Introduction

China is a very large country, extending some 5,500 km north to south. In the south, the rainfall is adequate for the needs of the population, but the north is deficient in rainfall, relative to the demands of the population and industry. In the Chongqing area, for instance, the annual rainfall is 1.15 m, while in Beijing it is only 0.428 m and more seasonal, falling mainly in summer.

As a consequence, the Chinese Government is investigating the practicality of diverting some water from the South to the North. In particular, it is planned to divert some of the flow from the Yangtze River into the Yellow River. Three diversion routes are being considered; the West Route, Centre Route and East Route (Fig. 1).

Fig. 1
figure 1

The route map of diverting water from Yangtze River to Yellow River

Full size image

This paper is concerned only with the West Route, part of which is between the Ma-ke and Jia-qu Rivers (tributaries of the Yangtze and Yellow Rivers). Here the topography generally varies between 3,450 and 4,600 m asl and is deeply incised with many high, steep slopes. The diversion into the Yellow River will be through the Bayankela Mountain.

Geology

The area, in south west China, is part of the Tibetan plateau. As such it has suffered/is experiencing significant geostresses related to the movement of the Indian Plate toward the Eurasian Plate.

As the Indian Plate is extending eastwards and being subducted beneath the Eurasian Plate, the study area on the edge of the Tibetan Plateau is rising in the order of 10–30 mm/year. A feature of its particular geographical position is that within the area in which the western route tunnel will be driven, the direction of the main stress field changes from ENE to become almost E–W (Fig. 2).

Fig. 2
figure 2

Tectonic characteristics of the research area

Full size image

A view of the Upper Triassic geology of the area is seen in Fig. 3. It consists of a calcareous sequence of gritstones (coarse sandstone) and mudrocks. In view of the regional stresses, the mudrocks have taken on a pronounced slatey cleavage. A feature of the geology is that the relative proportions of gritstones and slate will differ along the tunnel profile.

Fig. 3
figure 3

The alternating gritstone and slate

Full size image

In this area of high geostress, tight folding as well as extensive fracture development/faulting are common features and as it was anticipated a TBM would be used, the planning engineers were concerned that such fractures could be water bearing.

Method

In an attempt to better evaluate the geology, a geophysical examination of the bedrock geology was considered appropriate. However, standard geophysical techniques were unlikely to be appropriate in this terrain, hence in order to establish the geology at the depth of the tunnel, a controlled source audio-frequency magnetotelluric method (CSAMT) was considered. This method uses an artificial signal source at ground level or in a borehole with a working frequency of up to 10,000 Hz, allowing an exploration depth of 1,500 m. With the signal source with strong current controlled by a transmitter and the high signal–noise rate, the CSAMT is considered very stable. It has been used extensively since the 1970 s for mineral exploration (Boerner et al.1993; Basokur et al. 1997). It has also been used for underground water exploration (Wu and Shi 1996), geological disasters in mining (Di et al. 2002a, b); geothermal exploration (Sandberg and Hohmann 1982, Wannamaker 1997) and in volcanic areas (Batrel and Jacobson 1987). Zonge et al. (1986) and Zonge and Hughes (1991) have summarized the electromagnetic (EM) exploration.

Following its initial development (Goldstein 1971; Goldstein and Strangway 1975), research on the CSAMT method has led to significant improvements, particularly with near-field correction (Yamashita et al. 1985); terrain correcting (Macinner 1987); static effects (David et al. 1993; Jimmy et al. 2003) and inverse methods (Sasaki et al. 1992; Routh and Oldenburg 1996; Lu et al. 1999).

Survey line position

As seen in Fig. 4, the two areas assessed using CSAMT are part of the 173.7 km Shang-jia line. The survey was divided into two parts:

Fig. 4
figure 4

Survey line position for geophysical exploration

Full size image

Section A is 7.35 km long and orientated N074°E.

Section B (to the East of the River Zu-mu-zu) is 12.975 km long and orientated N060°E.

The tunnel portals at either end of Section A will be at 4,050 and 3,696 m asl, while the portals at Section B will be at 4,095 and 3,814 m asl. As the upper tributaries of the Yangtze River are some 80–500 m lower than the Yellow River, the diversion project involves establishing some dam/storage facilities in order to raise the water level, before it passes in a long tunnel through the Bayankela Mountain.

