range of applications of this method as in other fields, by developing the theory of modal analysis for stratified rotating lakes. We made a first step in theoretical development, and then applied it to Lake Biwa to investigate horizontal structure and excitation of the basin-scale motions.
Lake Biwa is the largest lake in Japan, with ~60-km length, ~20-km width, and ~100-m maximum depth. During the stratified period, Lake Biwa supports many basin-scale internal wave modes, as well as strong and persistent gyres (typical current speed ~0.1 m/s).
Basin-scale modes in Lake Biwa were computed using two-layer approximation. The internal wave modes are shown in Fig. 1, and the Rossby wave modes in Fig. 2. (The Rossby wave modes have very long periods, and effectively represent quasi-geostropic flows, such as gyres.)
Following previous theories for homogeneous rotating basins9)-12), basin-scale modes can be used as basis functions for generalized Fourier series. For example, averaged velocity in $k$th layer $\vec{v}_k (\vec{x}, t)$ ($k=1$ and $2$ for upper and lower layer) can be decomposed into modal components as
\begin{align} \vec{v}_k=\sum_n{\vec{v}_k^{(n)} a^{(n)}}, \end{align}where $\vec{v}_k^{(n)}(\vec{x})$ is the complex-valued layer-average velocity due to $n$th mode, and $a^{(n)}(t)$ is its amplitude. Then, neglecting nonlinearity and damping, $a^{(n)}$ follows an decoupled modal amplitude equation:
\begin{align} \frac{d a^{(n)}}{dt} =& i \omega^{(n)} a^{(n)} + \frac{f^{(n)}}{e^{(n)}}, \\ f^{(n)}=&\int{}\vec{v}_1^{(n)*} \cdot \vec{\tau} dS, \end{align}where $\vec{\tau}$ is the wind stress vector, $e^{(n)}$ is an arbitrary normalization factor of $n$th mode, and the integral is taken over the lake surface.
Based on this theory, we proposed the following new applications of modal analysis:
Details are available in
The modal analysis has also been applied to Lake Biwa in winter:
This study was conducted when I was a PhD candidate at Centre for Water Research, University of Western Australia, with financial support of Japanese government (MEXT) scholarship and Tokyo-Tech Long-term Overseas Study Support Program.