Effects of preferential flow induced by desiccation cracks on slope stability

https://doi.org/10.1016/j.enggeo.2021.106164Get rights and content

Highlights

  • A full-scale model test was carried out

  • Characteristics of desiccation cracks strongly depend on their positions at the slope and rainfall-evaporation cycles.

  • Desiccation cracks can significantly increase the infiltration depth and trigger slope failure or even landslides

  • Preferential flow can be identified from the response sequence of hydrologic sensors

Abstract

Desiccation cracks on a soil slope can significantly increase permeability, reduce shear strength, and potentially result in shallow landslides. To reveal the slope failure mechanism induced by desiccation cracks, a full-scale model test was conducted on a cracked soil slope under rainfall–evaporation cycles. Image processing techniques were used to quantify the crack characteristics at the slope crest (SC), around the slope shoulder (SS), and at the slope foot (SF), and hydrologic sensors were used to monitor the moisture content, matric suction, and pore water pressure at different depths in the crack areas. The results showed dynamic variations in the desiccation crack patterns in accordance with their position on the slope and the rainfall–evaporation cycle. Preferential flow induced by the desiccation cracks in response to rainfall was detected earlier by the lower hydrologic sensors than the upper ones, and the desiccation cracks significantly increased the infiltration depth by up to four or five times the crack depth. Experimental evidence confirmed that preferential flow through desiccation cracks can trigger slope failure or landslides by forming local perched water zones near the crack tips. Based on this investigation, the failure process of the cracked soil slope was separated into three stages according to the crack patterns and failure modes: (I) generation of desiccation cracks, with surface erosion as the failure mode; (II) development and transformation of cracks, with flow-slip and local failure as the failure modes; and (III) renewal and further development of cracks, with overall failure as the failure mode. These conclusions suggest that when simulating the seepage and stability of a cracked soil slope, further modifications should consider the dynamic changes that occur within desiccation cracks. In addition, the use of specific treatment measures to avoid slope failure during the different stages is suggested.

Introduction

Desiccation cracks occur in soil that contains large amounts of hydrophilic minerals, such as illite and montmorillonite, and the mechanical behaviour of such soil is sensitive to variations in the atmospheric environment (Yassoglou et al., 1994; Krisnanto et al., 2014; Kong et al., 2017, Kong et al., 2018). Rainfall-triggered shallow landslides occur frequently on the cracked soil cutting slopes near highways, which constitute hazards to infrastructures and cause significant maintenance problems (Titi and Helwany, 2007; Khan et al., 2016).

However, although shallow landslides can occur on a cracked soil slope with a gentle inclination (Hou et al., 2013; Dai et al., 2017; Xiao et al., 2017; Xie et al., 2020; Pei et al., 2020), the associated failure mechanism has not been determined. Numerous researchers have proposed that slope failure occurs because desiccation cracks act as pathways for rainwater to rapidly infiltrate deep soil and form preferential flow. ((Caris and Van Asch); Hencher, 2010; Li and Zhang, 2011; Khan et al., 2016; Damiano et al., 2017; Xie et al., 2020; (Zhang et al., 2021)). However, some field engineers have argued that desiccation cracks on the surface of most collapsed cracked soil slopes are fine and shallow, and they are thus not deep enough to enable rainwater to deep soil and cause a shallow landslide with a depth of several meters. In addition, (Favre et al., 1997) and (Liu et al., 2003) also found a significant reduction in the infiltration rate when the cracks were closed after wetting. Therefore, the mechanism involved in slope failure via desiccation cracks remains a subject of debate.

Researchers have studied the mechanism involved in rainfall-triggered landslides using numerical modelling (Gasmo et al., 2000; Cai and Ugai, 2004; Lora et al., 2016; Elkamhawy et al., 2018), in-situ experiments (Kong et al., 2007; Zhang et al., 2014; (Cui et al., 2017, Cui and Jiang, 2019)) and laboratory model tests (Take et al., 2004; Schnellmann et al., 2010; Wu et al., 2018). However, the first two methods have inherent problems. For example, numerical models often regard the slope as a homogeneous and isotropic medium, and it is difficult to simulate the non-uniform infiltration process caused by the dynamics of desiccation cracks. In situ experiments are time-consuming, and it is obviously difficult to impose environmental boundary conditions and induce landslides. In contrast, laboratory model tests enable the control of the soil properties and boundary conditions, and also facilitate the monitoring of the infiltration and failure process; therefore, this method is the most reliable for use in studying rainfall-triggered landslides.

