Engineering Geology 91 (2007) 100 116 www.elsevier.

com/locate/enggeo

Assessing slope protection methods for weak rock slopes in Southwestern Taiwan
Der-Her Lee a,b , Yi-En Yang a,c,, Hung-Ming Lin d
b a Department of Civil Engineering, National Cheng Kung University, Tainan, Taiwan Sustainable Environment Research Center, National Cheng Kung University, Tainan, Taiwan c Department of Construction Engineering, Nan-Jeon Institute of Technology, Tainan, Taiwan d Department of Construction Technology, Leader University, Tainan, Taiwan

Received 3 July 2006; received in revised form 15 December 2006; accepted 20 December 2006 Available online 10 January 2007

Abstract Failures of weak-rock slopes adjacent to roadsides in southwestern Taiwan most often occurred during or immediately after heavy rainfall. Field survey of weak rock slopes along the national South-2 Freeway conducted in this study showed that the slope protection methods employed in the study areas mainly included vegetation (82.0%), prestressed rock anchors with vegetation (6.8%), grille beam (5.9%), rock anchors with grille beam (3.6%). The highest failure rate occurred in the slopes that were protected by the vegetation method. The most frequently encountered weak rock formation along the South-2 Freeway is the alternating sandstoneshale formation (36.3%), followed in sequence by sandstone (24.4%), conglomerates (21.9%) and mudstone (17.4%). The field survey also found that the mudstone slopes present the highest failure rate among all rock types, and the most commonly encountered modes of failure were surface erosion and shallow slides. Factors affecting slope failure include inadequate drainage of storm water runoff, disparate rock types and vegetation on slope surfaces, slope angles and heights. This paper presents results of the field survey of the rate of failures of weak rock slopes in Southwestern Taiwan and examines the attributes of slope failures and the effectiveness of commonly used slope protection methods in the region. Requirements or essential features of an effective slope protection method are then presented along with the preliminary results of its field implementation. 2007 Elsevier B.V. All rights reserved.
Keywords: Weak rock; Mudstone; Slope failure; Slope characteristics; Slope protection methods

1. Introduction Weak rock or soft rock is a geological material harder and stronger than engineering soils, but not behaving as a hard rock. Whether a rock is a weak rock is commonly judged with the rheological properties of the rock, namely, the deformability, the uniaxial compressive
Corresponding author. Department of Construction Engineering, Nan-Jeon Institute of Technology, Tainan, Taiwan. E-mail address: yien@mail.njtc.edu.tw (Y.-E. Yang). 0013-7952/$ - see front matter 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.enggeo.2006.12.005

strength and shear strength and its time dependency (Oliveira, 1993). Mudstone is a poorly indurated weak rock having the texture and composition of shale, but lacking its fine lamination and fissility. According to Goodman (1993), mudstone is the preferred name for a silt/clay sedimentary rock that lacks lamination or fissility, although the name of mudrock is popular in many parts of the world (Cripps and Taylor, 1981; Dick and Shakoor, 1992). Durability is the most important property of mudrocks in projects that involve exposure of mudrocks to

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116

101

weathering. The presence of nondurable mudrocks in an environment of severe weathering is generally recognized as the main reason for the instability problems of mudrock slopes (Dick and Shakoor, 1992). Physical and engineering properties of mudstones have been studied and reported in many parts of the world (Dick and Shakoor, 1992; Chang et al., 1996). This paper presents the results of field study involving the collection of geological data along the South-2 (S-2) Freeway in the southwestern region of Taiwan detailing the effectiveness of existing slope protections. Because of the scarcity of land for economic

development in southwestern Taiwan, the S-2 Freeway and its network highways were routed through unstable weak rock foothills, characterized by geological deposits consisting mainly of Neogene debris sedimentary rock (Fig. 1). The rock types along the S-2 Freeway, to the south of Meishan, mainly consist of sandstone, alternating sandstoneshale formation, and mudstone. These weak rocks pose a significant challenge in terms of hillside roadway maintenance. Sandstones are loosely cemented, mudstone can easily be softened and slaked when it gets wet, and alternating sandstoneshale formation is geologically unstable (Lee et al., 1996). The

Fig. 1. Geological map along S-2 Freeway Central Geological Survey, MOEA (2000).

102

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116

adverse characteristics of weak rocks, coupled with the steep, high and long slopes created during the construction of the freeway, have exacerbated the already unstable nature of these foothills in southwestern Taiwan. Consequently, slope failures regularly occurred both during and following the construction of the S-2 Freeway. Among the weak rock slopes in the study area, mudstone slopes are particularly susceptible to erosion. Distinct wet and dry seasons characterize the weather in this area. Heavy rains during the summer and dry weather during the fall and winter severely impacts local terrain, resulting in hillsides nearly devoid of vegetation.

Consequently, severe erosion coupled with poor vegetation results in extremely unstable mudstone slopes. Large amounts of rainfall during the typhoon season compound these erosion difficulties, resulting in near constant slope failures and debris flows. For instance, a slide failure occurring on steep slopes in a rockanchored area at mileage post-300 km + 600 of the northbound South-2 Freeway in July 1997 was induced by heavy rainfall into the mudstone slopes. Similar incidents occurred on the northbound Freeway as well, between markers 362 km + 200 to 364 km + 800 in 1999 and 2000, and markers 373 km + 200 and 373 km + 700 during July to August 2001. Another slide occurred near

Fig. 2. Geological map of the Southwestern mudstone area near Tainan, Taiwan.

