Introduction

Liquefaction of soil is defined as saturated or partially saturated cohesionless soil deposits substantially losing their strength and stiffness in response to an applied stress, usually earthquake shaking. In this phenomenon, the soil behaves much like a dense, viscous liquid. Liquefaction is one of the most important, interesting, complex, and controversial topics in geotechnical earthquake engineering. In 1964, two major earthquakes that attracted the attention of geotechnical engineers occurred in Niigata (Japan) and Alaska. Both earthquakes induced soil liquefaction producing extensive structural damage and loss of life in urban areas. In the 50 years since these earthquakes, liquefaction has been studied extensively by hundreds of researchers around the world (Kramer 1996).

The liquefaction potential is evaluated by field tests in engineering geology and geotechnical engineering studies. Different types of in situ tests including the standard penetration test (SPT), the cone penetration test (CPT), measurement of in situ shear wave velocity (Vs), and the Becker penetration test (BPT) are used to predict the liquefaction potential of soil (Youd et al. 2001). Among different types of in situ tests, the SPT (ASTM D1586-99 2004) is generally preferred for the evaluation of liquefaction in most countries and is used in this paper. Seed and Idriss (1971) suggested a number of SPT-based simplified methods. This method has been modified and improved for over 45 years by several researchers (e.g., Seed et al. 1985; Youd et al. 2001; Idriss and Boulanger 2004; Cetin et al. 2004).

The study site, at the coordinates between 4,441,000–4,447,000N latitude and 449,000–453,000E longitude (UTM Zone 35N, ED50), covers an area of about 17 km2 (Fig. 1). Canakkale is one of the important cities in Turkey connecting the two continents (Europe and Asia) and is a developing city. The Canakkale settlement area contains three different soil types. Quaternary alluvial deposits are one of the soil types and are located everywhere in the Canakkale city. Distribution and generation of these alluvial deposits are linked to the impact of the Saricay River which flows through the city from east to west. The groundwater-level variation in alluvium is between 0.6 and 9.9 m below the surface; the mean groundwater depth in the Canakkale city centre is about 2.0 m (Fig. 2). Canakkale city, which is located in the northwestern part of Turkey, is bounded by north and central segments of the North Anatolian Fault Zone (NAFZ) (Yaltırak et al. 2012). The city of Canakkale is situated in a seismically active region according to the earthquake zone map of Turkey produced by the Ministry of Public Works and Settlement, General Directorate of Disaster Affairs (1996).

Fig. 1
figure 1

Location map of the study area

Full size image
Fig. 2
figure 2

Groundwater map of the study area

Full size image

The main purpose of this paper is to determine the soil vulnerability to liquefaction for different earthquake scenarios and to prepare liquefaction severity maps for the city of Canakkale using the liquefaction severity index (LS) and liquefaction potential index (LPI). The study involved three stages: field work, laboratory testing, and generation of the liquefaction severity maps. In the first step, geotechnical boreholes were drilled and SPTs were performed, and in addition, disturbed samples were taken using an SPT sampler. In the second step, grain-size distribution was determined by means of laboratory testing. Field-based SPT values were used to calculate the factor of safety against liquefaction (FL). The FL of a saturated soil is expressed as the ratio of the cyclic resistance ratio (CRR) to the cyclic stress ratio (CSR). The liquefaction-prone areas in the city of Canakkale were examined based on a simplified procedure by Youd et al. (2001). In addition, the calculation of probability of liquefaction (PL) and LS correlations was determined according to a recent method proposed by Sonmez and Gokceoglu (2005). In addition, the LPI modified by Sonmez (2003) was used and compared with the LS. These methods were applied to determine the liquefaction potential of different areas (Yilmaz and Bagci 2006; Sonmez and Ulusay 2008; Yalcin et al. 2008; Dixit et al. 2012).

