MATERIALS TRANSACTIONS
Online ISSN : 1347-5320
Print ISSN : 1345-9678
ISSN-L : 1345-9678
Characteristics of Fine Particle Distribution at Cut Slope and Fill Slope Comprising Weathered Granite
Hiroyasu OhtsuThirapong PipatpongsaTakafumi KitaokaShunichiro ItoMitsuru YabeSoralump Suttisak
Author information
JOURNAL FREE ACCESS FULL-TEXT HTML

2018 Volume 59 Issue 11 Pages 1723-1730

Details
Abstract

In this study, fine particle distribution in artificial slopes comprising weathered granite, which may affect rainfall-triggered landslide, was investigated comprehensively, based on electrical resistivity, soil composition and unsaturated soil properties. The results showed that while degree of saturation plays a key factor on electrical resistivity in unsaturated soil, it has close correlation to pore-size distribution. Therefore, it can be considered that electrical prospecting is an effective method to investigate distribution of both coarse particle and fine particle. In addition, it was also pointed out that there is possibility that fine particle fraction involved in soil poorly compacted in artificial slopes may be eroded due to rainfall infiltration.

 

This Paper was Originally Published in Japanese in J. Soc. Mater. Sci., Japan 67 (2018) 346–353.

1. Introduction

The frequency of sediment disasters due to rainfall-triggered landslides has been increasing recently in Asian countries, including Japan.14) Based on the classification of previous landslide cases in Thailand according to the geographical conditions, many occurrences have been reported for weathered granite as well as for weathered sedimentary rocks.5) In Asian countries, the geographical conditions for weathered granite are commonly classified into six grades, as shown in Fig. 1, in accordance with the degree of weathering and the unweathered core stone content by percentage.6,7) Jworchan8) pointed out that in terms of the correlation between these geographical classifications and their respective engineering properties, Grade VI and Grade V on the surface layer can be classified as soil, Grade IV under those two grades can be classified as a transition region, and Grade III to Grade I (below Grade IV) can be classified as bedrock. The weathering of granite is classified as primary weathering (mechanical weathering) and secondary weathering (chemical weathering). In other words, after sand and gravels are formed by mechanical weathering, fine-particle fractions (clay fraction and silt fraction) are formed by chemical weathering, mainly triggered by rainfall infiltration into the shallow layer. As a result, as shown in Fig. 2, although the fine-particle fractions are prevalent in Grade VI and Grade V, the sand and gravel fractions are prevalent in Grade IV.8) Also, the results of existing studies have shown that the weathered rock-forming minerals of Grade VI and the Grade V have weak interparticle forces, and thus the shear strength tends to decrease significantly due to the degree of saturation, which is increased by the influence of the fine-particle fractions (silt and clay).9,10) Thus, the authors11) conducted a multi-stage shear test by using undistributed specimens that were classified as Grade VI and taken from weathered granite cut slopes in Phuket (Thailand). Three target saturation values (60%, 80%, and 100%) were used. The results of the test, shown in Fig. 3, showed that the cohesion and internal friction angle decreased with increasing degree of saturation.

Fig. 1

Classification of granite in accordance with degree of weathering.6)

Fig. 2

Grain size distribution for each grade.8)

Fig. 3

Results of multi-stage shear test.11) (a) Cohesion, (b) Internal friction angle.

Based on the findings regarding the ground characteristics attributable to granite weathering as mentioned above, the mechanism of the formation of rainfall-triggered landslides on a natural slope of weathered granite was examined. As water retention and permeability due to rainfall in regions where fine-particle fractions (clay and silt fraction) predominate are higher than those in regions where coarse-particle fractions (gravel and sand fraction) predominate, the infiltration water is temporarily stored. This causes a decline in shear strength due to the increase in the degree of saturation, which decreases the stability of a slope. According to a report by Phien-wej et al.,10) who studied a landslide formed on a slope in Nakhon Si Thammarat in the southern part of Thailand in 1988, the main sliding surface was formed near the upper-limit surface of a Grade IV transition region. To determine the degree of risk of landslides caused by this weathering property, the application of electrical resistivity survey methods to identify fine-particle fraction distribution and water retention has been recently increasing.1214)

The factors affecting the ground resistivity values obtained from electrical resistivity surveys include types of ground materials, porosity, resistivity of pore water, degree of saturation, clay content due to weathering, and changes in rock quality.15) For example, Archie16) defined a function where the ground resistivity was proportional to the resistivity of the pore water and inversely proportional to the exponent of porosity and degree of saturation. However, it is known that the resistivity obtained as the result of surveys is lower than the value estimated by the Archie equation because the clay fraction formed due to weathering and changes in rock quality is conductive.17)