Wang and Sassa (2001), Olivares and Damiano (2007) and Damiano and Olivares (2010) conducted laboratory model tests to investigate the conditions required for the occurrence of flow slides, and they analysed the behaviour of pore water pressure and slope displacement. Tohari et al. (2007), Wu et al. (2015) and Chueasamat et al. (2018) conducted small-scale 1 g flume tests and Take et al. (2004) and Kohgo et al. (2011) conducted small-scale centrifuge slope model tests to investigate the effect of density, water level and inclination angle of slope on the deformation and failure mechanism of landslides during artificial rainfall. The results of these model tests indicated that increased pore water pressure and decreased effective stress in the soil caused by rainfall infiltration were the most critical mechanisms underlying slope failure or landslides (Zhang et al., 2019). However, the previous experiments have inherent problems, which are as follows: first, most of the slope models were constructed using loose granular materials because of the ease of triggering slides or flow (Pei et al., 2020), and this may not accurately reflect the real failure characteristics of clayey soil slopes with low permeability, such as cracked soils. Second, to investigate slope stability, only a few small-scale model tests. Second, to investigate slope stability, only a few small-scale model tests (Wang et al., 2005; Amenuvor et al., 2018; Chen, 2019) and in-situ experiments (Xie et al., 2020; Pei et al., 2020) were conducted under rainfall-evaporation cycles to induce desiccation cracks on the soil slopes. These experiments either failed to trigger landslides or they only focused on the developmental characteristics of the desiccation cracks on the slope crest and surface runoff; therefore, limited insights into the internal infiltration process and failure mechanism of the cracked soil slope were obtained. Third, reduced scale and centrifuge model tests are problematic with respect to the scale effects, similarity relations, and the disruptive effects of sensors and their cables (Moriwaki et al., 2004). In summary, it is extremely difficult to reproduce the correct stress field of soil slopes in the laboratory (Jia et al., 2009; Pei et al., 2020).

A full-scale model test was conducted in this study. A cracked soil slope was subjected to rainfall–evaporation cycles, and the characteristics of the desiccation cracks at the slope crest (SC), around the slope shoulder (SS), and the slope foot (SF) were quantified. In addition, the moisture content, matric suction, and pore water pressure at different depths under the crack areas were monitored using hydrologic sensors to identify the effects of desiccation cracks on the infiltration process. The results enabled elucidation of the failure mechanism of the cracked soil slope, and corresponding solutions were thus proposed.

Section snippets

Engineering geological and project background

The study site is located within the Jianghuai hilly region along the Yangtze River. The region spans a length of 194 km, covers an area of 7334 km2, and is a remnant of the Ta-pieh Mountains in Anhui Province, China (Fig. 1). It originates in Anqing City and extends northeastward through the cities of Chizhou, Tongling, Wuwei, Chaohu and finally reaches Ma'anshan City (Cheng et al., 2014). The area lies within a transitional zone from subtropical to warm temperate climate. The average annual

Design of model test system

The full-scale model in this study was designed based on an in-situ cutting slope (i.e., with a height of 3.2 m and an angle of 45°) at the maintenance centre of S22 road that failed between 22 June and 11 July 2017 (Fig. 1 e). The experiment was conducted at a site near the in-situ slope, and the set-up mainly consisted of a mobile artificial precipitation simulator, a steel flume with glass walls, a slope model, and a monitoring station(Fig. 2), which are described below.

The mobile artificial

General observation of slope failure

The entire slope failure process during the nine rainfall events is described in Table 4, and Fig. 8 shows typical photographs of the slope failure. Slope failure primarily included four modes: 1) surface erosion (during rainfall events 1–3); 2) flow-slip and local failure (during rainfall events 4–8), where local failure with a maximum sliding depth of approximately 26 cm (Fig. 8b2) was initiated during rainfall event 5 near the crack monitoring area of SS; 3) overall failure (during rainfall