Geological setting and seismotectonics of the Gallipoli and Biga Peninsulas

Canakkale settlement area is located in the Canakkale basin which is formed by Middle Miocene-Pliocene sedimentary units (“Canakkale Group”, Siyako 2006) and alluvial deposits. The Canakkale Group consists of the Gazhanedere Formation, Kirazli Formation, Camrakdere Formation, and Alcitepe Formation from bottom to top. The Kirazli and Camrakdere Formations are absent in the study area (Fig. 3). The Gazhanedere Formation outcrops in the southeastern part of the study area. It consists of conglomerate, sandstone, and reddish grey mudstone (Atabey et al. 2004).

Fig. 3
figure 3

Geological map of the study area

Full size image

The Alcitepe Formation outcrops in the northern part of the study area. It generally consists of sandstone, mudstone, sandy, and clayey limestone (Senturk and Karakose 1987). A large part of the Canakkale settlement area is located on alluvial deposits which were formed by the Saricay River. Typical cross sections depicting the subsurface profile of the study area are given in Fig. 4.

Fig. 4
figure 4

Engineering geological cross sections for the study area

Full size image

Turkey is located in seismically active region of the world. The NAFZ is a 1200 km-long transform fault forming the boundary between the Anatolian and Eurasian plates (Okay et al. 2000). Canakkale, which is situated in the northwest of Turkey, is approximately 40 km southeast of the Saros–Gazikoy segment which is the northern branch of the NAFZ in the Gallipoli peninsula region (Fig. 5). The Saros–Gazikoy segment is locally called the Ganos fault (Rockwell et al. 2001). In 1912, the Ganos fault generated an earthquake with a magnitude (Ms) of 7.3. The central branch of the NAFZ which is located in the Biga peninsula region is composed of many fault zones. The Yenice–Gonen fault is one of these. The southeast of the study area is 70 km from the Yenice–Gonen fault. In 1953, the Yenice–Gonen fault generated the Yenice–Gonen earthquake with a magnitude (Ms) of 7.2. In addition to the faults mentioned, the Edremit fault zone is located farthest from the study area is. It is 75 km from the south of the study area. In 1944, this fault zone produced an earthquake of magnitude (Ms) 6.8 (Fig. 5).

Fig. 5
figure 5

Active tectonic map of the Marmara region (Gurbuz et al. 2000)

Full size image

More recently, the 2014 Aegean Sea earthquake occurred in the northern Aegean Sea on May 24 at 12:25:01 (local time). It had moment magnitude (Mw) of 6.5 and depth of 25.02 km with epicentral coordinates determined as 36.75230 N, 36.03700 E (KOERI 2014; AFAD 2014). The epicentre of the 2014 Aegean Sea earthquake was 93 km away from Canakkale. The peak ground acceleration (amax) recorded was 141 gal (NS component) at the Canakkale station (AFAD 2014). This earthquake was strongly felt in the centre of Canakkale. However, no serious damage and liquefaction phenomena were reported in the city of Canakkale (AFAD 2014; Yildirim et al. 2015).

Determination of peak ground acceleration

Earthquakes with magnitudes of 4.0 and larger (Ms ≥ 4.0) occurring in the study area and within a radius of 100 km from 1900 to 2018 which were taken from a website based on Disaster and Emergency Management Presidency (AFAD) records are shown in Fig. 6. The equation (Eq. 1) proposed by Well and Coppersmith (1994) can be used to estimate the largest possible earthquake that can be produced by a fault in terms of Mw. In this proposed equation, the surface rupture length (SRL) is used as the input parameter.