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116 Table 2 Results of the swelling pressure tests in the study area Test method Swelling pressure (kgf/cm2) CPC Kaohsiung base

103

the south Jhongliao Tunnel portal, at 380 km + 700 to 381 km + 000 of the southbound S-2 Freeway involving a grille beam as a result of torrential rains loosening the already unstable soil (CEC, 1999; Chen, 2002). The attributes of slope failures and the effectiveness of commonly used slope protection methods in the region will be first examined followed by a subsequent treatment of the requirements and essential features necessary for effective slope protection design. The ultimate goal of the study was to develop an effective and economic slope protection method based on the characteristics of mudstone and climatic conditions in the Southwestern Taiwan. 2. Characteristics of mudstone and other weak rocks in Southwestern Taiwan 2.1. Distribution of mudstone in Southwestern Taiwan In Taiwan, exposed mudstone formations largely predominate in the foothills of the southwestern region of the country, the southern part of the East Coast Mountain Range, and the Hengchun Peninsula. The southwestern region has the largest area of mudstone exposure, ranging from the city of Chiayi in the north to Kaohsiung in the south and covering an area over 1000 m2. Fig. 2 shows the distribution of mudstone in this region (Lee et al., 1996; Wang and Huang, 2002). This mudstone range was largely formed in the Miocene, Pliocene and Pleistocene, and as such, it is prone to swelling and slaking with water, and has weak erosion resistance to stream flows, surface runoffs and rainfall. Severe erosion often results in bald landforms with gullies and exposed surfaces (Chang et al., 1996). 2.2. Physical and chemical properties of the mudstone Chemical composition of this mudstones includes SiO2 (63.49%), Al2O3.Fe2O3 (21.53%) and CaO (2.71%). Primary mineral components of the mudstone, based on X-ray diffraction testing results of Tsai (1984), include illite (30.54%), chlorite (28.70%) and quartz (28.45%). Secondary mineral constituents include feldspar, calcite and kaolinite. The particle size distriTable 1 Results of the slake durability test in the study area Tianliao mudstone Fresh mudstone Fresh mudstone Surface weathered mudstone Surface weathered mudstone Id1 (%) 73.6 81.3 68.5 69.2 Id2 (%) 51.7 65.6 44.5 51.1

Tianliao

Fresh mudstone Remolded mudstone Fresh mudstone Different pressure 2.86 method Strain-controlled 1.39 swelling test 0.90 0.57 4.0

bution of the weathered mudstone shows the average consistency is 49% of silt-size particles, 29% of claysize particles, and 22% of sand-size particles. The index properties of the mudstone are: LL (liquid limit) = 38, PL (plastic limit) = 24, Gs (specific gravity) = 2.74, d (dry unit weight) = 1.77 gf/cm3 (17.4 kN/m3), and e (void ratio) = 0.51. The weathered mudstone in southwestern Taiwan is a low-plasticity clay, classified as CL according to the Unified Soil Classification System (USCS) (Lee et al., 1996). 2.3. Slaking and swelling characteristics of the mudstone While possessing relatively high retention strength in dry conditions, the strength of mudstone can be drastically reduced when it absorbs too much water. As water infiltrates the mudstone, the diagenetic bonds between mudstone grains are gradually destroyed. As these bonds are destroyed, the recoverable strain energy stored in the compressed and deformed mudstone grains is released (Bjerrum, 1967). As the space between these particles is widened, an associated swelling of the clay minerals also occurs, resulting in swelling and slaking phenomena of the mudstone (Lee et al., 1994). Table 1 shows results of the slake durability test (Franklin et al., 1979) of the mudstone taken from Tianliao, Kaohsiung County. The first-cycle slake durability index values of fresh mudstone (i.e., Id1), are 73.6% and 81.3% respectively, falling within the range of 60 to 85% under the category of low durability as suggested by Gamble (Goodman, 1980). The secondcycle slake durability index values (Id2) are 51.7% and 65.6%, which are within the ranges of 30 to 60% and 60 to 85% respectively, so they fall under low to medium durability categories. The slake durability index values of surface weathered mudstone in both cycles fall within the range of low durability. If an external force is imposed to constrain the swelling of the mudstone, pressure resulting from this

104

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116

Table 3 Comparison of slope protection methods in the study area Stabilization methods Functions Erosion protection Slope protection methods Vegetation 1 Grille beam 1 Fabric form 1 Shotcrete 1 Green coating 1 Stone masonry 1 Geological stability methods Gabion 3 Prestressed rock anchor 3 Concrete retaining wall 3 Rating: 1-good, 2-acceptable, 3-bad. Geological stability 3 3 3 3 3 3 Ecological considerations 1 2 2 3 3 3 Natural material considerations 2 3 3 3 3 3 Inclination limit 3 2 2 1 1 2 b35 slope b45 ~60 slope b45 ~60 slope Preventing water from entering the slope Preventing water from entering the slope b45 ~60 slope Suitability

2 1 2

2 3 3

2 3 3

2 1 2

Stabilize slope toes Deal with geological instability Increase slope stability

swelling can be measured. Table 2 shows results of the swelling pressure. The mudstone taken from CPC Kaohsiung base with the different pressure method and the strain-controlled swelling test method (Proter and Nelson, 1980) show that the swelling pressure is in the range of 1.4 to 2.9 kgf/cm2 (137 kN/m2 to 284 kN/m2). The swelling pressure of the remolded mudstone is much lower than that of fresh mudstone (Lee et al., 1994, 1996). Furthermore, as with the different pressure method, the swelling pressure of fresh mudstone obtained from Tianliao in Kaohsiung County is approximately 4.0 kgf/cm2 (392 kN/m2). 2.4. Characteristics of other weak rocks 2.4.1. Sandstone In weak rock areas of the southwestern sections of Taiwan, sandstone formation is primarily exhibited in the exposure points of the Hsiangshan facies of the Toukoshan, Liushuang and Nanshihlun Sandstone formations. Sandstone within the Hsiangshan facies of the Toukoshan formation is cemented by silty sediments. Younger sandstone is poorly cemented, prone to slaking in water with an extremely weak weathering resistance. Rock types in the Liushuang Formation mainly consist of three variants; thick-bedded massive sandstone, thick sandstone interbedded with thin layers of mudstone, and interbedded sandstone and mudstone formations. The peak shear strength parameters (cp and p) are 3.8 kPa and 29.7, respectively as derived from the direct shear tests which the specimens are submerged in the shear box for 48 h before the test starts, while the parameters from residual shear strength tests, cr and r, are 0 and 26.3, respectively (Lin et al., 2005).