In terms of factors affecting the resistivity, as mentioned above, Giao et al.12) examined how the fine-particle fraction distribution formed by weathering changed according to the gradient of the slope, an idea originally proposed by Ruxton and Berry.18) Giao et al.’s12) study was based on the resistivity distributions in the electrical survey results for weathered granite slopes in Phuket (Thailand), as shown in Fig. 4. Ha et al.14) showed the difference distributions in electrical surveys conducted twice, once in rain and once in dry weather, for a natural slope of Hiroshima granite. The results, shown in Fig. 5, revealed that the decline in the degree of saturation is small because the infiltration water is temporarily stored at the border of the highly weathered layer and the slightly weathered layer. Thus, the change in resistivity is smaller than that in other areas.

Fig. 4

Distribution of electric resistivity on a weathered granite slope.12)

Fig. 5

Reduction rate for resistivity values in a weathered granite slope.14)

However, the study by Ha et al.14) described above did not clearly discuss the influence of the degree of saturation as a result of rainfall permeation in unsaturated soil. Accordingly, the current study examines the influence of degree of saturation in unsaturated soil on resistivity based on the principles described below.

It is common in geotechnical engineering to evaluate the engineering characteristics of unsaturated soil by using the soil–water characteristic curve (SWCC), which relates the degree of saturation of the ground to suction. Fredlund and Xing19) showed that the SWCC in the drainage process expresses a function of the pore-size distribution characteristic. Also, Elkady et al.20) related the characteristics to the pore-size distribution characteristics as described below. The water–suction relation (i.e., the water content in the drainage process of a sand–clay mixture in an unsaturated state, as shown in Fig. 6) changes depending on the clay content. In other words, ground pores are classified as macro-size pores, which are equivalent to the coarse-particle fractions, and meso-/micro-size pores, which are equivalent to the fine-particle fractions.21) According to the results shown in Fig. 6, Elkady et al.20) established a rough standard of suction ranges equivalent to macro-size pores and meso-/micro-size pores of 0.1–100 kPa and 0–1500 kPa, respectively. They examined the characteristics of the increase in the meso-/micro-size pores due to the increase in clay content (increase in the fine-particle fraction). Water content (w) is used as an indicator of water retention in Fig. 6. Given that w has a linear relation with the degree of saturation (Sr) (w = eSr/Gs, where e is the void ratio and Gs is the specific gravity of the soil), it can be said that a similar argument is valid for the relation between the degree of saturation and suction.

Fig. 6

Suction-gravimetric moisture content relationship.20)

Based on these principles, in this study, we investigated the resistivity obtained from electrical surveys conducted at a cut slope/fill slope of weathered granite by using the particle-size composition as the engineering property and linking the pore-size distribution to the relation between the degree of saturation and suction.

2. Results of Surveys at a Cut Slope

The target cut slope referred to in Section 2 was a slope constructed (height of slope: 6.0 m and gradient of slope: 37.5°) by cutting weathered granite along the highway in Phuket (Thailand).22) The tests described below were carried out approximately two years (in 2011) after the construction of the slope (in 2009). Figure 7 shows an overview of the target slope and the survey lines. On the target slope, test pores were installed at two places (TP-1: toe of the slope on survey line 4, TP-2: in the middle of survey line 4), and a KU-Miniature sampling device23) was used to collect undisturbed specimens in the pores for water retention tests at depths of 0.4 m, 0.6 m, and 1.0 m. Refer to the paper by Ohtsu et al.11) for details. Also, disturbed specimens for particle-size tests, etc., were collected at the intersection of survey line 3 and survey line B by implementing hand auger sampling. Both types of specimens were collected in a dry season.

Fig. 7

Field monitoring survey lines in Phuket site.22)

2.1 Results of electrical surveys

As shown in Fig. 7, electrical surveys (Wenner installation, at electrode intervals of 1.0–2.0 m) were performed at the set survey lines (survey lines A to D and survey lines 1 to 4) to determine the resistivity profile by inversion analysis.

Figure 8 shows the results of electrical resistivity surveys on survey lines 3 and 4 in the transverse direction. The resistivity in most areas was between 15 and 865 Ω·m; however, the resistivity exceeded 865 Ω·m in some areas. It is considered, as shown in the study by Little6) in Fig. 1, that these areas had some residual blocks (shown as a “Boulder” in Fig. 1) on the unweathered matrix as weathering advanced along the discontinuous surface of the granite. Moreover, in the resistivity range between 15 and 865 Ω·m, a range with relatively low resistivity was distributed along the blocks, as described above.