Fig. 6
figure 6

Distribution of earthquakes (Ms ≥ 4.0) in the study area and its surroundings (http://www.deprem.gov.tr/en/eventcatalogue)

Full size image
$${M_{\text{w}}}=5.08+1.16 \times \log ({\text{SRL}}).$$
(1)

In this study, the length of the fault is considered to be more realistic for the prediction of magnitudes, instead of SRL. The fault lengths were taken from the active fault map of Turkey produced by the General Directorate of Mineral Research and Exploration (MTA) (2012). In Table 1, the three major faults located within a radius of 100 km from the centre of the study area are considered (Tunusluoglu 2014). Based on this information, the seismicity of the study area is considered to be quite high. In liquefaction assessment methods, the triggering effect of an earthquake on liquefaction should be taken into consideration as amax produced by the earthquake. The amax considered for liquefaction assessment in the study area was evaluated with the attenuation equation proposed by Ulusay et al. (2004). This relationship is based on the large Turkish earthquake database and considers different ground conditions:

Table 1 Important faults belonging to the study area and its surroundings, and the largest earthquake moment magnitudes produced by these faults (Tunusluoglu 2014)
Full size table
$${a_{\hbox{max} }}={\text{ }}2.18{\text{ }}{{\text{e}}^{0.0218{\text{ }}(33.3{M_{\text{w}}}-{\text{ }}\operatorname{Re} {\text{ }}+{\text{ }}7.8427{\text{ }}{S_{\text{A}}} + {\text{ }}18.9282{S_{\text{B}}}}}^{)},$$
(2)

where SA = 0, SB = 0 (rock), SA = 1, SB = 0 (soil), and SA = 0, SB = 1 (loose soil).

A large part of the investigated area is composed of alluvium. For this reason, when taking SA = 0, SB = 1, the closest distance of each fault to the investigation area was considered to be Re. The largest horizontal ground acceleration values which can be produced by the three major faults were calculated using the attenuation equation (Eq. 2), and the results are given in Table 2. As seen from Table 2, the amax that can be produced by an earthquake of 7.5 moment magnitude in the Saroz–Gazikoy Fault Zone was calculated as 319 gal. In addition, a second possible earthquake scenario was considered at a magnitude (Mw) of 7.0 along the Saroz–Gazikoy Fault Zone. The amax was calculated as approximately 222 gal for the Saroz–Gazikoy Fault Zone (Table 2).

Table 2 Peak ground accelerations to be produced by faults located in the study area and its surroundings
Full size table

The 2014 Aegean Sea earthquake was recorded by three accelerometer stations in Canakkale city (AFAD 2014). The amax recorded was 141 gal (NS component) at the Canakkale centre station. In addition, using the Aegean Sea earthquake data, the amax value was calculated using the attenuation equation proposed by Ulusay et al. (2004). The calculated peak ground acceleration (amax = 46.5 gal) value is close to the values measured at Canakkale-Kepez station (amax = 45.3 gal NS component and 51.1 gal EW component). However, the peak ground acceleration (amax = 141 gal) value was used in this study.

Field studies and geotechnical evaluation

151 geotechnical boreholes were opened with a total depth of 3000 m to obtain disturbed and undisturbed soil samples, to implement the SPT, and to determine the static water level from the ground surface (Buyuksarac et al. 2013). The depth of the groundwater level varies between 0.6 and 17.1 m depth. In addition, groundwater depth was mapped (Fig. 2). During the 2 years of measurements, the variation in groundwater levels was determined to be between 0.1 and 3.5 m (Deniz 2005). The depths of 140 boreholes were 20 m and the depths of other boreholes range between 15 and 30 m. During the field studies, SPTs (ASTM D1586-99 2004) were conducted at every 1.5 m depth. The Canakkale settlement area comprises three different units. The Alcitepe and Gazhanedere Formations are described as rock materials. Quaternary alluvial deposits are located at the centre, the eastern, the northern, and the southern parts of Canakkale city (Fig. 3). 131 geotechnical boreholes were drilled in alluvial deposit areas. As shown in Fig. 4, most of the areas in the investigated region are alluvium. Sandy areas (SP, SW, SC, and SM) are shown as alluvium in general and clays are shown separately. In some places in the alluvial deposits, the SPT-N values are 10 at the ground surface with a depth of 5 m. At the lower strata, the SPT-N increases, but the density of the soils at these depths can be described as loose-to-moderately dense. In addition, at depths between 10 and 20 m, the average value of SPT-N counts varies between 10 and 25 (Fig. 7). However, in some locations in the alluvial deposits, drill rods were swamped by weight in the borehole at depths between 4.5 and 12 m.