2.4.2. Shale Shale of southwestern Taiwan is primarily exhibited in exposed Kaitzuliao Shale formations. Shale is a finegrained argillaceous rock with foliations that are usually grayish black or black, forming planes of fissility. Because this foliation constitutes a weak plane with little shear strength that is further reduced due to the effect of water infiltration and slides easily occur along foliation planes on dip shale slopes. For interbedded sandstoneshale formations consisting of sandstone and shale, water infiltrating into these formations accumulates between permeable sandstone and almost impermeable shale bedding surfaces. Thus the sandstoneshale interface tends to become a weak plane of extreme instability. 3. Erosion, vegetation and slope protection 3.1. Erosion characteristics of the mudstone In the study of slope erosion in the Nanhua mudstone area, Chen et al. (1984) and Chang et al. (1996) observed that south-facing slopes are more susceptible to erosion than those facing north and that the percentage of bald slopes facing south was also correspondingly higher. The angles on these denuded slopes range from 36 to 55, and it was observed that the longer the slope lengths, the more serious the erosion rate, with steeper slopes experiencing lower rates of erosion. The erosion depth of the bald mudstone slopes often exceeds 10 cm, while the infiltration depth of ground water may reach approximately 20 to 40 cm vertically below surfaces (Chang et al., 1996). Compared to bald slopes, slopes covered with vegetation generally experienced less erosion depth

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116

105

and land loss. Comparison between different vegetative plants indicates that herbaceous plants (e.g., Paspalum notatum and Cynodon dactylon) or herbaceous plants mixed with woody plants (e.g., Rhus semialata roxburgiana and Psidium guajava) can better protect slopes than woody plants alone (Chiu, 1999). In order to explore the effects of mudstone soil conservation with

the vegetation method in cut slopes, Kuo and Lin (1998) used erosion pins to conduct erosion experiments on mudstone slopes. They concluded that the erosion on mudstone slopes was largely affected by the rainfall, and that the erosion depth of bald mudstone slopes (at a slope angle of 33 with a southwest aspect) averaged about 7.94 cm.

Fig. 3. Slope protection methods used along S-2 Freeway: (a) vegetation, (b) prestressed rock anchors with vegetation, (c) grille beam, (d) rock anchors with grille beam, (e) concrete stairs with shotcrete, (f) precast cement frame, (g) gabion, (h) fabric form (upper).

106

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116

3.2. Vegetation on mudstone slopes Several investigators have previously studied the issue of using vegetation to prevent erosion of mudstone slopes in the south of Meishan region (e.g., Lee et al., 1996; Chen, 1999). A summary of events possibly leading to severe erosion is presented herein. In dry seasons, plant roots are prone to damage from significant shrinkages and cracks that generate exfoliation interfaces on the mudstone surfaces. Since the mudstone is loosely cemented, and is subjected to weathering with water, exfoliation interfaces may cause immediate slaking and loss of mudstone surfaces, which leads to further runoff erosion and exposure of plant roots. Mudstone is structured so closely that water cannot easily infiltrate. Since these mudstone material surfaces contain no geologic aquifers, existing plants usually die from lack of water. Future growth is also limited because plant roots are unable to penetrate these densely compacted surfaces. Large amounts of soluble salts in the soil make the ground nearly inhospitable to growth from all but the hardiest plants, thus increasing the incidents of more surface slides. According to Lin (1997), the annual dry season in southwestern Taiwan which lasts nearly 6 months, is the major obstacle preventing vegetation growth in these mudstone areas. Therefore, the species selected to revegetate these mudstone hillsides should be selected from creeping grasses (e.g., C. dactylon) that are both alkali and drought resistant and convenient in terms of both seed production or propagation. Close monitoring and maintenance of nascent plantings in this mudstone region is critical to the success of a re-vegetation process. As these seedlings take hold and the sites

become more hospitable to plant growth, more species may be introduced into the region. 3.3. Existing mudstone slope protection methods Rainfall, slope angle and slope height are considered the main factors for controlling the failure of mudstone slopes, while the slope aspect, or slant is the factor most affecting the successful growth of vegetation on these hillsides. Using the knowledge of mudstone slope failure characteristics, Lee (1992) recommended that both slope angle and slope height be considered in any mudstone slope protection and re-vegetation initiative. His research called for mudstone slopes (especially slopes with very steep and lengthy grades) to be redesigned into a terraced system. This new landscape design would involve cutting the larger slope into as many smaller terraced slopes as possible, each with a height of 5 m. For slopes with angles greater than 40, sophisticated civil engineering slope protection methods must be used to ensure slope viability. The slope protection methods commonly used by the National Expressway Engineering Bureau and other public agencies and private engineering firms include vegetation, grille beam, fabric form, shotcrete, green coating, stone masonry, gabion, prestressed rock anchor and concrete retaining wall. The following factors should be considered when selecting proper slope protection methods: the objectives and extent of protection to be achieved; the geological structure of the slope; the angle, height and length of the slope; and the constraints and advantages of various slope protection methods in terms of erosion protection, geological stability, ecological considerations and natural material considerations.

