Landslide investigations in the northwest section of the lesser Khingan range in China using combined HDR and GPR methods
- First Online:
- 1.4k Downloads
In the northwest section of the lesser Khingan range, located in the high-latitude permafrost region of northeast China, landslides are a frequent occurrence, due to permafrost melting and atmospheric precipitation. High-density resistivity (HDR) and ground-penetrating radar (GPR) methods are based on soil resistivity values and characteristics of radar wave reflection, respectively. The combination of these methods, together with geological drilling, can be used to determine the stratigraphic distribution of this region, which will allow precise determination of the exact location of the sliding surface of the landslide. Field test results show that the resistivity values and radar reflectivity characteristics of the soil in the landslide mass largely differ from those of the soil outside the mass. The apparent resistivity values exhibit abrupt layering at the position of the sliding surface, with a sudden decrease in apparent resistivity. In addition, the radar wave shows strong reflection at the position of the sliding surface, where the amplitude of the radar wave exhibits a sudden increase. Drilling results indicate that the soil has high water content at the location of the sliding surface of the landslide mass in the study area, which is entirely consistent with the GPR and HDR results. Thus, in practice, sudden changes in apparent resistivity values and abnormal radar wave reflection can be used as a basis for determining the location of sliding surfaces of landslide masses in this region.
KeywordsLandslide Drilling High-density resistivity (HDR) Ground-penetrating radar (GPR) Sliding surface
In recent decades, geophysical investigations for assessing stratigraphic distribution have become a common tool in geological research. In situ geophysical techniques are able to measure physical parameters directly or indirectly linked with the lithological, hydrological and geotechnical characteristics of the terrains related to the movement (McCann and Forster 1990; Hack 2000; Benedetto et al. 2013). These techniques, less invasive than direct ground-based techniques (i.e., drilling, inclinometer, laboratory tests), provide information integrated on a greater volume of the soil, thus overcoming the point-scale feature of classical geotechnical measurements. Among the in situ geophysical techniques, the high-density resistivity (HDR) and ground-penetrating radar (GPR) methods have been increasingly applied for landslide investigation (McCann and Forster 1990; Malehmir et al. 2013; Timothy et al. 2014).
HDR is based on the measure of electrical resistivity and can provide 2D and 3D images of its distribution in the subsoil. It is one of the standard methods of geophysical prospecting for investigating shallow geological problems (Cardarelli et al. 2003; Yamakawa et al. 2010; Donohue et al. 2011; Sauvin et al. 2014). Current applications of HDR focus on landslide recognition and permafrost detection, while investigations on debris thickness in arctic and alpine environments are comparatively sparse (Carpentier et al. 2012; Donohue et al. 2011; Perrone et al. 2014). GPR is based on the measure of reflection of radar waves in the subsoil, and is largely focused on the fields of natural resource exploration, hydrogeology, engineering and archaeological investigation (Sass 2007; Sass et al. 2008; Schrott and Sass 2008; Zajc et al. 2014).
HDR and GPR are useful for determining characteristics such as the landslide main body, geometry, and surface of rupture, and have been used in landslide investigations since the late 1970s (Hack 2000; Havenith et al. 2000; Bichler et al. 2004; Drahor et al. 2006; Rhim and Chul 2011). However, the applicability of the various geophysical methods, including regional limitations and reliability, in estimating stratum thickness and sliding surface location of landslides in high-latitude permafrost regions in northeast China has not been addressed in detail. Applications of geophysical prospecting on landslides in cold regions of the lesser Khingan range of China have been rare.
Using landslide K178 + 530 as an example, in this paper we present a combination of traditional methods (drilling and mapping) and geophysical techniques (HDR and GPR) applied to the landslide in the lesser Khingan range of northeast China in order to investigate the thickness and internal structure of the landslides that occur frequently as a result of melting permafrost and atmospheric precipitation, in an effort to ascertain the applicability of HDR and GPR on this regional type of landslide.
