Shallow and deep-seated regolith slides on deforested slopes in Çanakkale, NW Turkey

(1)

Yunus Levent Ekinci

a,

⁎

, Murat Türke

_ş

b

, Alper Demirci

a

, Ahmet Evren Erginal

c

a

Çanakkale Onsekiz Mart University, Engineering Faculty, Department of Geophysical Engineering, TR-17020 Çanakkale, Turkey

b_{Middle East Technical University, Faculty of Arts and Sciences, Afﬁliated Faculty at the Department of Statistics, TR-06800 Ankara, Turkey} c

Ardahan University, Faculty of Humanities and Letters, Department of Geography, TR-75000 Ardahan, Turkey

a b s t r a c t

a r t i c l e i n f o

Article history: Received 8 July 2011

Received in revised form 28 January 2013 Accepted 10 June 2013

Available online 22 June 2013 Keywords:

Regolith

Soil moisture content Shallow and deep-seated slides Electrical resistivity

Çanakkale Turkey

This study deals with the stripping of regolith on a steep slope by surface wash and shallow landslides and a deep-seated landslide at a lower slope that took place on 17 February 2003 at the village of Mazılık, east of Çanakkale, Turkey. Soil loss and shallow slides dominate on the deforested steep slopes in the study area and occur preferentially along slope-parallel sub-horizontal joint planes with clay coatings, particularly oxyhydroxides that are rich in Fe but poor in Mn as a result of weathering under well-drained conditions. Gully erosion also occurs where the regolith cover is relatively thick (up to ~ 4 m). The area of the deep-seated landslide, however, is dominated by silty clay (46%). A geoelectrical resistivity survey revealed a clay-rich zone at depths of ~ 3–10 m, corresponding to the slip surface of the slide, which was associated with excessive water content after the snowy day of 14 February 2003 with a daily precipitation of ~ 16.4 mm. Based on Thornthwaite's water budget analysis, the study area has a slide-prone condition with excess soil-moisture content, heavy rainfall events, snow accumulation and snow melting in winter months, and low soil permeability also favouring slope instability.

1. Introduction

Regolith, a termﬁrst suggested byMerrill (1897), is a weathering mantle formed by decomposition of the underlying host rock that can be of various origins and nature. As discussed in detail byEggleton and Taylor (2001), regolith is of great importance for agricultural activities as well as engineering projects. Nevertheless, regolith, depending on its thickness and other geo-environmental factors, may sometimes be subjected to intensive gully erosion and slope failures of various scales. Although rainfall is a common triggering factor of regolith slides (De Vita and Reichenbach, 1998; Corominas, 2001; Conte and Troncone, 2012; Montrasio et al., 2012; Salciarini et al., 2012), various hydrogeological factors as well as chemical and mineralogical heterogeneities are also important (Duzgoren-Aydin and Aydin, 2006; Che et al., 2012). Shallow landslides often affect long-term hillslope evolution and have a basal surface between regolith and underlying bedrock, where bedding, joints, and faults provide planes of pre-existing weakness with high clay concentration (Skempton and Petley, 1967; Cruden and Varnes, 1996).

We studied shallow and deep-seated regolith slides in the Biga Peninsula, which is underlain by various volcanic rocks including basalt, dacite, and agglomerate of Eocene–Miocene age. The study site, an abandoned small village called Mazılık, is located 20 km east

of Çanakkale City. The slope failures and gully erosion in the area date back to the 1950s. Shallow slides occurred on a steep deforested slope (N20°). In mid-February of 2003, a deep landslide also occurred below the shallow landslides, destroying a primary school.

The present study describes the petrography and structural features of weathered volcanic rocks on a deforested steep slope and discusses their relationship with gully erosion and shallow slides. The subsurface characteristics of the deep-seated landslide are also examined using data from electrical resistivity tomography (ERT) and vertical electrical sounding (VES) techniques. The failures are discussed in relation to sub-horizontal joint planes with clay coatings. Hydroclimatological analyses explaining the effects of precipitation on the removal of regolith are also described. Additionally, changes in the landslide area between 1954 and 2008 are documented with aerial photographs and satellite images.

2. The study area

The study area in the village of Mazılık (40°07′31″ N, 26°34′14″ E) (Fig. 1) is now abandoned because of limited arable land and the threat of slope failures and gully erosion. Precipitation in the study area is low from May to September, a typical characteristic of the Mediterranean climate (Türkeş, 1996, 1998). According to long-term meteorological data from the Çanakkale weather station (40°08′28″ N, 26°24′01″ E; 6 m a.s.l.; 20 km west of the landslide area) operated by the Turkish State Meteorological Service (TSMS),

⁎ Corresponding author. Tel.: +90 2862180018 2166; fax: +90 2862180541. E-mail address:ylekinci@comu.edu.tr(Y.L. Ekinci).

(2)

maximum precipitation occurs in December with a long-term average precipitation of ~ 109 mm, whereas minimum precipitation occurs in August with a long-term average of ~ 6 mm. The coldest month, with a long-term average temperature of 6 °C, is January, and the warmest, with an average of 24.8 °C, is July. The aridity index of the United Na-tions Convention to Combat Desertiﬁcation (UNCCD, 1995) indicates that a semiarid or a sub-humid climate prevails over the Çanakkale District (Türkeş, 1999).