Fig. 7
figure 7

Selected engineering logs illustrating the SPT-N values and groundwater conditions in the study area

Full size image

Laboratory tests were undertaken on undisturbed and disturbed soil samples taken by Shelby tube and SPT sampler from the 151 boreholes, to determine the index properties of the soil samples. They included natural moisture content, unit weight, grain-size distribution (sieve and hydrometer methods), and Atterberg limits. The laboratory tests were carried out following the procedures described by ASTM (2004). The study area is formed by alluvium with fine-grained soils in the upper levels and coarse-grained soils in the lower levels. Unit weight was determined from the undisturbed samples which were taken at the depth of 2.5–3.0 m. Unit weight values vary between 18.1 and 18.4 kN/m3. Natural moisture contents of the fine-grained soils varied between 7.3 and 67.6%. The fine-grained soils in the study area are formed, on average, of 1.4% gravel, 20.8% sand, and 77.8% fines (silt–clay). The liquid limit of the fine-grained soils changes between 24.0 and 73.5%. The plasticity index values vary between 7.0 and 44.0%. Coarse-grained soils are comprised of sand and form the bottom layer of alluvium. Natural moisture contents of the coarse-grained soils varied between 4.3 and 61.5%. The coarse-grained soils in the study area consist of, on average, of 3.0% gravel, 85.1% sand, and 11.9% fines (silt–clay). The liquid limit of the coarse-grained soils changes between 23.5 and 56.5%. The plasticity index values range between 5.9 and 27.2% (Table 3).

Table 3 Geotechnical properties of the Quaternary alluvium at the centre, the eastern, the northern, and the southern of Canakkale city
Full size table

The Unified Soil Classification System (USCS) was used to describe the soils which were taken from boreholes. Based on the results of the grain-size analysis, 80% of the samples are coarse-grained and 20% of the samples are fine-grained soils. The fine-grained fraction mostly consists of low plasticity clay of classification CL (10.1%) and CH (8.1%), and in a few locations, inorganic silt of classification, ML, MH, and CL–ML were also encountered. All the coarse-grained soils in the study area consist of sand and 25% of them are SM (silty sand) according to USCS. Figure 8 shows the upper and lower limits in the gradation curves for liquefiable and potentially liquefiable soils and the position of grain-size distribution of the soils collected from different depths by SPT tubes. It was seen that the obtained grain-size distribution curves fell within the range of liquefiable soils as specified by the Japanese Port and Harbor Research Institute (JPHRI 1989 from Lee et al. 2004).