Table 4 Lengths of sections protected by various methods along S-2 Freeway including branch line, No. 10 Highway (279 K to 391 K and 0 K to 33 K) Road section Length or Slope protection methods percentage Vegetation Rock anchor + Grille vegetation beam 0 0 1260 7.3% 1090

Rock anchor + Concrete stairs + Precast grille beam shotcrete cement frame 910 6.0% 520 3.1% 0 0 0 0 0 350 4.9% 350 0.9% 320 2.1% 0 0 0 0 320 0.8%

Section Total length total percentage 15,140 100% 17,200 100% 7160 38.3%

Meishan to Sinhua (279 K to 347 K)

Length 13,910 0 (m) Percentage 91.9% 0 Sinhua to Jiouru Length 12,750 2670 (347 K to 391 K) (m) Percentage 74.1% 15.5% Kaohsiung to Cishan Length 5720 0 (0 K to 33 K) (m) Percentage 79.9% 0 Total Total 32,380 2670 length (m) Percentage 82.0% 6.8%

43.6%

18.1%

15.2% 0 2350 1430 5.9% 3.6%

100% 39,500 100% 100%

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116

107

Table 3 provides a comparison of the commonly employed slope protection methods. 4. Field investigation of mudstone slope problems along South-2 (S-2) Freeway 4.1. Slope protection methods employed in S-2 Freeway The field survey shows that the slope protection methods used in road sections of S-2 Freeway to the south of Meishan include vegetation, prestressed rock anchors with vegetation, grille beam, rock anchors with grille beam, concrete stairs with shotcrete, precast cement frame, gabion and fabric form designs (Fig. 3). In 91.9% of the slopes, between Meishan and Sinhua, vegetation has been the primary method of holding the soil in place, followed by 6.0% of the slopes being

secured in place using the rock anchor with grille beam method and 2.1% of the slopes using other methods. Along the road section between the cities of Sinhua and Chiuju, 71.4% of the slopes have been secured using the vegetation method, followed by 15.5% with rock anchor with vegetation method, 7.3% with grille beams and 3.1% with rock anchor with grille beams. Of those slopes adjacent to the Kaohsiung beltway and Cishan branch line, 79.9% are secured with vegetation, 15.2% with grille beams, and 4.9% with concrete stairs and plus shotcrete. Along the freeway in the south of Meishan, on average, 82.0% of the slopes are secured using the vegetation, followed by 6.8%, 5.9% and 3.6% of slopes anchored with prestressed rock, vegetation, grille beam, and rock anchor with grille beam, respectively. The survey results detailing the use of various slope protection methods are summarized in Table 4.

Fig. 4. Slope protection problems along S-2 Freeway: (a) drainage effectiveness of reduced due to a weedy transverse ditch; (b) serious erosion due to the interface of the vertical ditch and the slope surface resulting in a large hole; (c) weedy and silt choked transverse ditch; (d) erosion and hollow space slope damage due to surface runoff overflows below the platform,; (e) sagging of both platform and slope due to water accumulation at the top of slope; and (f) severe slope erosion.

108

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116

Fig. 5. Slope protection problems along S-2 Freeway: (a) serious erosion of the vegetation in mudstone slope, (b) the occurrence of mudflow in the mudstone slope, (c) and (d) severe grille beam slides due to steep slope.

4.2. Slope failure modes Along the S-2 Freeway to the south of Meishan, slope failures occurred mainly in the form of surface erosion, shallow slides and side ditch cracks. Surface runoff caused surface erosion of the slopes; gullies could form first (Fig. 4e) and then developed into more severe erosion (Fig. 4f). Also, vegetation layers could slide easily due to severe erosion, which often occurred in the mudstone and the alternating sandstoneshale formation slopes (Figs. 4d and 5a). Shallow slides often occurred during the season of typhoons and heavy rains (Fig. 3a), especially for mudstone slopes. The depth of

slides was generally shallow (about 50 cm), and the size of the slides ranged from 5 m 4 m to 60 m 10 m. Based on field surveys along the road section between Meishan and Sinhua, there are 14 sites where surface erosion is prevalent and two sites where shallow slides have occurred. Thus, the percentages of the failure modes of surface erosion and shallow slide in this section of freeway are 87.5% and 12.5%, respectively. Along the road section between Sinhua and Chiuju, approximately 38 sites have surface erosion, 23 sites are shallow slide areas and 4 sites have side ditch cracks, accounting for 58.5%, 35.4% and 6.1% of total slope failures, respectively. Along the Kaohsiung beltway and

Table 5 Quantity of failures along S-2 Freeway including branch line, No. 10 Highway (279 K to 391 K and 0 K to 33 K) comparison of failure modes Road section Percentage Failure modes of total Surface length erosion Number Percentage Number Percentage Number Percentage Number Percentage 14 87.5% 38 58.5% 30 83.3% 82 70.1% Destruction ratio

Shallow slide 2 12.5% 23 35.4% 6 16.7% 31 26.5%

Side ditch crack 0 0 4 6.1% 0 0 4 3.4%

Total

Percentage of total number of failures

Meishan to Sinhua (279 K to 38.3% 347 K) Sinhua to Jiouru (347 K to 43.6% 391 K) Kaohsiung to Cishan (0 K to 18.1% 33 K) Total 100%

16 13.7% 100% 65 55.5% 100% 36 30.8% 100% 117 100% 100%

0.36 1.27 1.70

Note: Destruction ratio is the ratio of the percentage of total number of failures over the percentage of total road length.