Crack patterns on slope surface

Desiccation cracks will appear once the tensile stress, generated from the increased matric suction induced by water evaporation, exceeds the tensile strength of the soil (Qi et al., 2020, Yoshida and Adachi, 2004). The polygonal structure has been widely considered as a typical plane pattern of desiccation cracks (Li and Zhang, 2011; Tang et al., 2011a, Tang et al., 2011b). However, our experimental results revealed that the crack patterns were significantly affected by their positions on the

Conclusions

A full-scale slope model test was conducted under rainfall–evaporation cycles, and the characteristics of the crack patterns and hydraulic variations were studied in relation to the infiltration process and failure evolution. The following conclusions were made based on observations and analyses:

  • (1)

    The patterns of desiccation cracks were substantially affected by their positions on the slope and the progression of the rainfall–evaporation cycles. Desiccation cracks on the slope crest maintained a

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

This work was financially supported by the Plan of Anhui Province Transport Technology Progress (grant 2018030) and Anhui Transportation Holding Group Co., Ltd. (grant JKKJ-2017-20). It was also partially funded by the Fundamental Research Founds for National University, China University of Geosciences (Wuhan) (grant 1810491A24). China Railway Investment road and Bridge Co., Lt is acknowledged for their great help in building the experimental set up.

References (61)

  • P. Pei et al.

    A protective measure for expansive soil slopes based on moisture content control

    Eng. Geol.

    (2020)
  • W. Qi et al.

    Crack closure and flow regimes in cracked clay loam subjected to different irrigation methods

    Geoderma

    (2020)
  • R. Schnellmann et al.

    Effect of rising water table in an unsaturated slope

    Eng. Geol.

    (2010)
  • C.S. Tang et al.

    Desiccation and cracking behavior of clay layer from slurry state under wetting-drying cycles

    Geoderma

    (2011)
  • C.S. Tang et al.

    Experimental characterization of shrinkage and desiccation cracking in thin clay layer

    Appl. Clay Sci.

    (2011)
  • N. Yassoglou et al.

    Cracking in recent alluvial soils as related to easily determined soil properties

    Geoderma

    (1994)
  • S. Yoshida et al.

    Numerical analysis of crack generation in saturated deformable soil under-row-planted vegetation

    Geoderma

    (2004)
  • J. Zhang et al.

    Shallow slope stability evolution during rainwater infiltration considering soil cracking state

    Computers and Geotechnics

    (2020)
  • C.L. Zhang et al.

    Rainfall infiltration in Chinese loess by in situ observation

    J. Hydrol. Eng.

    (2014)
  • M. Zhang et al.

    Field model experiments to determine mechanisms of rainstorm-induced shallow landslides in the Feiyunjiang River basin, China

    Eng. Geol.

    (2019)
  • Z.B. Zhang et al.

    Characteristics of cracks in two paddy soils and their impacts on preferential flow

    Geoderma

    (2014)
  • B.A. Albrecht et al.

    Effect of desiccation on compacted natural clays

    J. Geotech. Geoenviron.

    (2001)
  • A.C. Amenuvor et al.

    Shrinkage cracking in physical model of undisturbed expansive clay slope subjected to wet-dry cycles

  • F. Cai et al.

    Numerical analysis of rainfall effects on slope stability

    Int. J. Geomech.

    (2004)
  • J.P.T. Caris et al.

    Geophysical, geotechnical and hydrological investigations of a small landslide in the French Alps

    Engineering Geology

    (2021)
  • K.S. Chen

    Evolution law of fissures of red clay slope under wet-dry cycles

    J. Archit. Civil Eng.

    (2019)
  • X.F. Cheng et al.

    Spatial-temporal change of flood disaster risk of area along the Yangtze River in Anhui Province

    J. Nat. Dis.

    (2014)
  • Y Cui et al.

    Investigation of the initiation of shallow failure in widely graded loose soil slopes considering interstitial flow and surface runoff

    Landslides

    (2019)
  • Y Cui et al.

    Experimental study on the moving characteristics of fine grains in wide grading unconsolidated soil under heavy rainfall

    J. Mt. Sci.

    (2017)
  • Z. Dai et al.

    The failure characteristics and evolution mechanism of the expansive soil trench slope

    PanAm. Unsaturated Soils

    (2017)
  • Cited by (0)

    1

    Professor Chikhotkin Victor, is a foreign teacher employed from other school, and he doesn’t possess an institutional e-mail address of the China University of Geosciences (Wuhan).

    View full text