Vegetation is quite sparse and is characterised mainly by some maquie elements (e.g., Quercus coccifera, Phillyrea latifolia, and Styrax ofﬁcinalis), pine species (Pinus brutia and Pinus nigra) and Quercus species (e.g., Quercus cerris, Quercus infectoria, and Quercus pubescens).

Field studies also showed the presence of S. ofﬁcinalis, Pyrus

amygdaliformis, Prunus amygdalus, and P. latifolia. The nearlyﬂat ridge surface behind the steep slope is dominated by P. brutia, Q. cerris L. var. cerris, and Q. pubescens. The sparse vegetation around the landslide area, with an altitude of between 350 and 400 m, is mainly a result of deforestation and overgrazing.

The slope with the shallow slides and gully erosion faces southwest with elevation ranges from 350 to 440 m, and angles up to ~ 27°. The slides occur in an area with a length of 240 m and width of 160 m (Fig. 2A), in which minor slide scars with heights between 10 and 50 cm occur along sub-horizontal joints (Fig. 2B). The regolith mantle on the slope is dissected by nine gullies with lengths between 205 and 514 m (Fig. 2C). These gullies areb3 m in width and their bottoms are congested with basalt blocks with

diameters of b1 m. The deep-seated failure extends down to

300 m a.s.l. (Fig. 2D). Besides the concavity of the slip surface and hummocky topography of the landslide toe with accompanying ridges, small ponds and distorted trees and fences typify rotational slides (Cruden and Varnes, 1996).

According to oral and recorded information from elderly villagers and the local administration and to technical reports from the Çanakkale Public Improvements Administration, the slope failures and land degradation of the study area began in the early 1950s. An aerial photograph acquired in 1954 shows that there had been a

Fig. 1. Location map of the study area.

71 Y.L. Ekinci et al. / Geomorphology 201 (2013) 70–79

(3)

village with ~ 40 houses (Fig. 2E, G). The slope behind the settlement area was covered with extensive vegetation; however, a Quickbird satellite image acquired in 2008 indicates that this vegetation had been destroyed as a result of human interference (Fig. 2F, H). According to the villagers, the shallow slides and gully erosion compelled the inhabitants to evacuate to a region 450 m northeast of the study area. The village's primary school was in use until the last destructive event of 2003.

3. Methods

3.1. Climatological and meteorological analyses

We analysed climatic data recorded at the Çanakkale station for the period of 1930 to 2008, daily weather bulletins, and daily surface

and upper air synoptic weather maps of theTSMS (2009a,b). The

precipitation and temperature data were originally compiled by

Fig. 2. Digital elevation model (DEM) and pictures from the landslide area. (A) DEM produced from contour lines with a 10 m interval. (B) Shallow slides on a deforested steep slope. (C) Gullyﬁlled with basalt blocks. (D) Shallow and deep-seated slides and the primary school later damaged. RS: Reference site. SS: Steep slope. R1: Regolith in the shallow slide area in the upper slope. R2: Regolith in the deep-seated slide in the lower slope. The black and white dashed lines indicate boundaries of shallow and deep-seated slide zones, respectively. (E) Aerial photograph taken in 1954. (F) Quickbird satellite image taken in 2008. (G and H) Closer views of (E) and (F), respectively. Dashed white lines show the study area.

(4)

Türkeş (1996, 1998)andTürkeş et al. (2002). We updated the data up to 2005 for the present study. The data consist of monthly precipitation (mm) and mean monthly temperature (°C).

Thornthwaite's (1948)climate classification and water budget for Çanakkale were calculated using the WATBUG programme developed byWillmott (1977). This programme outputs unadjusted potential evapotranspiration (UPE); adjusted potential evapotranspiration (APE); soil moisture storage (ST); actual evapotranspiration (AE); soil moisture deficit (D); and soil moisture surplus (S) (all in mm). The WATBUG calculates monthly or daily APE values (Türkeş and Akgündüz, 2011), which are adjusted for the length of daytime and the number of days in the month. It further provides precipitation minus potential evapotrans-piration and change in soil water or moisture storage from the preceding month (CST). AE is equal to APE when water storage is atfield capacity or when P− APE has a positive value; it is otherwise equal to the sum of P and CST. D is equal to APE− AE, and S is equal to the positive P − APE

when water storage is at capacity or higher (Türkeş, 2010).

Thornthwaite's (1948)moisture index (Im) is calculated as

Im¼ 100S−60Dð Þ=APE; ð1Þ

where S, D, and APE are the summations of the monthly amounts. Negative values of Im are obtained for dry climates, whereas positive values are obtained for moist climates.

Monthly precipitation totals measured at Çanakkale were normalised to allow objective comparisons between monthly totals with different long-term averages and variances. The normalised monthly precipitation anomaly (PAm) for a monthly total precipitation series is calculated as

PAm¼ Pm−P

=σ; ð2Þ

where Pmis the total precipitation amount (mm) for the m-th month (m); and P andσ are the long-term average and standard deviation of the total precipitation series for that month, respectively, for the period 1930–2008.

3.2. Sampling and analyses of slip surface material and regolith The nature of samples collected from clay-coated joint planes, of rock material, and of regolith was investigated based on energy dispersive X-ray spectrometry (EDX) with a Bruker AXS XFlash (Madison, Wisconsin), X-ray diffractometry (XRD) with an X'Pert Pro (Philips Analytical, Cambridge), and thin section interpretations. The elemental composition and clay types of the slip surface materials were examined from fourﬁne clay-rich samples extracted from the surface of joint planes. The analyses were carried out at the Center for Materials Research of the Izmir Institute of Technology, Turkey.