The geological structure of the study area belongs to the Khingan–Haixi fold belt. From the bottom up, the stratigraphy comprises Cretaceous mudstone, Tertiary pebbly sandstone, silty mudstone and powdery sandstone. From the late Tertiary to the early Quaternary, the lesser Khingan range experienced block uplift. Due to long-term erosion and leveling, loose sediments on the summit and slope of the hills have gradually thinned, and the current residual layer is generally only 1–2 m thick. The loose deposits accumulate mainly in the basin and valley areas between mountains, with a thickness of about 10 m. The soil is mainly composed of clayey silt, mild clay and gravelly sand, and the surface is covered with a relatively thick layer of grass peat and turf. The surface vegetation is grassland and woodland, and there are inverted trees in the woodlands.
Embankment: yellow, mainly composed of loosely mixed Tertiary pebbly sandstones, Cretaceous mudstones and sandy mudstones; soil is loose when dry, plastic when saturated with water. Clay: yellow, soft and plastic, of high strength and toughness when dry. The upstream region of the landslide mass has a depth distribution of 1.5–3.8 m. The downstream region has depth distribution of 0–6.7 m, and there is more than one sandwiched grit layer; the thickness of a single layer is approximately 1–10 cm, which greatly enhances the water seepage capacity of the soil.
Tertiary pebbly sandstones: distributed in the embankment at a depth of 2.0–3.4 m and in the upstream region of the landslide mass at a depth of 3.8–4.5 m, all weathered, composed mainly of feldspar stone and mineral sands, well-graded, highly permeable. Fully weathered siltstones: yellow, distributed in the upstream region of the landslide mass at a depth of 4.5–9.7 m, sandy, of bedding structure and poor water seepage capacity.
Fully weathered mudstone: yellow or gray-green, pelite, of layered structure, easy to soften with water, with poor water seepage capacity. Strongly weathered mudstones: dark gray, pelite, of layered structure, weakly cemented rock. Moderately weathered mudstones: brown, black and gray, pelite, layered structure.
The instrument used in this study was the WGMD-9 Super HDR system (Chongqing Benteng Digital Control Technical Institute, Chongqing, China). With this system, the WDA-1 super digital direct current electric device is used as measurement and control host, and with the optional WDZJ-4 multi-channel electrode converter, centralized high-density cables and electrodes, centralized two-dimensional HDR measurement can be achieved. The inversion of the apparent resistivity data sets that were obtained was performed using the RES2DINV software package, which produces a two-dimensional subsurface model from the apparent resistivity pseudo-section (Loke and Barker 1996). Data were acquired using a Wenner configuration. The method is based on the smoothness-constrained least-squares inversion of pseudo-section data (Tripp et al. 1984; Sasaki 1992; Loke and Barker 1996). In this algorithm, the subsurface is divided into rectangular blocks of constant resistivity. The resistivity of each block is then evaluated by minimizing the difference between observed and calculated pseudo-sections using an iterative scheme. The smoothness constraint leads the algorithm to yield a solution with smooth resistivity changes. The pseudo-sections can be calculated by either finite-difference or finite-element methods (Coggon 1971; Dey and Morrison 1979). In this case, a finite-element scheme was employed due to the topographical changes in the field.
The smoothness-constrained least-squares method used in the inversion model is essentially a method in which the model resistivity is constantly adjusted through model correction in order to reduce the difference between the calculated apparent resistivity and the measured resistivity, and to describe the degree of fit between the two using the mean squared error. This method has been widely applied, and has a number of advantages, including adaptability to different types of data and models, a relatively small influence of noise on the inversion data, high sensitivity to deep units, rapid inversion, and a small number of iterations. In tests using the HDR method, the spacing between unit electrodes was 3.0 m, and the maximum exploration depth of the survey lines was 30 m.
On the road section of landslide K178 + 530, a total of three HDR survey lines, I–I′, II–II′ and III–III′, were established, as shown in Fig. 4. The measuring date was 3 September 2012.