Investigation

A V6A system produced by Canada Phoenix Geophysics Ltd. was used. The basic principles are shown in Fig. 5. The system uses a 20 kW generator and transmitter with a working frequency varying between 0.125 and 8,192 Hz. There is a digital receiver linked to a computer, a magnetic coil and a non-polarization electrode. At any time a receiver can simultaneously accept the different frequencies of seven electrical components (Ex) and one magnetic component (Hy).

Fig. 5
figure 5

Sketch of CSAMT observation system

Full size image

As noted above, two lines were surveyed. It should be appreciated that in Section A, Station 0 is in the East and Station 294 is some 25 m away. In Section B, the soundings began in the West (Station 0) with Station 519 some 25 m away (see Fig. 7). The frequency ranges were between 211 and 21 Hz. Two transmitting stations were used for each line.

When undertaking CSAMT exploration in a mountainous region, it is essential to ensure that the horizontal difference and vertical distance between the measuring electrodes is accurate. As a consequence, a RTK double frequency GPS system was used which can record the accuracy to within 100 mm. In view of the isolation of the area, the background noise was low, such that a good, smooth data curve could be obtained.

Results and discussion

Figure 6 shows the frequency-apparent resistivity curve (ρ a) and the frequency-phase curve (ϕ) at one sounding point, based on the original data. The horizontal co-ordinate is the position of the measuring points with the vertical axis recording the frequencies. In the computer plot, the change in resistivity between blue (low) and red (high) indicates the variation in the geology. Figure 7 shows a typical plot for Sections A and B, with blue anomalies appearing almost as a series of fingers.

Fig. 6
figure 6

Curve of apparent resistivity and phase with frequency

Full size image
Fig. 7
figure 7

Field data pseudo section of frequency and apparent resistivity for profiles East and West: (a) Section A; (b) Section B

Full size image

The quality of the results depends on both the reliability of the data and the processing methods/techniques. Figure 8 shows the steps taken in the data processing. (1) Anomalous results were noted particularly where there was a pronounced topographic break (e.g., cliffs) and also where, in retrospect, it appeared the electrode contacts had been unsatisfactory. These “wild data” were omitted from the assessment. (2) In this mountainous area, the static effect is very strong, hence low-frequency filtering was undertaken using the Han-ning filter (Bostick 1986) The filter function is

$$ h(x) = \left\{ \begin{gathered} {\frac{1}{\omega }}\left(1 + \cos {\frac{2\pi x}{\omega }}\right)\mathop {}\nolimits_{{}} \mathop {}\nolimits_{{}} \left| x \right| \le {\omega \mathord{\left/ {\vphantom {\omega 2}} \right. \kern-\nulldelimiterspace} 2} \hfill \\ 0^{{}} \mathop {}\nolimits_{{}} \mathop {}\nolimits_{{}} \mathop {}\nolimits_{{}} \mathop {}\nolimits_{{}} \mathop {}\nolimits_{{}} \mathop {}\nolimits_{{}} \mathop {}\nolimits_{{}}^{{}} \left| x \right| > {\omega \mathord{\left/ {\vphantom {\omega 2}} \right. \kern-\nulldelimiterspace} 2} \hfill \\ \end{gathered} \right. $$
(1)

where ω is the length of the filter window and h(x) can be discretized into seven filter points. (3) Terrain correction in such a mountainous area is extremely important (Takeshhi 1953; Xu and Zhao 1985; Jiracek 1990). In order to assess the magnitude of the effect induced by the topography, a 2D model was prepared and the electric field (E x) and magnetic field (H y) and the Cagniard apparent resistivity calculated. In order to minimize the effect induced by the installation of the instrument, H y was measured horizontally and E x measured in the slope direction, such that the magnetic field (H y) is unchanged and the electric field (E x) is equal to

$$ E_{\text{x}} = E_{\text{l}} \cos \theta = {\frac{\Updelta V}{\Updelta l}}\cos \theta = {\frac{\Updelta V}{\Updelta x}}\cos^{2} \theta $$
(2)

where E x is the corrected electric field, θ is slope angle, ΔV is measured potential difference of the slope; ΔV is the slope distance interval; Δx is the horizontal distance interval. It can be seen from Eq. 2 that the influence of the installation can be bigger than the influence created by the topography. (4) Having removed the wild data and made the terrain correction, a 2D inversion was undertaken using trial and error methods. The results are shown in Fig. 9 which indicates that the static and terrain effects have been significantly reduced.