Seventeen samples were collected and their sediment size fractions were determined by the hydrometry technique (Bouyoucos,

1963). pH values were measured using a WTW multi-parameter

instrument based on the method of Grewelling and Peech (1960).

Four samples (massive basalt, partly weathered basalt, strongly weathered basalt, and regolith) were also subjected to thin-section examination with an Olympus BX51 petrographic microscope. 3.3. Geoelectrical resistivity survey

A two-dimensional (2D) ERT survey, frequently performed for shallow geophysical investigations in environmental and engineering applications, was carried out along the axis of the deep-seated landslide at the lower slope section of the study area in order to investigate the subsurface of the landslide. The technological development of automatic multi-dimensional data acquisition systems and multi-dimensional forward modelling and inversion

algorithms permits efﬁcient and reliable surveys and produces

high-resolution images of the subsurface resistivity variation. Therefore, ERT is widely used to depict the subsurface geometry of a landslide body (e.g.Suzuki and Higashi, 2001; Lapenna et al., 2003; Drahor et al., 2006; Göktürkler et al., 2008; Erginal et al., 2009).

The ERT survey was carried out using the Wenner–Schlumberger electrode array conﬁguration by means of the IRIS-Syscal R1 plus resistivity-meter system. The line length was 64 m, and electrodes were spaced every 2 m (Fig. 2D). Because the depth of penetration depends on the distance between current electrodes, 13 different distances were used to obtain information about both the shallower and deeper parts of the subsurface. The scheme of the tomographic survey, along with the data density, is shown inFig. 3. The number of pulses (repeated measurements) for each data point was initially set to four for the calculation of measurement error (standard deviation), and, when the error was greater than 2%, the number of pulses was increased to eight to enhance data quality, although this circumstance occurred at only a few points. The results were used to calculate the average apparent resistivity.

Fig. 3. Schematic representation of the Wenner–Schlumberger array with electrode locations and datum points.

Table 1

Climatic characteristics of the Çanakkale weather station according to the Thornthwaite's climate classiﬁcation.

Station Moisture index (Im) Thermal efﬁciency Humidity index (Ih) Summer concentration (%) Climate type with symbols Çanakkale −1.7 79.6 29.6 52.7 C1B′2s b′3 73 Y.L. Ekinci et al. / Geomorphology 201 (2013) 70–79

(5)

In order to transform the apparent resistivity data into the true resistivity distribution in the subsurface, the algorithm proposed by

Loke and Barker (1996)for an automatic 2D inversion scheme was implemented using RES2DINV software. The algorithm is based on smoothness-constrained least-squares inversion (Sasaki, 1992)

implemented by a quasi-Newton optimisation technique (Loke and

Barker, 1996). Because of the significant topographic relief along the survey line, elevation data were also considered. Elevation was recorded at the location of each electrode using optical levelling. The calculated apparent resistivity values were obtained by the finite-element method using four nodes per unit electrode spacing during the inversion process. The inversion process converged to a satisfactory solution after five iterations with a root-mean-square

(RMS) error of 4%. Thus, the ﬁfth iteration was considered to

have produced a geoelectrical image that approximated a realistic representation of the true resistivity distribution of the subsurface.

In addition to the ERT survey, a VES survey was performed in order to explore the depth of the bedrock. The technique determines the one-dimensional resistivity variation with depth and generally gives satisfactory results if the subsurface is horizontally layered with a little lateral variation. A line perpendicular to the main axis of the landslide was used for this survey (Fig. 2D). Apparent resistiv-ity data were collected using the Schlumberger electrode array con-ﬁguration, which is the most popular conﬁguration for electrical resistivity sounding. The half-current electrode separation (AB / 2)

was gradually increased from 1 to 80 m to delineate deeper resistiv-ity; the maximum length was set at 80 m due to space limitation. A total of 22 apparent resistivity data was acquired, and the stacking procedures applied to the ERT survey were also applied to enhance the quality of the data. The data were processed using an automatic inversion code based on the damped least-squares solution with singular value decomposition (Ekinci and Demirci, 2008). A homo-geneous earth model was used for the initial guess, and geoelectrical model parameters of the subsurface were recovered after 58 iterations with an RMS error of 8.5%.

4. Results and discussion

4.1. Precipitation and weather events

According to Im, a dry sub-humid climate type is dominant at the Çanakkale station (Table 1) and can be described as‘dry sub-humid, second mesothermal throughout the year, moderate winter water surplus, with a summer concentration of thermal ef_{ficiency equal to} a third mesothermal climate’. A soil moisture surplus occurs in winter months, whereas a soil moisture deficit occurs from May to October (Table 1,Fig. 4). The deficit is particularly pronounced in July and August. Details of the water budget of the study area (Table 2) can be summarised as follows: (1) a period of soil moisture surplus (total of 236 mm) from December to March; (2) a period of soil

Fig. 4. Thornthwaite water budget diagram of the Çanakkale weather station for the period 1930–2005. APE: Monthly adjusted potential evapotranspiration (mm). AE: Monthly actual evapotranspiration (mm). P: Monthly precipitation (mm).