The GPR instrument used was the RIS-K2 FastWave Ground Penetrating Radar (IDS Ingegneria Dei Sistemi S.p.A., Pisa, Italy). The radar antenna was a low-frequency 40-MHz unshielded dual antenna. The detection time window was set to 600 ns, the sampling rate to 1024, and the data acquisition track pitch was 0.05 m. Two GPR survey lines were established, as shown in Fig. 4, with positions coinciding with those of the HDR survey lines, but with different start and end points. The two survey lines (I–I′ and II–II′) were 150 and 118 m in length, respectively. The date of measurement was 1 October 2013. GPR raw data were processed using REFLEXW scientific software (Sandmeier geophysical research, Karlsruhe, Germany). Coordinates for each trace were calculated at equal distances. The surface signal reflection was set to the zero-time position. Low frequencies and noise in the spectrum were filtered using a de-wow and bandpass filter. Next, temporally consistent signals were eliminated utilising background removal, and topographic correction was applied. The picks were exported with the attribute of two-way travel time, and the velocity of propagation of the wave in this case appears to be about 0.10 m/ns.
Results, analysis and discussion
Survey line I–I′
Borehole ZK1 is 115 m from the starting point of the HDR survey line. At a depth of 0–2.1 m, the soil is silty clay and rather loose, containing approximately 15 % grass roots and other organic matter. Resistivity values range from 25 to 45 Ohm m. At a depth of 2.1–6.7 m, the soil is silty clay, and there are local weathered sand layers. Resistivity values range from 45 to 65 Ohm m. At a depth of 6.7–8.0 m, the soil is yellow mudstone. The permeability coefficient is small, and it is difficult for the water to infiltrate downward, forming a watertight layer, and water easily gathers here. The resistivity value is relatively low, 20–30 Ohm m. At a depth of 8.0–26 m, the soil is gray mudstone, close to or below the water table. The resistivity value is relatively low, 10–25 Ohm m. As shown in the curve, silty clay contacts mudstone at a depth of 6.7 m, and resistivity exhibits apparent layering, with a sudden decrease in resistivity value.
The borehole ZK2 is 175 m from the starting point of the survey line. At a depth of 0–4.5 m, the soil is rather loose. Resistivity values range from 45 to 80 Ohm m. The resistivity of the surface embankment soil dominated by silty clay (depth 0–3.8 m) ranges from 60 to 80 Ohm m, and that of gravelly sand (depth 3.8–4.5 m) ranges from 45 to 60 Ohm m. At a depth of 4.5–9.7 m, the soil is siltstone and composed of rather small particles. The permeability is poor, forming a watertight layer, with water easily gathering here. The resistivity value ranges from 25 to 35 Ohm m. At a depth of 9.7–14.6 m, the soil is sandstone, and resistivity values range from 15 to 25 Ohm m. As is evident in the curve, gravelly sand contacts siltstone at a depth of 4.5 m, and there is apparent resistivity layering, with a sudden decrease in resistivity value.
The above-mentioned HDR profiles and resistivity curves show that soil resistivity values above and below the sliding surface of the landslide mass are clearly different and exhibit abrupt stratification. Based on this typical characteristic of the sliding surface, the positions of the major sliding surfaces along line I–I′ were deduced, as shown by the thick black dashed line in Fig. 5. The type of sliding for this landslide was characterized as propelled sliding, and the sliding power originated from the trailing edge of the landslide. The slip rate of the trailing edge was the greatest, followed by the middle part of the landslide; the minimum slip rate occurred at the leading edge (Shan et al. 2015). As a result, secondary sliding occurred in the landslide mass. By combining the changes in soil resistivity values for different positions in the landslide mass and drilling exploration, we obtained the secondary sliding surface, as shown by the thin black dashed line in Fig. 5.
Survey line II–II′
This landslide belongs to a recurring old landslide that slipped again (Shan et al. 2015). The black dashed line in Fig. 8 shows the current sliding surface. Based on the site geological survey, combined with characteristics of resistivity changes, we can infer the position of the sliding surface for the paleo-landslide, as shown by the yellow dashed line in Fig. 8.