Fig. 8

Results of electrical prospecting.22) (a) Resistivity at survey line 3, (b) Resistivity at survey line 4.

Accordingly, it is considered that weathering on the target slope reflected the natural weathering state on the discontinuous surface before cutting and that most parts consisted of material that changed into soil and sediment (Grade VI to Grade V) according to the classification provided in Fig. 1. We will see that the resistivity near the slope surface is higher than that of the areas underneath by linking resistivity to the engineering properties.

2.2 Engineering properties

In this section, we examine the correlation between resistivity and fine-particle content. We carried out particle-size tests using the specimens collected with the hand auger (near the intersection of survey line 3 and survey line B). Figure 9 shows the distributions of the clay fraction and the sand/gravel fraction in the direction of depth based on the test results.24) The figure also shows the classification according to the degree of weathering estimated from the particle-size composition (Grade VI–Grade IV). As shown in Fig. 9, the fine-particle fraction (clay fraction) had a relatively small dispersion, between approximately 10 and 20%. It peaked to 30–35% at the maximum between GL −1.0 m and GL −2.0 m on the surface layer of the slope, whereas the fraction tended to decrease with increasing depth. As a result, the sand/gravel fraction reached the lowest value between GL values of 1.0 m and 2.0 m, where the permeability range was relatively low.

Fig. 9

Distribution of soil particle composition.24) (a) Clay fraction, (b) Sand/gravel fraction.

The tendency of particle-size composition distributions in Fig. 9 almost matched the results obtained at a natural slope, as shown in Fig. 2. However, on the target cut slope, the fine-particle fraction (clay fraction) near the surface of the slope was different from the results shown in Fig. 2, and the fine-particle fraction was lower than the areas underneath, showing a relatively small dispersion. This tendency agrees with the results shown in Fig. 8, where the resistivity near the surface of the slope is higher than that in the areas underneath. Therefore, it is considered that the fine-particle fraction in the surface layer on the target slope exhibited outflow in a relatively short period of time or approximately two years after cutting, as shown with the arrow in Fig. 9. As the permeability of the surface layer increased due to this outflow of the fine-particle fraction, rainfall infiltration was promoted in this area. Furthermore, it is estimated that the phenomenon of temporary storing of infiltration water is relatively enhanced in regions where the fine-particle fraction predominates in the lower part with the outflow of the fine-particle fraction.

Next, we examined the relation between suction and degree of saturation in the drainage process according to the measurements with undisturbed specimens collected from the two test pores (TP-1 and TP-2), as described earlier. In this test, suction within the specimens was changed by sealing the specimens collected in the undisturbed state in a PVC pipe and by decreasing (or increasing) the pressure in the pipe. For suction measurements, a miniature tensiometer developed by Kasetsart University was used. Refer to the paper by Jotisankasa et al.25) for details.

Recent studies on SWCC pointed out that the drainage process of unsaturated soil is affected by the changes in the void ratio due to increased suction (essentially equivalent to the volume change).26,27) Therefore, in this study, three types of undisturbed specimens, PK-18, PK-40, and PK-04, were collected as typical examples for relatively great, relatively small, and intermediate void ratios after soaking, respectively, as shown in Table 1.

Table 1 Soil properties.

Figure 10 shows the suction–void ratio relation of the three types of undisturbed specimens. The void ratios of specimens with relatively great void ratios after soaking (PK-18 and PK-04) tended to slightly decrease with increased suction; however, the overall changes were small. Accordingly, for the unsaturated soil collected from the target slope with a suction of 100 kPa or less, the influence of changes in the void ratio on the SWCC in the drainage process was very small.

Fig. 10

Suction-void ratio relationship.

Next, Fig. 11 shows the suction–degree of saturation relationship of the three types of undisturbed specimens. In addition to the survey results, a curve to fit the results by using the Van Genuchten29) (noted as VG in the figure) model by the SWRC fit28) method is also shown. Furthermore, the figure shows the suction range, although as a rough standard, equivalent to macro-size pores and meso-/micro-size pores in Fig. 6.

Fig. 11

Suction-degree of saturation relationship.

The suction–degree of saturation according to the pore-size distribution shown in Fig. 11 is in harmony with the results shown in Fig. 6, as described below.