Table 2

Thornthwaite water budget of the Çanakkale station for the period 1930–2005. T: Air temperature (°C). UPE: Unadjusted potential evapotranspiration (mm). APE: Adjusted potential evapotranspiration (mm). P: precipitation (mm). DIFF: P minus APE (mm). ST: Soil moisture storage (mm). CST: Change in storage from the preceding month (mm). AE: Actual evapotranspiration (mm). D: Soil moisture deﬁcit (mm). S: Soil moisture surplus (mm).

Variables Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec Total

T 6.0 6.5 8.0 12.3 17.2 22.0 24.8 24.6 20.7 16.0 11.7 8.2 UPE 13 15 21 41 69 100 121 119 91 61 38 22 APE 11 13 21 45 85 126 153 141 94 58 31 17 796 P 93 73 65 44 32 22 12 6 23 48 88 109 616 DIFF 82 61 44 −1 −53 −104 −141 −135 −72 −10 57 92 ST 100 100 100 99 58 20 5 1 1 1 58 100 CST 0 0 0 −1 −41 −38 −15 −4 −1 0 57 42 AE 11 13 21 45 73 60 27 10 23 48 31 17 380 D 0 0 0 0 12 66 126 131 71 10 0 0 416 S 82 61 44 0 0 0 0 0 0 0 0 49 236

(6)

moisture deﬁcit (416 mm) from May to October with severe dryness in July and August; and (3) a transition in November during which soils begin to store water due to the beginning of frontal precipitation associated with mid-latitude and Mediterranean cyclones.

The relationship between precipitation/weather events and shallow landslide occurrences in the study area was investigated using monthly and daily precipitation totals, daily maximum and minimum surface air temperatures, and daily weather events associ-ated with synoptic-scale weather conditions. According to our interviews with villagers and a local administrator, together with ofﬁcial technical reports from the Çanakkale Public Improvements Administration, the deep-seated landslide occurred on 17 February 2003 during a short period of sequential cold and wet days following a dry spell. In the study area, February is characterised by rainy weather and moist soil conditions due to the dominance of the Mediterranean-type rainfall regime (Table 2).

The normalised monthly precipitation anomalies may show a general relationship between the precipitation conditions and landslide activities in the study area (Fig. 4). Like the deep-seated landslide, the shallow slides occurred on 17 February 2003. This corresponds to a moderately high standardised positive anomaly of rainfall (0.7;Fig. 5).

Because the precipitation appeared insufﬁcient to produce such devastating failures, we investigated the details of the daily

meteoro-logical measurements and weather events for the ﬁrst half of

February 2003 (Fig. 6). A total precipitation of 103.2 mm was

recorded during theﬁrst two weeks of the month. This value exceeds the long-term average total for February (73 mm;Table 2). Weather events for the period include effective rain and snow events and rain/snow showers (Fig. 6). The number of snowy days in February 2003 (eight days) is markedly greater than the long-term average (two days). It is likely that snowfall, showers, and snowy days in the study area are more frequent than those measured at the Çanakkale station located at a lower elevation.

More interestingly, the daily time series of precipitation–temperature values and major weather conditions in February 2003 also show cycles of approximately three days of colder and rainy/snowy conditions followed by three days of warmer and dry/sunny conditions (Fig. 6). This change along with the increased soil water content from rain showers likely increased the pore water pressure, which is a well-known trigger of landslides (Kuriakose et al., 2008). Indeed, the deep-seated landslide movement of 17 February 2003 occurred after an increase in the minimum temperature of ~6 °C on that day and in the maximum temperature of ~5 °C since the last snowy day of 14 February (Fig. 6).

Fig. 5. Variability of normalised monthly precipitation anomalies for the Çanakkale weather station from October 2001 to September 2003. To calculate these anomalies, monthly precipitation totals recorded at the station were normalised using the long-term average precipitation amount and standard deviations for 1930–2008.

Fig. 6. Variability of daily precipitation (mm), maximum and minimum surface air temperatures (°C) and weather events recorded during theﬁrst three weeks of February 2003 at the Çanakkale weather station. The international synoptic meteorological symbols are used: — moderate or heavy rain shower; — moderate or heavy snow shower; — in-termittent fall of snowﬂakes; — slight rain shower — shower of hail, with or without rain; — slight snow shower;

•

— intermittent rain, not freezing; and — slight rain or drizzle and snow.

(7)

The rainy and snowy conditions of theﬁrst half of February 2003 were regional and closely linked to persistent mid-latitude and Mediterranean frontal cyclones over Turkey and/or the eastern Mediterranean basin, which are controlled by upper-air westerlyﬂows.

4.2. Regolith and joint planes as slip surfaces

Slope failures in the abandoned village of Mazılık occurred in the Şahinli Formation, which consists predominantly of Upper Eocene volcanics (Dönmez et al., 2005). Thin section images showed that the rock consists of plagioclase, titanite, opaque minerals, and lesser hornblende and biotite arranged in a hypocrystalline texture (Fig. 7A). Clay coatings are very common in strongly weathered

surfaces (Fig. 7B), which is favourable for slope failures (Morgenstern and Tchalenko, 1967; Shuzui, 2001; Wen and Aydin, 2003).