Fig. 8
figure 8

Sketch of data processing

Full size image
Fig. 9
figure 9

Interpretation of West (a) and East (b) profiles

Full size image

Geological interpretation

The CSAMT method is an electromagnetic method based on resistivity. In general, in addition to the lithology, the geological structure and the degree of saturation will influence the resistivity values obtained. The previous work (Wu and Shi 1996; Di et al. 2002a, b) supported by measurements of resistivity on rock outcrops within a short distance of the site have indicated

  1. (a)

    Quaternary sediments, <100 ohm-m

  2. (b)

    sandstone/gritstone, 4,000–5,000 ohm-m

  3. (c)

    alternating layers of sandstone and slate, 1,000–2,000 ohm-m

  4. (d)

    limestone, 100–8,000 ohm-m

  5. (e)

    Tertiary sands and pebbles—300–800 ohm-m

Low-resistivity values were found in areas of karstic limestone and where the sandstones were fractured but were fully saturated. The resisitivity of the dominantly slatey material varied significantly, depending on the relative proportion of gritstone present.

Figure 9 indicates a close agreement between the integrated interpretation along the West and East profile and the engineering geology assessed in the field.

West profile

Based on the resistivity characteristics, the stratum along the West profile was divided into three segments (Fig. 9a).

  1. (a)

    In the section between Stakes 294 and 170, four low-resistivity anomalies exist—WL1′, WL2′, WL3′ and F2′. The slate component is less and there is a cleavage in the sandstone layer. The profile is approximately parallel to the strike. The lower anomaly may be a volcanic dyke. In view of the obvious intraformational slip and fracture, this may be a water bearing horizon.

  2. (b)

    Close to Stake 118 is a thin, straight, low-resistivity feature (STD1′) which may indicate a high-density zone.

  3. (c)

    Adjacent to Stake 55, there is a low-resistivity feature (STD2′) considered to be a fracture.

  4. (d)

    Between Stakes 42 and 23, there is a significant low-resistivity anomaly (WL4′) which would warrant further investigation.

East profile

The stratum along the East profile can be divided into four segments, including one fault and five crush belts. The anomalies (Fig. 9b) can be explained as:

  1. (a)

    Near Stake 114, there is a clear low-resistivity zone which inclines eastwards and extends downwards compared with the high-resistivity background. The zone is referred to as STD1.

  2. (b)

    Between Stakes 160 and 183, the low-resistivity zone expands and the resistivity character changes. Fault FTD1 and fracture STD2 are part of this anomaly zone, WL1. The cementation of the fault zone is limited, hence there is a replenishment of surface water. Fault FTD1 is also apparent as a lower-resistivity area.

  3. (c)

    There is a discontinuous resistivity zone around Stake 334, referred to as STD3.

  4. (d)

    There is an obvious low-resistivity strip near Stake 39 (named STD4). The potential influence of this should be considered in future studies;

  5. (e)

    Adjacent to Stake 443, there is an obvious low-resistivity anomaly (STD5); this is particularly clear near the planned tunnel. Again this should be further investigated.

Conclusions

CSAMT has been used to detect the geology and any weak structures at depth along a line where a shield tunnel will be constructed between the Ma-ke and Jia-gu Rivers. The survey line has been divided into the East and West sections. After static and terrain corrections, an overall assessment of the geology indicated five fractures, five faults and one anomalous zone in the East profile; and two fractures, three faults and four anomalous zones in the West profile. The results are consistent with the outcrop geology as determined during fieldwork.

The results presented in this paper show the combination of CSAMT geophysical prospecting and engineering geological mapping can be an effective way to study the lithology of the rocks and give valuable information for tunnel design and construction.