Fig. 8
figure 8

Grain-size distribution of the soils

Full size image

Liquefaction assessment

The liquefaction potential of an area can be evaluated by laboratory tests or in situ tests and empirical methods. Due to the difficulties associated with obtaining good soil samples, empirical approaches based on the in situ penetration test methods are widely used. Among different types of in situ tests, the SPT is generally preferred for evaluating liquefaction potential in most countries (Chu et al. 2004; Vipin et al. 2010; Dixit et al. 2012; Sharma and Hazarika 2013; Kang et al. 2014; Rahman et al. 2014) and in Turkey (Ulusay et al. 2000, 2007; Ulusay and Kuru 2004; Yilmaz and Yavuzer 2005; Yilmaz and Bagci 2006; Hasancebi and Ulusay 2006; Yalcin et al. 2008; Sonmez and Ulusay 2008; Sonmez et al. 2008; Ulamis and Kilic 2008; Tosun et al. 2011; Akin et al. 2013; Duman et al. 2015). The liquefaction potential in this study was investigated by the simplified SPT-based method proposed by Seed and Idriss (1971) and Seed et al. (1985). It also considered the modifications suggested by Youd et al. (2001). In this study, corrected SPT-N values were used using the equation suggested by Liao and Whitman (1986). Seed and Idriss (1971) proposed the cyclic stress ratio (CSR) which is defined as cyclic shear stress required to cause liquefaction. Youd et al. (2001) proposed a slight modification in the calculation of CSR. Seed et al. (1985) presented an empirical correlation which is determined from the correlation between corrected SPT (N1)60 and the CSR. The empirical correlation curves, which are the same as the liquefaction triggering curves, represent the capacity of the soil to resist liquefaction referred to as the cyclic resistance ratio (CRR). Youd et al. (2001) modified the CRR curves from Seed et al. (1985). This CRR modification is comprised of clean sands and magnitude 7.5 earthquakes. CRR curves represent limiting conditions that determine whether liquefaction will occur for a magnitude of 7.5. In addition, CRR curves were developed for granular soils with the fines contents of 5% or less, 15, and 35% (Youd et al. 2001). In this study, the CRR equation proposed by Youd et al. (2001) was taken into consideration. To assess the liquefaction analyses, the equation for factor of safety (FS) against liquefaction is written in terms of CRR7.5, CSR, and MSF as follows (Eq. 3):

$${\text{FS }}={\text{ }}\left( {{\text{CR}}{{\text{R}}_{{\text{7}}.{\text{5}}}}{\text{/CSR}}} \right){\text{MSF.}}$$
(3)

The magnitude scaling factor (MSF) in Eq. 4 was employed in this study (Youd et al. 2001):

$${\text{MSF}}=\frac{{{{10}^{2.24}}}}{{M_{{\text{w}}}^{{2.56}}}}.$$
(4)

In this study, the distributions of the corrected SPT (N1)60 data versus the CSR data are compared with the CRR curves modified by Youd et al. (2001) (Fig. 9). This comparison was made for two different amax. In Fig. 9, 100 of 131 borehole data have high and very high liquefaction intensity indices for 319 gal. These borehole data fall to the left of the curve (FC = 15%) and 8 borehole data are located between the two curves (FC = 35% and FC = 15%). For 222 gal, 34 borehole data from 131 borehole data have high and very high liquefaction intensity indices. Twenty-one borehole data fall to the left of the curve (FC = 35%) and 3 borehole data are located between the two curves (FC = 15% and FC = 5%).

Fig. 9
figure 9

Cyclic stress ratio and SPT (N1)60 plot based on Youd et al. (2001)

Full size image

In addition, the hydrometer and the Atterberg limits experiments were carried out on 177 soil samples taken from different depths of 73 boreholes in alluvial deposit areas. Fifty-two samples from a total of 177 samples are composed of silty clay (CL = 48 unit), silty sand (ML = 3 unit) and silty clay–silty sand (CL-ML = 1 unit) soils. The liquefiability of these soils was also investigated using the method proposed by Seed et al. (2003) (Fig. 10). Six samples falling into the A region and 46 samples falling into the B region were defined using this method (Fig. 10). According to the procedure of Seed et al. (2003), 6 samples falling into zone A and 34 samples falling into zone B were defined as liquefiable.

Fig. 10
figure 10

Liquefaction potential of silty soils (Seed et al. 2003)

Full size image

In the literature, the liquefaction potential of the loose sand layers which are susceptible to liquefaction up to a depth of 20 m from the surface is assessed (Seed and Idriss 1971; Sonmez 2003). Studies indicate that the liquefaction phenomenon occurs when the factor of safety is less than 1.0. In addition, it is stated that the liquefaction phenomenon is defined as marginally liquefiable with the factor of safety ranging between 1.0 and 1.2, but the liquefaction phenomenon will not occur above 1.2 (Tosun and Ulusay 1997; Ulusay and Kuru 2004; Duman et al. 2015). However, Seed and Idriss (1982) indicated that the admissible factor of safety value for the liquefaction phenomenon is between 1.25 and 1.5.