Samples taken from unweathered or partly weathered basalt, regolith, and clays accumulated on joint planes provided sound evidence for the important role of lithology as a causative factor of the slope failures (Table 3). In terms of mass percentage, the EDX data show that Si and Al, which constitute 44.42% and 24.24% of the total concentration, respectively, are most abundant. The remainder of the composition consists of Fe, Cu, Ca, Mo, Sn, Ru, Zn, Mn, and Na in decreasing order. At this locality, Al, Si, Ca, and Na decrease with

an increasing extent of weathering. As suggested by Tobe and

Chigira (2006) for shallow landslides, this change in elemental composition could be associated with rapid chemical alteration

of plagioclase (Durgin, 1977; Duzgoren-Aydin and Aydin, 2006)

and removal of Ca and Na from the plagioclase lattice (Nixon,

1979). The presence of Fe and Mn coatings, particularly along joints,

Fig. 7. Regolith and joint characteristics. (A, B) Thin section images showing weathered minerals. The greater weathering and clay formation is visible in B. (C) Joint-controlled weathering within regolith proﬁle. (D) Joint plane without regolith as a result of regolith stripping. (E) Rose diagram showing main joint dipping orientations.

Table 3

Results of EDX analysis for the weathering proﬁle of the upper slope with shallow landslides. Elements Unweathered basalt (%) Partly weathered basalt (%) Regolith (%) Joint plane material (%) Na 1.61 0 0 0 Al 28.09 28.49 26.31 22.08 Si 47.12 43.96 42.66 43.94 Ca 5.45 1.97 2.16 2.02 Mn 1.44 0.73 0.38 0.53 Fe 7.30 15.03 11.93 18.57 Cu 1.26 3.45 7.13 4.81 Zn 1.60 1.27 1.35 1.67 Mo 1.41 2.61 3.97 2.38 Ru 1.15 1.44 2.21 1.80 Sn 3.59 1.06 1.89 2.21 Table 4

Results of EDX analysis for three samples collected from slip surfaces. Elements Sample 1 (%) Sample 2 (%) Sample 3 (%) Al 13.21 12.38 10.55 Si 27.95 22.50 26.39 Ca 2.34 1.47 2.29 Fe 12.71 13.67 8.91 C 11.72 11.24 11.97 O 29.19 36.06 36.60 Mg 1.27 1.50 1.08 K 1.61 1.18 2.20

(8)

is also conspicuous, as suggested previously by several researchers (Weaver, 1977; Schwertmann and Taylor, 1989).

Along the steep slope exposed to shallow slides, sub-horizontal planes of orthogonal joint sets with spacing ranging from a few cm to 1 m are visible in many places (Fig. 7C) and the failure surfaces of the shallow landslides coincide with joint faces dipping south-southwest (Fig. 7D). These failure surfaces are covered by a thin mantle of regolith and have a fresh appearance at places where the regolith is entirely stripped. Measurements showed that the principal and secondary joint dips are north–south and east–west, respectively (Fig. 7E).

The weathered residues, especially those at joint planes, contain Fe-rich oxyhydroxides characterised by increased Fe and decreased Mn, suggesting an early stage of weathering under well-drained conditions (Duzgoren-Aydin and Aydin, 2006). Similar results were obtained for three samples collected from slip surfaces of the shallow slides (Table 4), which correspond to the slide-parallel faces of the sub-horizontal joints measured morphometrically at 101 locations. On these weak surfaces with dips ranging from 20° to 30°, EDX showed the predominance of Si, Al, and Fe. Quartz and titanium magnetite were detected by XRD.

Table 5points to significant changes in pH and grain size distribu-tions; the former shows a contrast between steep slopes and the deep-seated slide area. Except for the deep-seated slide, samples taken from various depths of the reference site, the steep slope, and the regolith profile yielded pH values lower than 7, indicating weathering of basalt under slightly acidic conditions. However, the lower slope with the deep-seated main landslide is alkaline. Silty clay makes up ~46% of the textural composition in this area, which is more than that on the reference flat surface (~37%), on the steep slope (~27%), and in the regolith profile (~27%) in the northern shallow slide area. Thus, effective weathering and subsequent removal offine materials from sub-horizontal joint faces (Fig. 7D) to the deep-seated landslide is ongoing. Furthermore, the silty clay mixture of the slip surface material is likely of significance for landsliding, as described elsewhere (Thevanayagam, 1998; Youd and Gilstrap, 1999; Vallejo and Mawby, 2000; Wang and Sassa, 2003; Erginal et al., 2008). We assume that the nature of joint sets and other structural weaknesses, such as bedding planes, faults, and schistosity, is significant for shallow sliding (Santaloia and Cancelli, 1997; Seno and Thuring, 2006; Erginal et al., 2008, 2009; Türke_{ş et al., 2011}). Many recent studies have also shown that susceptibility to sliding is affected mostly by structural factors that allow fast penetration of surface water into the deeper ground (Chigira et al., 2003; Wang et al., 2003).

4.3. Subsurface structure of landslide mass

Fig. 8shows theﬁnal electrical resistivity tomogram with corrected topography obtained by the 2D tomographic inversion process. The image displays a maximum depth of ~10_{–11 m for the resistivity} distri-bution. The resistivity values vary between 1.5 and 70 ohm.m. As the data were acquired in the winter season, higher precipitation caused in-creased moisture, which thus caused a decrease in the overall resistivity. The high clay content of the lowermost material also reduced the resis-tivity. Additionally, the northern upper slope is morphologically hilly and steeper, which allows rapid surface runoff and more inﬁltration at the lower gentler slope.