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116

109

Table 6 Quantity of failures along S-2 Freeway including branch line, No. 10 Highway (279 K to 391 K and 0 K to 33 K) comparison of slope protection methods Slope protection method Total length (m) Total Failure modes Destruction length ratio Surface erosion Shallow slide Side ditch crack Total Total failure percentage (%) percentage (%) 72 10 0 0 0 0 82 28 3 0 0 0 0 31 1 2 0 1 0 0 4 101 15 0 1 0 0 117 86.3 12.8 0 0.9 0 0 100 1.05 1.88 0 0.25 0 0

Vegetation 32,380 82.0 Rock anchor 2670 6.8 + vegetation Grille beam 2350 5.9 Rock anchor + grille 1430 3.6 beam Concrete stairs 350 0.9 + Shotcrete Precast cement frame 320 0.8 Total 39,500 100

Note: Destruction ratio is the ratio of the percentage of total number of failures over the percentage of total road length.

Cishan branch line, 83.3% (30 sites) of the failures showed evidence of surface erosion and 16.7% (6 sites) were considered shallow slide areas. Of the road sections investigated along S-2 Freeway to the south of Meishan, 70.1% (82/117) of the slope failures were due to surface erosion, 26.5% (31/117) of the slope failures were due to shallow slides, and 3.4% (4/117) of the slope failures were due to side ditch cracks. Surface erosion was determined to be the most common slope failure mode along the S-2 Freeway. In terms of the ratio of destruction, defined as the ratio of failure percentage to length percentage, the slopes along the Kaohsiung beltway and the Cishan branch line have the highest ratio of destruction of the three road sections investigated, while the section between Meishan and Sinhua has the lowest ratio of destruction. The modes and numbers of slope failures are summarized in Table 5. 4.2.1. Distribution of slope failure modes by protection methods Field investigations along the S-2 Freeway to the south of Meishan showed that the largest number of

slope failures occurred on vegetation-protected slopes at 101 sites, which accounted for approximately 86.3% of all recorded slope failures. Of these, surface erosion was prevalent at 72 sites, shallow slides were prevalent at 28 sites and 1 site was the result of a side ditch crack. The rock anchor with vegetation securing method had the second highest number of failures, accounting for 12.8% (15 sites) of all failures recorded. Of these 15 failures, 10 were due to surface erosion, two failures were due to shallow slides and two failures were due to side ditch cracks. A summary of the effectiveness of the slope protection methods based on the data collected from field investigation of the three road sections along the S-2 Freeway is presented in Table 6. Comparing all methods in terms of the destruction ratio, the ratio of destruction was highest for those slopes protected using rock anchor plus vegetation and using just vegetation (1.88 and 1.05, respectively). The ratio of destruction was lowest for those slopes secured with the grille beam method and that of concrete stairs combined shotcrete (both ratios are essentially zero). These results suggest the inevitability of slope failures

Table 7 Quantity of failures along S-2 Freeway including branch line, No. 10 Highway (279 K to 391 K and 0 K to 33 K) comparison of rock types Rock types Total length (m) Total Failure modes length Surface Shallow percentage erosion slide (%) 18 34 20 10 82 8 10 12 1 31 Destruction ratio

Side ditch crack 2 0 2 0 4

Total Total failure percentage (%) 28 44 34 11 117 23.9 37.6 29.1 9.4 100

Sandstone 9630 24.4 Alternating 14,330 36.3 sandstoneshale formation Mudstone 6890 17.4 Conglomerate 8650 21.9 Total 39,500 100

0.98 1.04 1.67 0.43

Note: Destruction ratio is the ratio of the percentage of total number of failures over the percentage of total road length.

110

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116 Table 8 List of possible effects of water on slope surface Cause of failure Adverse effect Countermeasures Essential feature of effective protection method Prevent rain from directly impacting the mudstone slope Uses filter on the mudstone slope; placement of base layer for vegetation on the filter. Uses filter on the mudstone slope; placement base layer for vegetation on the filter and sets transverse ditch. Increases crosssectional area of the transverse ditch; extend the filter to the platform and slope below the platform. Extends filter to the platform and slope below the platform; Shorten the interval of vertical ditch.

when vegetation is exclusively used as protection for soft-rock slopes along the S-2 Freeway. The survey results also show that of the slopes along the S-2 Freeway exhibiting the most severe erosion, the grille beam slope protection method was the most commonly employed remedial measure. This method involves the use of grille beams filled with soil bags or vegetative bags for plants as the vegetative base layer. Good stability can generally be achieved when slopes initially treated with vegetation are reinforced with grille beam protection. 4.2.2. Distribution of slope failure modes by rock types The rock types along the S-2 Freeway to the south of Meishan are very complex. The S-2 Freeway was built through formations of sandstone, sandstone interbedded with mudstone, alternate layers of sandstone and mudstone, sandstone interbedded with shale or alternate layers of sandstone and shale, mudstone, mudstone interbedded with sandstone, and conglomerate soils interbedded with sandstone. To facilitate the analysis, the rocks adjacent to S-2 are summarized into four major types; sandstone, the alternating sandstoneshale formation (also referred to as alternation of sandstone and shale), mudstone, and conglomerate. Among all rock types along S-2, the alternating sandstoneshale formation predominates, accounting for 36.6%; sandstone and conglomerate accounts for 24.4% and 21.9%, respectively; while mudstone is the least prevalent, accounting for 17.4%. The proportion of both rock types and slope failure modes and quantity are summarized in Table 7. Among the four rock types, slopes composed of alternating sandstoneshale formation have the largest number of failures, accounting for 37.6% of the total. This is sequentially followed by slopes comprising mudstone, sandstone, and conglomerate, accounting for 29.1%, 23.9%, and 9.4% of total number of respective slope failures. However, in terms of destruction ratios, mudstone slopes have the highest ratio (1.67), followed by the slopes comprising alternating sandstoneshale formation (1.04), slopes comprising in sandstone (0.98), and slopes composed of conglomerate (0.43). All previously presented data indicate that the primary problem of slope failures along S-2 Freeway lies in slope surface erosion and shallow slides in mudstone slopes, and that the secondary problem involves slope surface erosion in the alternating sandstoneshale formation. 5. Attributes of slope failures and possible remedial measures Results of the field investigation indicate that slope failures in weak rock areas in southwestern Taiwan is

Impact by rainfall

Accelerate the erosion of mudstone slope.