Iwasaki et al. (1982) proposed a liquefaction potential index (LPI) to remove the limitations of the factor of safety (FL). LPI was defined in four categories of liquefaction potential (very low, low, high, and very high). The LPI is calculated by the following equations (Eqs. 5a, 5b, 5c, 5d, 5e):

$${\text{LPI}}=\int\limits_{0}^{{20}} {F\left( z \right)W\left( z \right){\text{d}}z} ,$$
(5a)
$$F\left( z \right)=1 - {F_{\text{L}}}\quad {\text{for }}{F_{\text{L}}}<{\text{ }}1.0,$$
(5b)
$$F\left( z \right)=0\quad {\text{for }}{F_{\text{L}}} \geqslant {\text{ 1}}.0,$$
(5c)
$$W(z)=10 - 0.5z\quad z{\text{ }}<{\text{ 2}}0{\text{ m,}}$$
(5d)
$$W(z)=0\quad z{\text{ }} \geqslant {\text{ 2}}0{\text{ m,}}$$
(5e)

where z is the depth of the midpoint of the soil layer in metres and FL is the factor of safety against liquefaction. Boundary values of LPI are given in Table 4 with liquefaction susceptibility descriptions.

Table 4 Liquefaction potential classification suggested by Iwasaki et al. (1982)
Full size table

However, this classification has some limitations. These limitations which are “non-liquefiable” and “moderate” liquefaction potential categories are not defined in the LPI. Therefore, Sonmez (2003) modified the classification by including these two categories in the classification (Table 5). In addition, Sonmez (2003) modified the F(z) term. F(z) is expressed by the following equations (Eqs. 6a, 6b, 6c):

Table 5 Liquefaction potential classification suggested by Sonmez (2003)
Full size table
$$F\left( z \right)=0\quad{\text{for }}{F_{\text{L}}} \geqslant {\text{ 1}}.{\text{2,}}$$
(6a)
$$F(z)=2 \times {10^6}{{\text{e}}^{ - 18.427{F_{\text{L}}}}}\quad{\text{for 1}}.{\text{2 }}>{\text{ }}{F_{\text{L}}}<{\text{ }}0.{\text{95,}}$$
(6b)
$$F(z)=1 - {F_L}\quad{\text{for }}{F_{\text{L}}}<{\text{ }}0.{\text{95.}}$$
(6c)

The FL = 1.2 threshold value used in this modification is considered to be the lowest value at which liquefaction does not occur. However, Sonmez (2003) stated that the threshold value of FL = 1.2 from marginally liquefiable to non-liquefiable could change according to the results of future studies.

The probability of soil liquefaction (PL) depends on the value of FL suggested by Juang et al. (2003). Sonmez and Gokceoglu (2005) proposed the liquefaction severity index (LS) and the authors also preferred to use PL in this index equation. It is stated by Sonmez and Gokceoglu (2005) that the use of the liquefaction probability equation in the concept of liquefaction index will be more consistent. The LS is calculated by the following equations (Eqs. 7a, 7b, 7c, 7d, 7e):

$${L_{\text{S}}}=\int\limits_{0}^{{20}} {{P_{\text{L}}}\left( z \right)W\left( z \right){\text{d}}z}$$
(7a)
$${P_{\text{L}}}=\frac{1}{{1+{{({F_{\text{L}}}/0.96)}^{4.5}}}}\quad {\text{for }}{F_{\text{L}}} \leqslant {\text{ 1}}.{\text{411,}}$$
(7b)
$${P_{\text{L}}}(z)=0\quad {\text{for }}{F_{\text{L}}}>{\text{ 1}}.{\text{411,}}$$
(7c)
$$W(z)=10 - 0.5z\quad z{\text{ }}<{\text{ 2}}0{\text{ m}}$$
(7d)
$$W(z)=0\quad z{\text{ }} \geqslant {\text{ 2}}0{\text{ m,}}$$
(7e)

where z is the depth of the midpoint of the soil layer in metres and FL is the factor of safety against liquefaction. Boundary values of LS are given in Table 6 with liquefaction susceptibility descriptions.

