**2. Experiment and Data**

In this work, data were collected during the 2016 Internal Solitary Wave Cooperative Observation Experiment. The experiment was conducted in the north of the South China Sea from July 2016 to July 2017. Cooperative observation, consisting of oceanographic and acoustic instrumentation, was adopted in the experiment. A subsurface mooring system was deployed at a station labeled IW5 (117.87◦ E, 21.11◦ N) with a water depth of 1000 m, which was shown in Figure 1. Oceanographic data measured by thermistor chains and the acoustic doppler current profilers (ADCPs) were continuously observed. Both upward and downward-looking ADCPs were mounted on the mooring system at depth of 511 m in order to record flow velocity information. The temporal resolution was 3 min, and the vertical resolution was 16 m, covering a depth range of 110~850 m. The mooring

system was equipped with thermistor chains of temperature loggers and Conductivity-Temperature-Depth (CTD) recorders between 110–930 m to collect temperature and salinity data. The temporal resolution was also 3 min. was equipped with thermistor chains of temperature loggers and Conductivity-Temperature-Depth (CTD) recorders between 110–930 m to collect temperature and salinity data. The temporal resolution was also 3 min.

which was shown in Figure 1. Oceanographic data measured by thermistor chains and the acoustic doppler current profilers (ADCPs) were continuously observed. Both upward and downward-looking ADCPs were mounted on the mooring system at depth of 511 m in order to record flow velocity information. The temporal resolution was 3 min, and the vertical resolution was 16 m, covering a depth range of 110~850 m. The mooring system

*J. Mar. Sci. Eng.* **2022**, *10*, x FOR PEER REVIEW 3 of 14

**Figure 1.** The topography in the South China Sea. The red pentagram indicates the location of the experimental site (IW5 station). The right figure shows the subsurface mooring system. **Figure 1.** The topography in the South China Sea. The red pentagram indicates the location of the experimental site (IW5 station). The right figure shows the subsurface mooring system.

The acoustic data were measured by a mooring-mounted hydrophone array attached on the cable. The array consisted of five independent acoustic hydrophones, spanning a depth of 200–950 m (Figure 1). The commercial hydrophones used on the array were independent acoustic hydrophones manufactured by the Institute of Acoustics of the Chinese Academy of Sciences. All hydrophones on the array were calibrated before experiment. Their received voltage responses were characterized. The sampling rate of the hydrophone was 4 kHz, and the sound pressure sensitivity of the hydrophone was −175 dB. The sensitivity value was fulfilled between 8 Hz to 2000 Hz without decaying more than −2 dB. Considering the storage and device power, the hydrophone was operated in intermittent mode (15 min every 3 h). The acoustic data were measured by a mooring-mounted hydrophone array attached on the cable. The array consisted of five independent acoustic hydrophones, spanning a depth of 200–950 m (Figure 1). The commercial hydrophones used on the array were independent acoustic hydrophones manufactured by the Institute of Acoustics of the Chinese Academy of Sciences. All hydrophones on the array were calibrated before experiment. Their received voltage responses were characterized. The sampling rate of the hydrophone was 4 kHz, and the sound pressure sensitivity of the hydrophone was −175 dB. The sensitivity value was fulfilled between 8 Hz to 2000 Hz without decaying more than −2 dB. Considering the storage and device power, the hydrophone was operated in intermittent mode (15 min every 3 h).

ISWs can cause large depressions on the isothermal surface and notable increases in the flow velocity. As shown in figure 2, the duration time of temperature and velocity segments was 1 h, spanning a depth of 110−850 m. The influence time of ISWs on seawater temperature was about 15 min. The amplitudes of different ISWs varied greatly, from approximately 90 m to 160 m. The characteristic data for the three different ISWs focused in the study are shown in Table 1. ISWs can cause large depressions on the isothermal surface and notable increases in the flow velocity. As shown in Figure 2, the duration time of temperature and velocity segments was 1 h, spanning a depth of 110−850 m. The influence time of ISWs on seawater temperature was about 15 min. The amplitudes of different ISWs varied greatly, from approximately 90 m to 160 m. The characteristic data for the three different ISWs focused in the study are shown in Table 1.


**Table 1.** Characteristic data for ISWs on three different days. **Table 1.** Characteristic data for ISWs on three different days.

The mooring-mounted hydrophone array and thermistor chains were configured together. It took more than ten minutes for the internal wave to pass the mooring system, so we needed to find the noise data, with internal waves occurring and acoustic records, The mooring-mounted hydrophone array and thermistor chains were configured together. It took more than ten minutes for the internal wave to pass the mooring system, so we needed to find the noise data, with internal waves occurring and acoustic records, for analysis. In spite of the experiment period of one year, we only found a total of 6 events that matched the requirement, all of which were analyzed. Three with large internal wave amplitudes were selected for analysis, and the rest of the data also showed similar phenomena (Figure 2). Figure 3 depicts the data received by the C058 hydrophone at