Slope runoff

Surface runoff causes the slope erosion; accelerate the weathering of mudstone.

Reduce surface runoff to protect the slope surface.

Transverse ditch facilities improperly constructed or are ineffective due to weedy vegetation Interval of vertical (longitudinal) ditch is too wide (large)

Transverse ditch unable to quickly and effectively drain; surface runoff overflows to the slope below platform, causing erosion.

Remove soil weeds from transverse ditch to prevent slope erosion below platform.

Transverse ditch unable to quickly and effectively drain; surface runoff overflows to the slope below platform, causing erosion. Leakage Serious erosion from the of the interface vertical ditch of the vertical ditch and slope; penetration to the slope below the platform.

Proper design of vertical ditch interval to prevent slope erosion below platform.

Raise slope surface slightly higher than vertical ditch; avoids loss of soil at the interface.

Extends filter to the bottom of the vertical ditch.

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116 Table 9 Problems to vegetation along mudstone slopes Cause of failure Adverse effect Countermeasures Essential feature of an effective protection method Reduce evaporation; prevent slide of vegetation and weathering layer. Use filters on mudstone slope; placement of base layer for vegetation on filter.

111

Slope cracking due to evaporation of mudstone

Slope absorbing water and slaking

Forms a peeled interface, likely occurrence of both vegetation slide and weathering layer slide in rainy season. Accelerates erosion of the mudstone slope; weathering penetrates deeper into mudstone. Pressure from swelling may destroy slope protection facilities.

Prevent mudstone particle loss and weathering.

Slope absorbing water and swelling

Enhance slope protection facility to sustain swelling pressure.

Use filter on mudstone slope; placement of base layer for vegetation on filter. Slope protection facility able to tolerate some deformation.

failure. In fact, under unfavorable geological, topographical and climatic conditions, which are prevalent in the southwestern region, vegetation protection methods are not effective at all. For example, on mudstone slopes, erosion-related problems are severe even with substantial vegetation cover. Mudstone is prone to weathering and runoff erosion can facilitate the infiltration of water into the interface between weathered mudstone and fresh mudstone, which destabilize the entire mudstone slope. Once the mudstone slope becomes unstable, the vegetation and weathered layers migrate over the interface to induce further vegetation layer slides. The potential problems of mudstone slopes with vegetation protection are summarized in Table 9. Examples of mudstone slope failures, including those augmented with vegetation protection are shown in Fig. 5(a) and (b). From these results, it becomes unequivocally apparent that the use of vegetation without proper engineering measures is totally ineffective in protecting mudstone slopes. 5.3. Slope angles and slope height Excessive angles and lengths of slopes increase the possibility of surface erosion and shallow slides. Among those road sections surveyed, the roadside slope grade of 1:1.2 (approximately 40) was designed for the section between Sinhua and Chiuju, the Kaohsiung beltway and the Cishan branch line. The survey results show that vegetation slope angles generally ranged from 39 to 41, which is greater than 35 that is considered the upper bound of the angles that allow for natural invasion and propagation of plants. Moreover, the slope height was generally in the range of 7 to 9 m. The designed
Table 10 Problems related to slope angle and height on mudstone slopes Cause of failure Vegetated slope is too steep. Adverse effect Countermeasures Essential feature of an effective protection method Reduce slope angle or rebuild vegetation layer on the slope. Place a base layer for vegetation in the grid-frame.

due to water erosion and drainage, the use of only vegetation on slopes to protect and secure them in place, and slope angle and slope height. These attributes are further discussed and a new slope protection method is then presented. 5.1. Water and drainage Water is the most important and destructive element causing both slope instability and slope erosion along the S-2 Freeway. Table 8 lists all the possible effects of water on slopes and the recommended countermeasures. Because the study area has distinct dry and rainy seasons, including heavy rains brought about by typhoons, slope protection measures are destined to fail, if the surface and subsurface water cannot be effectively drained away. Fig. 4 shows examples of slope failures due to improper or inadequate drainage. 5.2. Vegetation on weak rock slopes Based upon the presented data, slopes along the S-2 Freeway protected by only vegetation are susceptible to

Grille beam slide due to steep slope angle.

Vegetation is difficult to maintain on a steep slope; vegetation layers can slide easily. Sliding of the grille beam and vegetation layer is likely to affect traffic; extremely difficult to repair grille beam.

Reduce weight Use lightweight of grid-frame. grid-frame to increase stability and to facilitate rapid repair, if necessary.

112

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116

Fig. 6. Sketch of a new slope protection method (Soil-TireVegetation Method) that has been tested during this study.

slope grade along the section between Meishan and Sinhua measured mostly within the range of 1:1.5 (approximately 34), with vegetative slope angles between 33 to 36. The survey also indicated a substantial ratio of slope destruction along the road section between the Kaohsiung beltway and the Cishan branch line. Indeed this destruction ratio was 1.70,

higher than any other road section studied. The destruction ratio of slopes next in sequence to the Kaohsiung/Cishan section was between Sinhua and Chiuju with 1.27, while the section with the least amount of measured destruction was between Meishan and Sinhua, with 0.36. This data infers that slopes with greater angles are more likely to fail, as summarized in

Fig. 7. Plan view of the new slope protection method (Soil-TireVegetation Method) that has been tested during this study.