Table 6 Liquefaction severity classification suggested by Sonmez and Gokceoglu (2005)
Full size table

Liquefaction severity map of Canakkale settlement area

Preparation of the liquefaction severity maps is a valuable solution to evaluate the effects of soil liquefaction in urban planning and earthquake-resistant design of constructions. In this study, the liquefaction potential of the Canakkale settlement area was investigated based on the liquefaction severity index and liquefaction potential index using the methods proposed by Sonmez and Gokceoglu (2005) and Sonmez (2003), respectively. Two possible earthquake scenarios with an Mw and amax of 7.5 and 319 gal and 7.0 and 222 gal, respectively, were considered. Furthermore, these two methods were analysed using the amax value measured at the Canakkale station during the 2014 Aegean Sea earthquake. According to both liquefaction methods, for the amax of 141 gal, liquefaction analyses were also performed. All liquefaction analyses were carried out for 131 boreholes (Tables 7, 8, 9). The liquefaction analysis results obtained for the three different amax values are shown in the pie charts (Fig. 11). According to liquefaction severity index and liquefaction potential index analyses, for amax = 319 gal, the Canakkale settlement area is in the categories of high severity (60%) and very high potential (91%), respectively. Moderate areas comprise 16% severity and 1% potential, respectively. For amax = 222 gal, the Canakkale settlement area is in the categories of moderate severity (62%) and very high potential (59%), respectively. High areas comprise 21% severity and 30% potential, respectively. For amax = 141 gal, the Canakkale settlement area is in the categories of moderate (33%) severity and high (36%) potential, respectively. Low areas are 47% severity and 25% potential, respectively. However, no liquefaction phenomena were reported after the 2014 Aegean Sea earthquake (AFAD 2014; Yildirim et al. 2015). For this reason, it is seen that the liquefaction severity index method is more consistent when the two methods are compared. The liquefaction severity and liquefaction potential index maps of the Canakkale settlement area were created by means of geographic information systems (GIS) software (Figs. 12, 13, 14, 15, 16, 17). As shown in Figs. 12, 13, 14, 15, 16, and 17, the Alcitepe and Gazhanedere Formations are described as non-liquefiable soil (LS = 0) in the study area. Quaternary alluvial deposits in the study area were divided into five classes from very low to very high liquefaction.

Table 7 Liquefaction severity index (LS) and liquefaction potential index (LPI) values for amax = 319 gal (Fm: formation, Qal: alluvium, Tmal: alcitepe Fm, and Tmg: gazhanedere Fm)
Full size table
Table 8 Liquefaction severity index (LS) and liquefaction potential index (LPI) values for amax = 222 gal (Fm: formation, Qal: alluvium, Tmal: alcitepe Fm, and Tmg: gazhanedere Fm)
Full size table
Table 9 Liquefaction severity index (LS) and liquefaction potential index (LPI) values for amax = 141 gal (Fm: formation, Qal: alluvium, Tmal: alcitepe Fm, and Tmg: gazhanedere Fm)
Full size table
Fig. 11
figure 11

Pie charts showing the areas of the liquefaction severity index and liquefaction potential index. a amax = 319 gal, b amax = 222 gal, and c amax = 141 gal

Full size image
Fig. 12
figure 12

Liquefaction severity map (amax = 319 gal)

Full size image
Fig. 13
figure 13

Liquefaction severity map (amax = 222 gal)

Full size image
Fig. 14
figure 14

Liquefaction severity map (amax = 141 gal)

Full size image
Fig. 15
figure 15

Liquefaction potential map (amax = 319 gal)