In PK-04 and PK-40, there were many meso-/micro-size pores and the water retention was high, which led to a small decline in the degree of saturation due to the increase in suction. However, in PK-18, the number of meso-/micro-size pores was smaller (equivalent to many macro-size pores) and the water retention was lower than that of the other two specimens, which led to a significant decline in the degree of saturation due to the increase in suction. In the function developed by Archie,16) as mentioned earlier, resistivity is inversely proportional to the exponent of the degree of saturation. Accordingly, the resistivity was low for materials such as PK-04 and PK-40, where the fine-particle fraction predominated because the amount of decline in the degree of saturation even in a dry state is small. On the other hand, the resistivity increased for materials such as PK-18 due to drying (equivalent to an increase in suction).

As a result, when the resistivity of pore water is constant, changes in the degree of saturation are dominant in the changes in resistivity due to drying/infiltrating in unsaturated soil. However, the changes themselves depend on the distribution of pore sizes in the soil, namely, the fine-particle fraction, and it is thus considered that argument by separating the changes in degree of saturation and the fine-particle fraction is not appropriate.

3. Results and Surveys at a Fill Slope

The target fill slope referred to in Section 3 was a small-scape fill slope that was constructed (height of slope: 6.0 m and gradient of slope: 30.3°) with weathered granite residual soil at a test farm of Kasetsart University near the top of Mt. Doi Pui in the northwestern part of Chiang Mai (Thailand). Figure 12 shows an overview of the target fill slope and the survey lines. On the target slope, sampling with a hand auger was carried out at the place marked with a circle in Fig. 12 to collect disturbed specimens for particle-size tests, etc. The collecting of undisturbed specimens for water retention tests was carried out using the same method as in the tests in Phuket. Both types of specimens were collected in a dry season.

Fig. 12

Field monitoring survey lines in Chiang Mai site.

Although no detailed construction records on the target slope remain, it was confirmed on a satellite photo that at least twelve years had passed since the construction. Also, according to interviews with the parties concerned, the initial ground form was a mild slope of weathered granite on a mountain ridge. The ridge was drilled to construct a reservoir (concrete water tank), and the residual soil was used for fill after the construction of the water tank. The filling work was carried out manually, without using any construction equipment such as a rolling machine. According to the results of ground surveys implemented by Soralump et al.1) on the target slope,2) the fill layer on the upper part of the weathered granite was 2.25 m thick, the property was classified as silty sand (SM)–clayey sand (SC), and the void ratio was 1.519, and thus there was a relatively large fine-particle fraction.

3.1 Results of electrical surveys

As shown in Chap. 2, although the ground structure of the cut slope reflected the natural weathering state before cutting, the fill slope was structured with remolded soil materials. Therefore, in order to focus on changes in the ground state on the shallow part of the slope due to rainfall infiltration, the electrode intervals to one survey line were set to 0.2 m to ensure high density, and electrical surveys were performed thrice for one survey line (dry season: May 2015, rainy season 1: September 2015, and rainy season 2: July 2016). A pole–pole arrangement was used for the surveys, and inversion analysis was used to determine the resistivity profile.

Figure 13(a) shows the changes in resistivity between the dry season and rainy season 1. A significant decline in resistivity occurred in the 0.2–1.0 m depth range. Although Fig. 13(b) shows the analysis results using the measurement data for electrode intervals of 1.0 m alone, any region with a clear decline in resistivity was not confirmed in this figure. Next, Fig. 14 shows the changes in resistivity between the dry season and rainy season 2, and a significant decline in resistivity is confirmed. We will examine these resistivity distribution properties based on the engineering properties in the next section of this paper.

Fig. 13

Difference on electric resistivity between dry season and rainy season 1.

Fig. 14

Difference on electric resistivity between dry season and rainy season 2.

3.2 Engineering properties

Figure 15 compares the grain-size accumulation curves of the specimens collected at three depths (GL −0.2 m, GL −0.4 m, and GL −0.6 m) on the target slope. Although a significant fine-particle fraction exists at GL −0.2 m, the fine-particle fractions at GL −0.4 m and GL −0.6 m were less than that at GL −0.2 m. Given that a significant decline in resistivity is associated with the GL range of 0.4 m to 0.6 m, as shown in Fig. 14, the fine-particle fraction is lost and is lower than that in shallower places due to certain factors (such as insufficient compacting of the target slope by manual construction and the resultant outflowing of the fine-particle fraction).

Fig. 15

Comparison of grain size accumulation curves.

Next, this section describes the unsaturated property in the drainage process of undisturbed specimens collected from GL −1.0 m, which is equivalent to the region with a clear decline in resistivity. The test method used to identify the unsaturated property was the same as that used for the cut slope described in the preceding section.