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116

113

Table 10. Fig. 5(c) and (d) provide photographic evidence for this data. 5.4. Requirements for effective slope protection methods Based on the characteristics of the mudstone and the results of the field survey, the following recommendations are essential for effective slope protection countermeasures:

1) A slope protection method must be able to resist erosion, prevent fine-grained soil loss, and facilitate adequate drainage, 2) A surface water drainage system is required to effectively control the rainwater, 3) The slope protection facilities must have adequate strength to resist the swelling pressure of the mudstone, and 4) The slope protection facilities must increase slope stability and facilitate vegetation growth.

Fig. 8. The field implementation of the Soil-TireVegetation Method: (a) slope cut to the design shape; (b) forming multi-stage slope and platform; (c) laying the fabric drains; (d) laying the nonwoven geotextile sheet; (e) constructing transverse ditch; (f) laying the separator of the Soil-Tires; (g) laying the Soil-Tire and vegetative bags; (h) placing the vegetative mat.

114

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116

Fig. 9. The arrangement of test slopes.

A new slope protection method, Soil-TireVegetation Method, implementing the requirements previously described, was developed for this weak-rock mudstone region (see Figs. 6 and 7 for implementation sketches). To determine its effectiveness, a field experiment has been conducted at a mudstone slope near the proposed Tungshan Rest Station of National Highway No. 3. Fig. 8 shows the steps taken for field implementation of the Soil-TireVegetation Method. The construction of the test slopes began on 17 September, 2004 and ended on 27 November, 2004. The original slope in the experimental zone consists of the upper and lower slope surfaces. As shown in Fig. 9, the test slopes are divided into original slopes (A1), contrast slopes (B1, B2 and G1, G2) and Soil-Tire slopes (C1, C2 and D1, D2; E1, E2 and F1, F2). The attitude, area and vegetative treatment of slopes of various groups are shown in Table 11. The strike of slope surfaces in the experimental zone is N40W with a SW dip direction. The dip angles of the original slopes are about 42, while those of contrast slopes and Soil-Tire slopes are approximately 36 and 34, respectively. The width of the slopes of various groups is about 15 m, and the slope lengths gradually lessen from right to left. With regard to vegetative treatment for slopes of various groups, the original slopes are covered with originally vegetative species, while the contrast slopes are laid with vegetative mats after slope cut to the design specification, and the SoilTire slopes are placed with Soil-Tires as the base layer for vegetation. As for the vegetation on Soil-Tire slopes, four Vativeria zizanioides are planted in the central space of Soil-Tire on the slopes of C1 and C2, and the remaining slopes (D1, D2, E1, E2, F1, and F2) are covered with vegetative mats which consist of nonwoven and straw vegetative mats. In addition, the seeds of C. dactylon and P. notatum are sowed on these slopes. The total project cost was about 35,000 U.S. dollars.

Table 12 shows the erosion of test slopes in the first year of experimentation. During this period, there had been several typhoons and heavy rainstorms and the accumulated rainfall reached 3553 mm. As shown in Table 12, the original slopes suffered the most serious erosion (with erosion of 27665 g/m2 on Slope A1). The contrast slopes showed significant improvement as far as the erosion is concerned. Finally, the slopes protected with the Soil-TireVegetation Method drastically reduced the erosion and the mud slide potential. This
Table 11 Types of vegetation and slope stabilization methods at test slopes No. Original slope Contrast slope A1 B1 B2 Soil-Tire slopes C1 C2 D1 D2 E1 E2 F1 F2 Contrast Slope G1 G2 Attitude N40W/ 42SW N40W/ 36SW N40W/ 36SW N40W/ 33SW N40W/ 33SW N40W/ 33SW N40W/ 33SW N40W/ 34SW N40W/ 34SW N40W/ 34SW N40W/ 34SW N40W/ 36SW N41W/ 35SW Area (m2) 69.38 112.01 82.50 31.84 18.43 35.30 20.70 40.62 22.20 38.15 21.88 158.78 121.80 Vegetation treatment Original slope vegetation Vegetative mat Vegetative mat Soil-tire + Vativeria zizanioides fertilized Soil-Tire + V. zizanioides Soil-tire + vegetative mat Fertilized Soil-Tire + vegetative mat Soil-tire + vegetative mat Fertilized Soil-Tire + vegetative mat Soil-tire + vegetative mat Fertilized Soil-Tire + vegetative mat Vegetative mat Vegetative mat

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116 Table 12 Soil erosion as measured at test sites

115

Note: original slope, contrast slope (vegetative mat); all others: slopes protected with the Soil-TireVegetation Method.

preliminary result shows the promise of the Soil-Tire Vegetation Method in the protection of the mudstone slopes in the southwestern Taiwan region. 6. Concluding remarks Along the S-2 Freeway in southwestern Taiwan, vegetation protection was the primary slope protection in 70% to 90% of all cases. Other methods used included the prestressed rock anchor plus vegetation, grille beams, rock anchor with grille beam, concrete stairs with shotcrete, precast cement frame, gabion and fabric form methods. Field surveys indicated that the vegetation protection method was ineffective, particularly on mudstone slopes. The most commonly observed mode of failure was surface erosion, accounting for approximately 70% of all slope failures, followed by shallow slides accounting for approximately 27%. The largest number of failures (86%) occurred on vegetation-protected slopes, followed by slope failures where rock anchor plus vegetation was the preferred method (13%). In terms of the destruction ratio, slopes secured with rock anchor and vegetation had the highest ratio (1.88) and the vegetation-protected slopes had the second highest ratio (1.05). Among the rock types encountered in the S-2 Freeway in southwestern Taiwan, the most commonly seen was the alternating sandstoneshale formation, which accounted for 36.3%, followed in sequence by sandstone (24.4%), conglomerate (21.9%), and mud-