Full size image
Fig. 16
figure 16

Liquefaction potential map (amax = 222 gal)

Full size image
Fig. 17
figure 17

Liquefaction potential map (amax = 141 gal)

Full size image

In Turkey, earthquake–soil–building interaction and the importance of liquefaction were proven after the 1998 Ceyhan-Adana and 1999 İzmit earthquakes. Afterwards, Turkish earthquake codes were published in different years. Unfortunately, many buildings were built before these codes without taking necessary precautions. The province of Canakkale is also located in a seismically active region. The population of the Canakkale settlement area has increased in the last 2 decades. Especially after establishment of the university in 1992, the city started to develop and housing demands increased. When we look at the residential areas in Fig. 12, it is seen that almost the entire city area is open for settlement. Although not all of these areas have been constructed, these areas are open for construction and housing will be built in the future. Residential areas in Canakkale were assessed and classified according to their physical vulnerability to natural hazards (earthquakes, landslides, and stream floods) by Basaran et al. (2014). In this study, the building density areas and few-multi storey housing areas were evaluated in five categories. In the study by Basaran et al. (2014), semi-organized and medium density housing developed to the south of Sarıcay River, and organized, multistorey housing remained to the north of Sarıcay River. In addition, housing areas in Canakkale were grouped into four risk zones according to their physical vulnerability to natural hazards by Basaran et al. (2014). It is clear that the liquefaction maps produced in this study match the high-risk zones map of Basaran et al. (2014). As a result, when we overlay the liquefaction map and development plan, it is obvious that some areas may be affected negatively by a possible earthquake. Therefore, understanding the ground conditions and construction of buildings in accordance with earthquake regulations may reduce structural damage.

Conclusion

In this study, geological and geotechnical investigations were carried out in the Canakkale settlement area and the liquefaction severity of the Quaternary alluvial deposits was determined using the methods of liquefaction severity index and liquefaction potential index using two probable earthquake scenarios. In addition, the peak ground acceleration value measured at the Canakkale station during the 2014 Aegean Sea earthquake was used to analyse these two methods. In addition, the possible environmental impacts of liquefaction were discussed. The main results obtained are the following:

  1. 1.

    The Quaternary alluvial deposits are generally susceptible to liquefaction and cover most of the Canakkale settlement area. It was also determined that the groundwater level is rather shallow. In addition, Canakkale city is located in a seismically active region in Turkey. All these conditions are compatible with the occurrence of the liquefaction phenomenon.

  2. 2.

    The liquefaction severity and liquefaction potential index maps for the Canakkale settlement area were prepared. According to the liquefaction severity index for amax = 319 gal, the category of very high and high susceptibility for liquefaction was observed in the eastern and south parts of Canakkale city. The moderate category of the liquefaction severity index was mainly observed at the centre, the southern, and the eastern parts of Canakkale city for the amax = 222 gal.

  3. 3.

    According to the liquefaction potential index, almost all of the alluvial deposits are in the very high liquefaction category for amax = 319 gal. The category of very high liquefaction potential was mainly observed at the centre and the southern parts of Canakkale city for the amax = 222 gal.

  4. 4.

    Almost all of the Quaternary alluvial deposits are in the category of moderate and low for amax = 141 gal according to the liquefaction severity index. However, according to the liquefaction potential index, the Quaternary alluvial deposits are in the category of high and moderate for the same peak ground acceleration value. Thus, it is considered that the liquefaction severity index is more reliable than the liquefaction potential index.

  5. 5.

    In the study area, the Gazhanedere and Alcitepe Formations consist of sandstone and mudstone. Therefore, these formations were accepted as non-liquefiable soil.

The results of this study and liquefaction susceptibility maps can be used for urban-regional planning and risk management practices in Canakkale. In addition, the geotechnical properties of the liquefiable soils must be considered for proper foundation design of buildings in the Canakkale settlement area.