The conspicuous difference in resistivity values from the topsoil to the deeper parts likely indicatesﬁning of grain size with depth: lower resistivity values occur at the lowermost part. It is also clear that the resistivity values decrease suddenly at a few metres depth, which can be attributed to clay-rich material with high water content. The values exceeding 10 ohm.m in the upper parts may be associated with coarser-grained regolith that is rich in quartz. The ERT image also shows that the clay-rich slip surface highlighted with white arrows inFig. 8dips ~15° downslope and that its depth varies from 3 to 10 m. Thus, we consider that the structural damage to the primary

Table 5

pH and texture characteristics of regolith samples. Locations of R1, R2, RS, and SS are shown inFig. 2D. Depth (cm) pH Clay (%) Silt (%) Sand (%) Shallow slides Reference site (RS) 20 6.50 23.76 17.63 58.61 40 6.69 33.96 15.59 50.45 60 6.41 25.80 13.55 60.65 80 6.57 9.47 9.47 81.06 Mean 6.54 23.24 14.06 62.69 Steep slope (SS) 10 6.86 13.55 19.67 66.78 20 6.11 13.55 15.59 70.86 30 6.28 15.59 7.63 66.78 40 6.21 15.59 9.47 74.94 Mean 6.36 14.57 13.09 69.84 Regolith (R1) 50 6.52 18.37 5.39 76.24 100 6.45 14.29 11.51 74.20 150 6.07 16.33 11.51 72.16 200 6.12 18.37 9.47 72.16 250 6.46 18.37 9.47 72.16 Mean 6.32 17.14 9.47 73.38 Deep seated slide Regolith

(R2) 50 6.20 24.49 17.63 57.88 100 7.23 30.61 15.59 53.80 150 7.27 30.61 13.55 55.84 200 7.20 34.69 15.59 49.71 Mean 6.97 30.10 15.59 54.30

Fig. 8. ERT image produced from the 2D inversion of apparent resistivity data. Dashed line: Clay-rich slip surface. Solid line: Bottom of coarser material. Dotted line: Widening of secondary scarps behind the slipped blocks.

(9)

school occurred due to the separation of the upper coarser-grained layers with higher resistivity from the underlying clay-rich unit. The increase in resistivity values (_{N35 ohm.m) below the school delimited} by thin broken lines inFig. 8is likely due to the widening of the secondary scarp behind the slipped blocks (Fig. 2D). The 1–2 m thick uppermost level between the horizontal distances of 34 and 62 m, as indicated by the solid black line inFig. 8, has relatively high resistivity values due to the presence of coarser materials, observed while locating the electrodes.

The VES survey displays a decrease in resistivity values with depth but ends in a resistive basal layer. The inversion process of the apparent resistivity data revealed a three-layered earth model.

The _{ﬁt between the observed and calculated data is shown in}

Fig. 9A. Fig. 9B shows the recovered ﬁnal earth model after the inversion process. The top layer represents ~2.5 m of regolith material (19 ohm.m) overlying a conductive water-saturated clay-rich layer (2.5 ohm.m) that extends to a depth of ~15 m. These results agree well with those of the ERT survey. The lowermost layer is represented by a resistivity value of 374 ohm.m. This increase in resistivity may reﬂect increased compaction immediately above the bedrock. Field observations also showed that the underlying bedrock crops out at the lowermost part of the deep-seated slide near the valley bottom (Fig. 2D).Fig. 10shows a

cross section of the landslide that summarises the results of theﬁeld observations and the ERT and VES surveys.

5. Conclusion

Shallow and deep-seated regolith slides in a deforested rural area

of Turkey were studied based on ﬁeld observations, microscopic

observations, meteorological data analysis, and geoelectrical resistivity surveys, including ERT and VES techniques. The main conditioning and triggering agents of the slides were discussed. In the upper slope, affected by shallow slides, the removal of weathered residues with Fe-rich oxyhydroxides occurs along sub-horizontal joint faces. The deep-seated landslide at the lower slope, however, caused a destructive displacement and was associated with a total daily precipitation amount of about 16.4 mm on 14 February 2003. A three-day cumulative precipitation (30.1 mm) from 12 to 14 February also facilitated the triggering of the slide. This rotational slide followed a concave-upward slip surface deﬁned between a lower clay-rich unit and an upper coarser unit, based on the electrical resistivity surveys. The clay-rich slip surface developed on the bedrock can also be observed in theﬁeld at the lowermost part of the landslide.

Fig. 9. Results of the VES survey. (A) Observed and calculated apparent resistivity curves. (B) Resistivity–depth distribution of the initial and calculated earth models.

(10)

Acknowledgements

We thank Muhammed Zeynel Öztürk, Rezzan Ekinci, Taner Gürer, Ercan Bilir, and Kenan Arar for their ﬁeld assistance. Dr. Mustafa Bozcu and Dr. Ali Sungur assisted with the analyses of thin sections and regolith, respectively. We thank Graham Lee for linguistic correc-tions of a draft of this paper. We also thank Gökhan Erdoğan and

Evrim Yakut of the İzmir Institute of High Technology for their

support in EDX/SEM and XRD analyses. Professor Takashi Oguchi, the journal Editor, and three anonymous reviewers provided many helpful comments that improved our paper.