stone (17.4%). In terms of the number of slope failures, slopes composed of alternating sandstoneshale formation had the highest number of failures, accounting for 37.6% of all failures. The mudstone slopes had the second highest number of failures (29.1%), followed by the sandstone slopes (23.9%), and the conglomerate slopes (9.4%). In terms of slope destruction ratios, the mudstone slopes had the highest ratio (1.67), followed in sequence by slopes composed of the alternating sandstoneshale formation (1.04), slopes composed of sandstone (0.98), and slopes composed of conglomerate (0.43). For mudstone slopes in Southwestern Taiwan, specifically those with high angles and length, the vegetation protection method alone was ineffective. Therefore, additional engineering protection methods must be employed along with this method. Considering the slope failure characteristics of the study area, the following attributes were deemed essential in designing effective slope protection methods: the ability to resist erosion, fine-grained soil loss prevention, proper surface water drainage systems to control stormwater runoff; the ability to resist the swelling pressure of the mudstone; and proper features to increase the stability of the slope and facilitate vegetation growth. A new slope protection method, called the Soil-TireVegetation Method, implementing these requirements, was developed and the preliminary field test results demonstrated its effectiveness in protecting the mudstone slopes in the southwestern Taiwan region. This field experimentation

116

D.-H. Lee et al. / Engineering Geology 91 (2007) 100116 Goodman, R.E., 1980. Introduction to Rock Mechanics. John Wiley and Sons, N.Y., p. 37. Goodman, R.E., 1993. Engineering Geology Rock in Engineering Construction. Wiley, New York, N.Y. Franklin, J.A., Vogler, U.W., Szlavin, J., Edmond, J.M., Bieniawski, Z.T., 1979. Suggest methods for determining water content, porosity, density, absorption and related properties and swelling and slake-durability index properties, for ISRM Commission on Standardization of Laboratory and Field Tests. Int. J. Rock Mech. Min. Sci. 16 (2), 141156. Kuo, C.C., Lin, S.H., 1998. Soil conservation effectiveness comparison of different vegetation engineering methods for cut slopes in the mudstone area. J. Soil Water Conserv. 30 (3), 187205. Lee, D.H., 1992. An investigation of protection work on highway slopes in mudstone area. Disaster Prevention Investigation, Report No. 8072. National Science Council, Taipei, Taiwan. Lee, D.H., Jhin, Y.Y., Tien, K.G., 1994. Characteristics of mudstone and the methods for slope protection. Sino-Geotechnics 48, 3547. Lee, D.H., Tien, K.G., Juang, C.H., 1996. Full-scale field experimentation of a new technique for protecting mudstone slope, Taiwan. Eng. Geol. 42 (1), 5163. Lin, C.Y., 1997. Re-vegetation of bare mudstone slope land for a watershed in southern Taiwan. J. Soil Water Conserv. 29 (2), 169181. Lin, H.M., Liao, C.J., Wu, C.H., Lee, D.H., 2005. Investigation the residual strength of soft rock in southern Taiwan using ring shear test. Proceedings of the Symposium of the Technical Treatments on Slope Stabilization for the Soft Rock in Southern Taiwan, Tainan, pp. 157175. Oliveira, R., 1993. Weak rock materials. In: Cripps, et al. (Ed.), The engineering geology of weak rock. Balkma, Rotterdam, pp. 515. Proter, A.A., Nelson, J.D., 1980. Strain controlled testing of expansive soils. Proceedings, 4th Int. Conf. on Expansive Soils, Denver. ASCE, New York, pp. 3444. Tsai, C.L., 1984. On the Mineral Components of the Mudstone in Southwestern Taiwan. Report No. NSC-73-0414-P006-011. National Science Council, Taipei, Taiwan. Wang, T.T., Huang, T.H., 2002. An experience of tunneling in the mudstone area in southwestern Taiwan. Tunn. Undergr. Space Technol. 17, 425436.

project showcases a successful application of engineering geology. Acknowledgments The study on which this paper is based was supported by National Science Council through Grant No. NSC912211-E-006-047. This financial support is greatly appreciated and acknowledged. References
Bjerrum, L., 1967. Progressive failure in slopes of overconsolidation plastic clay shales. J. Soil Mech. Found. Div., ASCE 93, 149. CEC, 1999. Slope Stabilization for the Follow-up Plan of the No. 3 Freeway in Keelung Sijhih Section, Tainan Section, Kaohsiung Section and Pintung Section. Technical Report, China Engineering Consultants (CEC), Inc., Taipei, Taiwan. Chang, H.W., Tien, T.M., Juang, C.H., 1996. Formation of southfacing bald mudstone slopes in southwestern Taiwan. Eng. Geol. 42 (1), 3749. Chen, C.S., 1999. Study on vegetation treatment of mudstone slope. Proceedings, the Symposium of the Rebuilding and Vegetation Greening on Mudstone Area, Tainan, pp. 7889. Chen, H.M., 2002. A study on failure mechanisms of weathering sedimentary rock in southern of Taiwan and mitigation. M.S. Thesis, National Taiwan University of Science and Technology, Taipei, Taiwan. Chen, C.Z., Lee, Y.F., Liu, I.T., 1984. A study on the erosion characteristics of the mudstone slopes in southwestern Taiwan. Disaster Prevention Investigation, Report No. 73-07. National Science Council, Taipei, Taiwan. Central Geological Survey, MOEA, 2000. Geological map of Taiwan scale 1:50,000. Chiu, C.Y., 1999. The method of re-vegetation of mudstone. Proceedings, the Symposium of the Rebuilding and Vegetation Greening on Mudstone Area, Tainan, pp. 6470. Cripps, J.C., Taylor, R.K., 1981. The engineering properties of mudrocks. Q. J. Eng. Geol., London 14, 325346. Dick, J.C., Shakoor, A., 1992. Lithological controls of mudrock durability. Q. J. Eng. Geol. 25, 3146.