1、Project description
On July 8, 2021, the towed seismic and TEM methods were used to test the stability of a dam in Yinhushan Country Park, Shenzhen, China. For towed seismic data acquisition, the length of the measurement line is 100 m, the offset distance is 3.0 m, the channel spacing is 0.5 m, the gun spacing is 1 m, the sampling interval is 41.7 us, and the sampling points are 8192. Each gun position is hit twice, and a total of 176 shots data are collected. A schematic diagram of field data acquisition is shown in Figures 1-1 and 1-2. For TEM data acquisition, the transmit current waveform was set as a bipolar square wave, with a duty cycle of 50% and a frequency of 16 Hz. The acquisition frequency was 1.25MHz, and the duration of a single sampling was about 16ms. The operating mode of the instrument was continuous mode, and the superposition point was 0.5m apart. There were 683 equivalent measurement points for Jigongshan Dam.

Figure 1-1. Schematic diagram of towed seismic data acquisition in the field.

Figure 1-2. Schematic diagram of towed transient electromagnetic data acquisition in the field.

2、Data Processing
2.1 Towed seismic data
In view of the original towed seismic data, the inverse scattering, surface wave and reflected wave superposition methods are applied to process the data, and the imaging results shown in Figure 2-1 are obtained. The inverse scattering profile (FIG. 2-1, (a)) showed anomalous body structures at lateral positions 5 m-10 m, 45 m, and 88 m, but the specific depth position was not determined. The shear wave velocity profile (Figures 2-1, (b)) shows that the depth position has a clear velocity interface at about 10 m (D), and the velocity variation of the upper and lower layers is bounded by 350 m/s (E). There is an obvious low-velocity area in the transverse position of 85 m-90 m. Combined with the structural data of the dam, this low-velocity area is the location of the culvert pipe. The lateral position of 5 m-10 m, and the position of 45 m have abrupt changes in velocity, which correspond to the position of strong energy in the inverse scattering profile (Figure 2-1, A and B).
The back-scattered profile has high lateral resolution but low longitudinal resolution. The effective depth of surface wave inversion method is up to 15 m, which has not reached the interface between silty clay and quartz sandstone. Based on the analysis of S-wave velocity, the average p-wave velocity was calculated to be 1200 m/s. At a depth of 15 m, there is a messy reflection axis (G), the reflection interface inside the silty clay. At a depth of 15 m, the silty clay is divided into upper and lower layers. The continuous reflection axes H and I represent the interface between silty clay and quartz sandstone, where I varies in depth from 21 m to 30 m. J is the reflection inside the quartz sandstone, and the depth varies from 29 m to 36 m.

Figure 2-1. Towed seismic Detection Profile: (a) Inverse scattering profile; (b) shear wave velocity profile; (c) Superimposed section.

2.2 TEM data
The inversion results of the TEM data are shown in Figure 2-2. The depth is from 20 m to 30 m, and the transverse position is from 20 m to 60 m (A). The upper layer is low resistance, and the lower layer is high resistance. Combined with the structural data of the dam, the low resistance layer is silty clay layer, and the high resistance layer is quartz sand rock layer. The transverse position is 75 m to 90 m, and the depth ranges from 10 m to 30 m. The upper layer is high resistance, while the lower layer is low resistance. Combined with the structural data of the dam, the high resistance layer is silty clay, and the continuous low resistance is the influence from the culverts. The lateral location of the culvert is approximately in area C.

Figure 2-2. TEM inversion section

3、Conclusions
Due to the low water level of the reservoir, the water content on both sides of the dam is low, and the transient electromagnetic detection results have no obvious low resistance abnormal characteristics. The silty clay interior is generally high resistance, and local areas are affected by the metal culcults, which is a continuous reflection of low resistance. Combined with the towed seismic and transient electromagnetic inversion profiles, the following conclusions are drawn:
(1) There were abnormal positions from 5 m to 20 m and 45 m in the lateral position, and the depth position was about 10m, but the specific abnormal body was not determined. The culvert is located about 88 m laterally.
(2) The interface between silty clay layer and quartz sand layer is obviously reflected in seismic superposition section and transient electromagnetic inversion section.
(3) The towed earthquake has high resolution and deep detection depth, and can distinguish the reflection axis at a depth of about 15 m in the silty clay layer and the reflection axis at a depth of 30 m to 36 m in the quartz sand rock layer.