References

Bouyoucos, G.J., 1963.Directions for making mechanical analyses of soil by hydrometer method. Soil Science 42, 25–30.

Che, V.B., Fontijn, K., Ernst, G.G.J., Kervyn, M., Elburg, M., Van Ranst, E., Suh, C.E., 2012.

Evaluating the degree of weathering in landslide-prone soils in the humid tropics: the case of Limbe, SW Cameroon. Geoderma 170, 378–389.

Chigira, M., Wang, W.N., Furuya, T., Kamai, T., 2003.Geological causes and geomorpho-logical precursors of the Tsaoling landslide triggered by the 1999 Chi-Chi earthquake, Taiwan. Engineering Geology 68, 259–273.

Conte, E., Troncone, A., 2012.Simpliﬁed approach for the analysis of rainfall-induced shallow landslides. Journal of Geotechnical and Geoenvironmental Engineering 138, 398–406.

Corominas, J., 2001.Landslides and climate. In: Bromhead, E.N. (Ed.), Keynote Lectures Delivered at 8th Int Symp on Landslides, Cardiff, June 2000, pp. 1–33.

Cruden, D.M., Varnes, D.J., 1996. Landslide types and processes. In: Turner, A.K., Shuster, R.L. (Eds.), Landslides: Investigation and Mitigation. Transp Res Board, Spec Rep, 247, pp. 36–75.

De Vita, P., Reichenbach, P., 1998.Rainfall triggered landslides: a reference list. Environmental Geology 35, 219–233.

Dönmez, M., Akçay, A.E., Genç,Ş.C., Acar, Ş., 2005.Biga Yarımadasında Orta-Üst Eosen volkanizması ve denizel ignimbiritler. Maden Tetkik ve Arama Dergisi 131, 49–61 (in Turkish with English abstract).

Drahor, M.G., Göktürkler, G., Berge, M.A., Kurtulmus, T.Ö., 2006.Application of electri-cal resistivity tomography technique for investigation of landslides: a case from Turkey. Environmental Geology 50, 147–155.

Durgin, P.B., 1977.Landslides and the weathering of granitic rocks. Geological Society of America Reviews in Engineering Geology 3, 127–131.

Duzgoren-Aydın, N.S., Aydın, A., 2006.Chemical and mineralogical heterogeneities of weathered igneous proﬁles: implications for landslide investigations. Natural Hazards and Earth System Sciences 6, 315–322.

Eggleton, R.A., Taylor, G., 2001.Regolith Geology and Geomorphology. John Wiley & Sons (375 pp.).

Ekinci, Y.L., Demirci, A., 2008.A damped least-squares inversion program for the interpretation of Schlumberger sounding curves. Journal of Applied Sciences 8, 4070–4078.

Erginal, A.E., Türkeş, M., Ertek, T.A., Baba, A., Bayrakdar, C., 2008.A geomorphological investigation of the excavation induced Dundar Landslide, Turkey. Geograﬁska Annaler Series A: Physical Geography 90, 109–123.

Erginal, A.E., Öztürk, B., Ekinci, Y.L., Demirci, A., 2009.Investigation of the nature of slip surface using geochemical analyses and 2-D electrical resistivity tomography: a case study from Lapseki area, NW Turkey. Environmental Geology 58, 1167–1175.

Göktürkler, G., Balkaya, Ç., Erhan, Z., 2008.Geophysical investigation of a landslide: the Altındağ landslide site, İzmir (western Turkey). Journal of Applied Geophysics 65, 84–96.

Grewelling, T., Peech, M., 1960.Chemical soil test. Cornell University Agricultural Experiment Station Bulletin no. 960.

Kuriakose, S.L., Jetten, V.G., van Westen, C.J., Sankar, G., van Beek, L.P.H., 2008.Pore water pressure as a trigger of shallow landslides in the Western Ghats of Kerala, India: some preliminary observations from an experimental catchment. Physical Geography 29, 374–386.

Lapenna, V., Lorenzo, P., Perrone, A., Piscitelli, S., Sdao, F., Rizzo, E., 2003.High-resolution geoelectrical tomographies in the study of Giarrossa landslide (southern Italy). Bulletin of Engineering Geology and the Environment 62, 259–268.

Loke, M.H., Barker, R.D., 1996.Rapid least-squares inversion of apparent resistivity pseudosections using a quasi-Newton method. Geophysical Prospecting 44, 131–152.

Merrill, G.P., 1897.A Treatise on Rocks, Rock-weathering and Soils. MacMillan Company, New York (411 pp.).

Montrasio, L., Valentino, R., Losi, G.L., 2012.Shallow landslides triggered by rainfalls: modeling of some case histories in the Reggiano Apennine (Emilia Romagna Region, Northern Italy). Natural Hazards 60, 1231–1254.

Morgenstern, N.R., Tchalenko, J.S., 1967.Microstructural observations on shear zones from slips in natural clays. Proc Geotechnical Conf. Oslo, vol. 1, pp. 147–152.

Nixon, R.A., 1979.Differences in incongruent weathering of plagioclase and microcline-cation leaching versus precipitates. Geology 7, 221–224.

Salciarini, D., Tamagnini, C., Conversini, P., Rapinesi, S., 2012.Spatially distributed rainfall thresholds for the initiation of shallow landslides. Natural Hazards 61, 229–245.

