Fig. 1. A bird's-eye-view of obliquely interacting solitary waves (left) with and (right) without nonlinear amplitude enhancement.
Fig. 2. Relationship between extended Miles interaction parameter $\kappa$ and amplification factor $a_s$. Solid line shows relationship for $\sigma$=0.3, and dashed lines for $\sigma$=0.0, 0.6, and 0.9. Vertical dotted lines indicate lower limit of applicable $\kappa$ for given $\sigma$. Figure taken from Shimizu and Nakayama (2017, J. Geophys. Res.-Oceans).
Fig. 3. A snapshot of modelled vertical velocity in the Andaman Sea (Full Coriolis case). (a) Model output from 0.005° grid, and (b) that from 0.0025° grid in region of interest, shown in rotated coordinates defined in (a). Note that only negative velocity is shown and that colour scale is nonlinear. Click image to see the movie (1.9MB, not for small screen).
Fig. 4. Parameters relevant to oblique ISW interaction. (a) Amplitudes of incident and stem-like ISWs, (b) incident angle, (c) extended Miles interaction parameter, and (d) amplification factor. Note that we use negative amplitude for ISW of depression. Shading indicates uncertainty in Full Coriolis case. Uncertainties in Traditional and No Coriolis cases are similar to Full Coriolis case. Reformatted from Shimizu and Nakayama (2017, J. Geophys. Res.-Oceans).
Fig. 5. Sea-surface-height (SSH) expression of modelled oblique ISW interaction under M2 tidal forcing (Full Coriolis case, amplitude unadjusted for topographic variation). (a) SSH in the region shown in Fig. 3b, and (b) profiles of SSH along solid, dashed, and dotted lines shown in panel (a). Green line in (a) shows ISW trough line calculated from vertical velocity. Mean sea level and M2 tide are removed by subtracting mean of modelled SSH time-series at each grid point, and then by least-square fitting of sinusoidal function with M2 tidal frequency to SSH time-series. Figure taken from Shimizu and Nakayama (2017, J. Geophys. Res.-Oceans).
In the fields of surface water waves and plasma, it is well-known that the oblique interaction of two solitary waves could enhance the amplitude by a factor of up to 4 (i.e., twice the linear superposition) through the development of a (Mach) stem1),8),12),14),15),17)(Fig. 1). However, this process has seldom been investigated for internal solitary-like waves (ISWs, or internal solitons) in the ocean, despite observations of its surface signature by remote sensing4),11),16),18),19). We investigate oblique ISW interaction in the Andaman Sea with particular focus on the effects of topography and Earth's rotation, because they make the oblique ISW interaction in the ocean different from that in other fields of physics. This study began when my collaborator, Prof. Keisuke Nakayama, identified the occurrence of oblique ISW interaction in early results of my MITgcm ISW modelling, and introduced me to the topic.
To analyse oblique ISW interaction under realistic oceanic conditions, we use the Miles theory13) but with the following extensions:
The theory for the asymptotic state (after sufficiently long interaction) depends on the following parameters: \begin{align} a_s&=\frac{A_s}{\bar{A}}, \\ \kappa&=\frac{\text{tan} \Psi}{\sqrt{2\alpha \bar{A}/c} \text{cos} \Psi}, \\ \sigma&=\frac{1}{2}\sqrt{\frac{\Delta A}{\bar{A}}}, \\ \sqrt{\alpha \bar{A}}&=\left( \sqrt{\alpha A_2}+\sqrt{\alpha A_1} \right)/2, \\ \sqrt{\alpha \Delta A}&= \sqrt{\alpha A_2}-\sqrt{\alpha A_1}, \end{align}
where $A_1$, $A_2$, $A_s$ are respectively the amplitude of the first incident, second incident, and stem-like ISWs, $\alpha$ is the quadratic nonlinear coefficient in the KdV theory, and $\Psi$ is the incident angle. The theory shows that the maximum amplification $a_s=4$ occurs when $\kappa=1$ (Fig. 2).
For our 3-D ISW modelling, we use MITgcm10), realistic topography and stratification, and (spatially-varying) monochromatic M2 tidal forcing. To investigate the effects of Earth’s rotation, we consider the following 3 cases:
The model output for Full Coriolis case (Fig. 3) shows that the interaction of ISWs with ~10-m and ~20-m amplitudes ('Incident 1' and 'Incident 2') produced an ISW with >50-m amplitude (labelled as 'Stem-like'), confirming the nonlinear enhancement of ISW amplitudes under realistic topographic and Coriolis effects. After adjusting the amplitudes for significant lateral topographic amplitude variation (by a factor of up to 2) and considering the transitional response, the simulated interaction is found to be approximately consistent with the extended Miles theory (i.e., $a_s \approx 4 $ when $\kappa \approx 1$ after sufficiently long interaction in Fig. 4). Although Earth's rotation has little effect on ISW propagation due to the low latitude (~9°N), Earth's rotation has considerable effects on the oblique interaction.
Details are available in
Data are available in
This modelling project was partly funded by Research Connections Grant from Australian Department of Industry, Innovation, and Science, RPS MetOcean Pty Ltd., and CSIRO.