Figure 16 shows the suction–void ratio relationship of the specimens (noted as CM in the figure). It also shows the results of specimens on the cut slope described in the preceding section (noted as PK-18 in the figure) for comparison. In the suction–void ratio relationship of CM, the void ratio as a function of suction in the drainage process is almost constant. This result indicates that the changes in the void ratios of the unsaturated soil collected from the target slope within the range of suction of 100 kPa or less had little influence on the SWCC.

Fig. 16

Comparison of suction-void ratio relationship.

Figure 17 shows the suction–degree of saturation relationship of the specimens on the target slope (noted as CM in the figure). The suction range, although as rough standard equivalent to macro-size pores and meso-/micro-size pores in Fig. 6, is shown. Also, the results of specimens on the cut slope in the preceding section (denoted as PK-18 and PK-40 in the figure) are shown for comparison. Note that original suction–degree of saturation relationship of unsaturated soil varies depending on the sampling location. However, in this study, we included the results of specimens collected from different places, as shown in Fig. 6, to compare the distribution properties among macro-size pores and meso-/micro-size pores.

Fig. 17

Comparison of Soil Water Characteristic Curves.

The results shown in Fig. 17 indicate that CM had more macro-size pores than PK-18 and PK-40, which led to a significant decline in the degree of saturation due to the increase in suction. Accordingly, it is estimated that CM has a high resistivity in a dry state. However, as mentioned in the preceding sections, in materials with many meso-/micro-size pores such as PK-40, the decline in the degree of saturation due to the increase in suction is small and the resistivity is low, even in a dry state.

According to the examination of the engineering properties as described above, the resistivity variation properties as shown in Figs. 13 and 14 are reviewed as follows. Although the resistivity increases in the dry season in the range with a small fine-particle fraction due to a significant decline in the degree of saturation, the resistivity decreases in the range containing a large fine-particle fraction. However, although the resistivity decreases due to a significant increase in the degree of saturation in the rainy season in the range with a small fine-particle fraction due to rainfall infiltration, the increase in the resistivity is very small because of the small increase in the degree of saturation in the range containing a large fine-particle fraction. Accordingly, as shown in the difference distributions of resistivity between the dry season and the rainy season, it is considered that the fine-particle fraction causes a clear difference in the properties. However, the range where the fine-particle fraction is lost on the target slope may be formed by the outflowing of the fine-particle fraction, and explicit descriptions of this formation mechanism require additional examination.

As the region with a small fine-particle fraction on the target slope was located in the shallow layer, as shown in Figs. 13(a) and (b), it is estimated that it was detected only when the electrode interval was 0.2 m in a high-density state. Therefore, it is estimated that the setting of electrode intervals is an important factor to identify the ground structure in an extremely shallow region, as in this study.

4. Conclusion

In this study, fine-particle distributions formed by weathering in cut/fill slopes comprising weathered granite, which may affect slope stability, were investigated comprehensively by relating the resistivity results of electrical surveys, the grain-size composition as the engineering property, the suction–degree of saturation relationship, and the pore-size distribution characteristics. The findings obtained as a result are summarized as follows:

  1. (1)    The dominant factor on resistivity in unsaturated soil is the degree of saturation. However, changes in the degree of saturation depend on the size distribution characteristics of pores in soil; that is to say, although macro-size pores equivalent to the coarse-particle fraction have a significant impact on the degree of saturation, meso-/micro-size pores equivalent to the fine-particle fraction have a small impact on the degree of saturation. These characteristics allow us to identify the planar distribution characteristics of the coarse-particle fraction and fine-particle fraction in the direction of depth on a slope by focusing on the difference in resistivity between the dry state and the wet state. It is estimated that identifying these distribution characteristics contributes to the identification of the relative distribution of rainfall infiltration, especially in the two-dimensional direction on a slope, namely, to the estimation of a sliding plane. As in this study, it is also estimated that setting of electrode intervals is an important factor to identify the ground structure of an extremely shallow range.
  2. (2)    Given the interpretation of the resistivity distributions at the target cut/fill slopes in this study (based on the findings obtained in,1) it is surmised that the fine-particle fraction may have been lost by surface runoff/infiltration of the slope surface/shallow layer part. The estimated factors for this are loosening on the surface due to cutting on the cut slope and insufficient compacting on the fill slope. However, it is also possible that frequent heavy rain with high rainfall intensity specific to the region may have induced the outflow of the fine-particle fraction.

REFERENCES
 
© 2018 The Society of Materials Science, Japan
feedback
Top