a depth of 700 m between 18:00 UTC+8 to 18:15 UTC+8 on 1 October 2016. Figure 3a shows the time-domain signal and the obvious noise burst when the internal wave arrived at 18:03 UTC+8. The time–frequency spectrum (Figure 3b) was obtained by short-time Fourier transform of the data. This analysis used a Hamming window of length 4096, with a 4096-point fast Fourier transform (FFT) and an overlap of 50%. The label of the time–frequency spectrum was power spectral density (PSD). Noise at about 500 Hz and some high frequency lines were generated by the shaking of the hydrophone and the collision of suspended particles in the water against the hydrophone casing when ISWs flowed past the mooring system. of 700 m between 18:00 UTC+8 to 18:15 UTC+8 on 1 October 2016. Figure 3a shows the time-domain signal and the obvious noise burst when the internal wave arrived at 18:03 UTC+8. The time–frequency spectrum (Figure 3b) was obtained by short-time Fourier transform of the data. This analysis used a Hamming window of length 4096, with a 4096 point fast Fourier transform (FFT) and an overlap of 50%. The label of the time–frequency spectrum was power spectral density (PSD). Noise at about 500 Hz and some high frequency lines were generated by the shaking of the hydrophone and the collision of suspended particles in the water against the hydrophone casing when ISWs flowed past the mooring system. of 700 m between 18:00 UTC+8 to 18:15 UTC+8 on 1 October 2016. Figure 3a shows the time-domain signal and the obvious noise burst when the internal wave arrived at 18:03 UTC+8. The time–frequency spectrum (Figure 3b) was obtained by short-time Fourier transform of the data. This analysis used a Hamming window of length 4096, with a 4096 point fast Fourier transform (FFT) and an overlap of 50%. The label of the time–frequency spectrum was power spectral density (PSD). Noise at about 500 Hz and some high frequency lines were generated by the shaking of the hydrophone and the collision of suspended particles in the water against the hydrophone casing when ISWs flowed past the mooring system.

for analysis. In spite of the experiment period of one year, we only found a total of 6 events that matched the requirement, all of which were analyzed. Three with large internal wave amplitudes were selected for analysis, and the rest of the data also showed similar phenomena (Figure 2). Figure 3 depicts the data received by the C058 hydrophone at a depth

for analysis. In spite of the experiment period of one year, we only found a total of 6 events that matched the requirement, all of which were analyzed. Three with large internal wave amplitudes were selected for analysis, and the rest of the data also showed similar phenomena (Figure 2). Figure 3 depicts the data received by the C058 hydrophone at a depth

*J. Mar. Sci. Eng.* **2022**, *10*, x FOR PEER REVIEW 4 of 14

*J. Mar. Sci. Eng.* **2022**, *10*, x FOR PEER REVIEW 4 of 14

**Figure 2.** Temperatures and flow velocities for ISWs passing by the IW5 station on three different days. (**a**,**c**,**e**) Temperatures measured by thermistor chains. (**b**,**d**,**f**) Flow velocities measured by the ADCPs. The flow velocity from west to east is positive and the reverse is negative. The flow velocity in this work refers to the zonal velocity. All data sampled at UTC+8. **Figure 2.** Temperatures and flow velocities for ISWs passing by the IW5 station on three differentdays. (**a**,**c**,**e**) Temperatures measured by thermistor chains. (**b**,**d**,**f**) Flow velocities measured by the ADCPs. The flow velocity from west to east is positive and the reverse is negative. The flow velocity in this work refers to the zonal velocity. All data sampled at UTC+8. **Figure 2.** Temperatures and flow velocities for ISWs passing by the IW5 station on three different days. (**a**,**c**,**e**) Temperatures measured by thermistor chains. (**b**,**d**,**f**) Flow velocities measured by the ADCPs. The flow velocity from west to east is positive and the reverse is negative. The flow velocity in this work refers to the zonal velocity. All data sampled at UTC+8.

**Figure 3.** The signal received by the C058 hydrophone at a depth of 700m from 18:00 UTC+8 to 18:15 UTC+8 on 1 October 2016. (**a**) Time–domain signal. (**b**) Time–frequency spectrum. (**c**) Time–frequency spectrum between 0 Hz and 100 Hz. **Figure 3.** The signal received by the C058 hydrophone at a depth of 700m from 18:00 UTC+8 to 18:15 UTC+8 on 1 October 2016. (**a**) Time–domain signal. (**b**) Time–frequency spectrum. (**c**) Time–frequency spectrum between 0 Hz and 100 Hz. **Figure 3.** The signal received by the C058 hydrophone at a depth of 700m from 18:00 UTC+8 to 18:15 UTC+8 on 1 October 2016. (**a**) Time–domain signal. (**b**) Time–frequency spectrum. (**c**) Time–frequency spectrum between 0 Hz and 100 Hz.

It was notable that the low-frequency noise presented an identifiable frequency fluctuation at frequencies below 100 Hz, and the acoustic intensity increased with decreasing frequency. A further finding was that low-frequency noise had uniform harmonics, as depicted in Figure 3c (bright stripes). The generation mechanism for these features is discussed in the next section.