Santaloia, F., Cancelli, A., 1997.Landslide evolution around Mt Campastrino (Northern Apennines, Italy): a complex and composite gravitational movement. Engineering Geology 45, 217–232.

Sasaki, Y., 1992.Resolution of resistivity tomography inferred from numerical simula-tion. Geophysical Prospecting 40, 453–464.

Schwertmann, U., Taylor, R.M., 1989.Iron oxides, In: Dixon, J.N., Weed, S.B. (Eds.), Minerals in Soil Environments, 2nd edition. Soil Science Society of America, SSSA Book Series. Madison, Wisconsin, pp. 179–438.

Seno, S., Thuring, M., 2006.Large landslides in Ticino, southern Switzerland: geometry and kinematics. Engineering Geology 83, 109–119.

Shuzui, H., 2001.Process of slip surface development and formation of slip surface clay in landslide in Tertiary volcanic rocks, Japan. Engineering Geology 61, 199–219.

Skempton, A.W., Petley, D.J., 1967.The strength along structural discontinuities in stiff clays. Proc Geotechnical Conf. Oslo, vol. 2, pp. 29–46.

Suzuki, K., Higashi, S., 2001.Groundwaterﬂow after heavy rain in landslide-slope area from 2-D inversion of resistivity monitoring data. Geophysics 66, 733–743.

Thevanayagam, S., 1998.Effects ofﬁnes and conﬁning stress on undrained shear strength of silty sands. Journal of Geotechnical and Geoenvironmental Engineering 124, 479–491.

Thornthwaite, C.W., 1948.An approach toward a rational classiﬁcation of climate. Geographical Review 38, 55–94.

Tobe, H., Chigira, M., 2006.Causes of shallow landslides of weathered granitic rocks from the view point of weathering styles and petrologic textures. In: Marui, H.E.A. (Ed.), Disaster Mitigation of Debris Flows, Slope Failures and Landslides. Universal Academy Press, Tokyo, pp. 493–501.

TSMS (Turkish State Meteorological Service), 2009a. Climatological Bulletin and Monthly Precipitation and Temperature Series of the Çanakkale Meteorology Station, Ankara.

TSMS (Turkish State Meteorological Service), 2009b.Daily Weather Bulletins and Synoptic Weather Maps, Ankara.

Türkeş, M., 1996.Spatial and temporal analysis of annual rainfall variations in Turkey. International Journal of Climatology 16, 1057–1076.

Türkeş, M., 1998.Inﬂuence of geopotential heights, cyclone frequency and Southern Oscillation on rainfall variations in Turkey. International Journal of Climatology 18, 649–680.

Türkeş, M., 1999.Vulnerability of Turkey to desertiﬁcation with respect to precipitation and aridity conditions. Turkish Journal of Engineering and Environmental Sciences 23, 363–380.

Türkeş, M., 2010.Climatology and Meteorology. Physical Geography Series, 1. Kriter Publisher,İstanbul (650 + XXII pp. (in Turkish)).

Türkeş, M., Akgündüz, A.S., 2011.Assessment of the desertiﬁcation vulnerability of the Cappadocian district (Central Anatolia, Turkey) based on aridity and climate-process system. International Journal of Human Sciences 8, 1234–1268.

Türkeş, M., Sümer, U.M., Demir, İ., 2002.Re-evaluation of trends and changes in mean, maximum and minimum temperatures of Turkey for the period 1929–1999. International Journal of Climatology 22, 947–977.

Türkeş, M., Erginal, A.E., Demirci, A., Ekinci, Y.L., 2011.Geophysical, geomorphological and climatological analysis of the Ambaroba and Mazılık landslides in the Çanakkale district. 5th Atmospheric Science Symposium Proceedings, Istanbul Technical University, 27–29 April, İstanbul, Turkey, pp. 461–474 (in Turkish with English abstract).

UNCCD, 1995.The United Nations Convention to Combat Desertiﬁcation in those Countries Experiencing Serious Drought and/or Desertiﬁcation, Particularly in Africa. Text with Annexes.United Nations Environment Programme (UNEP), Geneva.

Vallejo, L.E., Mawby, R., 2000.Porosity inﬂuence on the shear strength of granular material–clay mixtures. Engineering Geology 58, 125–136.

Wang, G., Sassa, K., 2003.Pore-pressure generation and movement of rainfall-induced landslides: effects of grain size andﬁne-particle content. Engineering Geology 69, 109–125.

Wang, W.N., Chigira, M., Furuya, T., 2003.Geological and geomorphological precursors of the Chiu-Fen-Erh-Shan landslide triggered by the Chi-Chi earthquake in central Taiwan. Engineering Geology 69, 1–13.

Weaver, C.E., 1977.Mn–Fe coatings on saprolite fracture surfaces. Journal of Sedimentary Research 48, 595–610.

Wen, B.P., Aydin, A., 2003.Microstructural study of a natural slip zone: quantiﬁcation and deformation history. Engineering Geology 68, 289–317.

Willmott, C.J., 1977.WATBUG: a FORTRAN IV algorithm for calculating the climatic water budget. Publications in Climatology, 30. Water Resources Center, University of Delaware, Newark, Delaware 1–55.

Youd, T.L., Gilstrap, S.G., 1999.Liquefaction and deformation of silty andﬁne-grained soils. Proc 2nd Int Conf on Earthquake Geotechn Eng Lisbon, Portugal, vol. 3, pp. 1013–1020.