Survey line I–I′
Survey line II–II′
As shown in Fig. 13, in positions C, D, ZK1, E and F, amplitude values were substantially increased, exhibiting abrupt change, at depths of 6, 9.5, 9, 7 and 3.5 m, respectively. These positions can be used to deduce the position of the sliding surface of the paleo-landslide, as denoted by the yellow dashed lines in Fig. 12. The magnitude of the abrupt change in radar wave amplitudes was smaller at the position of the sliding surface of the paleo-landslide than that of the current sliding surface.
The GPR results show that the moisture content of soils at the sliding surface of the landslide mass is relatively high. The drilling data also show very high moisture content of the sliding surface of the landslide mass in the study area, which is completely consistent with the results obtained from the GPR and HDR profiles.
Analysis of the mechanism underlying landslide development
Water seepage generated from the melting of permafrost, combined with the infiltration of concentrated summer precipitation, increases the local moisture content in the soil on hillsides, and is the main reason for landslide formation in the northwest section of the lesser Khingan range in China. The landslide in the study area is a shallow creeping consequent landslide in the high-latitude permafrost region. The highly permeable surface soil, the sand and gravel layer, and the silty clay containing a weathered sand interlayer provide a convenient channel for water infiltration, whereas the permeability of the mudstone and siltstone below is very low, forming a watertight layer. Water generated from permafrost melting, precipitation, melting snow and fractured springs rapidly increases the local water content as it infiltrates downward and along the interface between the permeable and impermeable layers to form a slip zone.
Soil resistivity values above and below the sliding surface of the landslide mass are clearly different and exhibit an abrupt stratification. There is apparent resistivity layering at the position of the sliding surface, with a sudden decrease in resistivity value. On the GPR profile, the sliding surface is manifested as a low-frequency, high-amplitude sync-phase axis, and there is a sudden increase in the amplitude of the radar wave. In practice, these abrupt, abnormal changes revealed in the HDR and GPR results can be used as characteristic markers for identifying the sliding surface position of shallow landslides in this region.
Prospecting of landslides in the study area using three methods, namely HDR, GPR and drilling, reveal basically the same sliding surface position, which suggests that HDR and GPR are fast, economical and reliable methods for site prospecting of landslides. They can be applied to shallow landslides in the high-latitude permafrost region for fast and accurate determination of the sliding surface position, providing a useful reference for engineering projects and for ensuring that appropriate measures are taken.
This work was supported in part by the Science and Technology project of Chinese Ministry of Transport (No. 2011318223630) and the Fundamental Research Funds for the Central Universities (No. 2572014AB07).
- Donohue S, Gavin K, Tolooiyan A (2011) Geophysical and geotechnical assessment of a railway embankment failure. Near Surf Geophys 9(1):33–44Google Scholar
- Havenith HB, Jongmans D, Abdrakhmatov K, Trefois P, Delvaux D, Torgoev A (2000) Geophysical investigations of seismically induced surface effects: case study of a landslide in the Suusamyr valley, Kyrgyzstan. Surv Geophys 21:349–369Google Scholar
- Jol HM (2009) Ground penetrating radar theory and applications. Elsevier, AmsterdamGoogle Scholar
- Malehmir A, Bastani M, Krawczyk CM, Gurk M, Ismail N, Polom U, Persson L (2013) Geophysical assessment and geotechnical investigation of quick-clay landslides—a Swedish case study. Near Surf Geophys 11(3):341–350Google Scholar
- Shan W, Hu Z, Guo Y, Zhang C, Wang C, Jiang H, Liu Y, Xiao J (2015) The impact of climate change on landslides in southeastern of high-latitude permafrost regions of china. Front Earth Sci 3(7):1–11Google Scholar
- Zajc M, Pogacnik Z, Gosar A (2014) Ground penetrating radar and structural geological mapping investigation of karst and tectonic features in flyschoid rocks as geological hazard for exploitation. Int J Rock Mech Min Sci 67:78–87Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.