The 2023 Kahramanmaras x earthquake sequence produced severe damage in southeastern Tu¨rkiye and resulted in significant loss of life. Extensive evidence of liquefaction-related damage was identified by the National Science Foundation (NSF) funded Geotechnical Extreme Events Reconnaissance (GEER) teams that worked with researchers from the Middle Eastern Technical University (METU) and other engineers. This liquefactionrelated damage is documented in Cetin et al. (2023) and Moug et al. (2023). In addition, Moug et al. (2024a) and Bassal et al. (2024) provide post-earthquake field observations in _ Iskenderun of liquefaction-induced effects on 26 buildings and seven liquefaction-induced lateral spreading transects in _ Iskenderun.
Post-earthquake reconnaissance following the moment magnitude (M w ) 7.8 Kahramanmaras x earthquake provided a remarkable opportunity to develop numerous high-quality case histories of liquefaction-induced damage in _ Iskenderun, Tu¨rkiye. Welldocumented field case histories are crucial for developing and validating methods for assessing liquefaction triggering and its consequences. This study utilized the cone penetration test (CPT) as it is the preferred method for subsurface characterization of liquefaction case histories involving primarily sand materials due to its nearly continuous profiling, accuracy, and repeatability.
This study presents CPT data collected in March 2024. The CPT profiles characterize the subsurface conditions at liquefaction-induced lateral spread transects (LS) and building settlement sites. Through this effort, 40 CPT, 7 seismic CPTs (SCPTs), and 3 soil borings were performed. Subsurface conditions are documented at 17 buildings that experienced liquefaction-induced settlements from less than 1 to 74 cm and includes three sites where building-ground interactions resulted in hogging ground deformation (i.e. when building settlement drags the surrounding earth downward) for several meters around the building. In one case, a single-story building was damaged by the ground deformation induced by adjacent taller buildings. The CPT investigations also consider seven liquefaction-induced lateral spread transects with observed accumulated deformations from 20 to 147 cm. In addition, this study provides estimates of earthquake ground motion intensity measures to characterize the seismic loading at these field case histories. CPT-based liquefaction assessment of liquefaction triggering, post-liquefaction volumetric settlement, and lateral displacement are assessed using several procedures currently used in engineering practice to provide insights. The critical layers are characterized at eight level ground sites. All geolocated field testing data reported in this study are available at DesignSafe (Macedo et al., 2025).
The Boulanger and Idriss (2014) CPT-based liquefaction triggering procedure is used to calculate the factor of safety of liquefaction triggering (FS liq ) in this study. A soil behavior type index (I c , Robertson and Wride, 1998) value <2.6 is used to identify soils susceptible to liquefaction. The sensitivity of the results to this threshold value was evaluated by considering a range of I c from 2.5 to 2.7. The subsequent effects of varying I c on the assessment of liquefiable layers, as well as liquefaction indices, are generally <10% as compared to the results using an I c threshold of 2.6. The one-dimensional (1D) free-field post-liquefaction volumetric-induced reconsolidation settlement (S v ) is estimated using Zhang et al. (2002), and the lateral spreading displacement index (LDI) is estimated using Zhang et al. (2004). The liquefaction severity number (LSN) is calculated using van Ballegooy et al.
(2014), the liquefaction demand parameter (L D ) is calculated using Hutabarat and Bray (2022), which can be used to estimate ejecta-induced settlement using Bray and Olaya (2023), and the liquefaction building settlement index (LBS) is calculated using Bray and Macedo (2017).
Tu¨rkiye is located at the junction of the Eurasian, African, Anatolian, and Arabian tectonic plates. The East Anatolian Fault Zone (EAFZ) surrounds a roughly 700 km long strike-slip fault that runs southwest to northeast in southeastern Tu¨rkiye. On February 6, 2023, Tu¨rkiye experienced two major left-lateral earthquake events along the EAFZ followed by numerous aftershocks, which is referred to as the Kahramanmaras x earthquake sequence. The first event had an M w of 7.8 and occurred in the Pazarcık section of the EAFZ. This rupture included surrounding fault segments (e.g. Amanos and Erkenek faults). The second event had an M w of 7.5 and occurred 9 hours after the initial event along the C xardak and Dog˘ans xehir faults within the EAFZ. As shown in Figure 1, the first event occurred along a northeast-southwest alignment, and the second event along a westnortheast alignment. The M w 7.8 and M w 7.5 earthquakes had focal depths of 8.6 and 7.0 km, respectively. Figure 1 of the reclaimed shoreline, which will be discussed later. Moug et al. (2024a), using field reconnaissance, and Taftsoglou et al. (2023), using satellite imagery, noted most liquefaction manifestations from the 2023 Kahramanmaras x earthquake occurred within the reclaimed shoreline area of _ Iskenderun.
Previous investigations in the _ Iskenderun area give some insight into the underlying subsurface conditions. Eryılmaz and Eryılmaz (2003) 2019) or Eryılmaz and Eryılmaz (2003) were located within the reclaimed shoreline where most liquefaction evidence was observed. A micro-zonation study by Denge Mu¨hendislik Ltd. Sti. (2011), included borings to 30 m depth with Standard Penetration Test (SPT), test pits, laboratory tests, and surface seismic surveys, showed the presence of gravels, sands, silts and clays in the upper 30 m; the sand fraction was observed to increase in areas closer to the shoreline. The Denge Mu¨hendislik Ltd. Sti. (2011) report also documented a shallow GWL between 1.4 and 2 m depth in the explorations closest to the _ Iskenderun shoreline.
Figure 2a shows the seismic strong motion stations in the vicinity of _ Iskenderun and the 2023 rupture of the East Anatolian Fault provided by Wang et al. (2023). The closest stations to _ Iskenderun are TK-3112; TK-3115 and TK-3116, as labeled in Figure 2b. Figure 3a presents the shear wave velocity (V s ) profiles for these stations, with timeaveraged shear wave velocity values over the top 30 m (V s30 ) reported as 233, 424, and 870 m/s, respectively (METU Earthquake Engineering Research Center).
Seismic station TK-3112 is located 2 km west of the C xay District of _ Iskenderun (Figure 2b). Of the nearby seismic stations, TK-3112 is at a site most representative of the subsurface conditions in the study areas. A SCPT was advanced at TK-3112 to measure V s ; its data are shown in comparison to those from METU EERC in Figure 3a. Figure 3b illustrates the 5%-damped acceleration response spectra of the East-West horizontal component of the recorded motions from the M w 7.8 mainshock at these stations. Acceleration, velocity, and displacement time histories are also shown. The recording at TK-3112 stopped early during the M w 7.8 event, likely before the strongest shaking, so response spectra are not provided for this station.
The V s profile at station TK-3116 (with values exceeding 800 m/s) indicates these records can be considered rock motions. The recordings at the TK-3115 and TK-3116 stations show peak ground acceleration (PGA) values ranging from 0.15 g to 0.25 g and cumulative absolute velocity (CAV) values ranging from 1.5 gÁs to 3.0 gÁs, which are significant. For context, the highest CAV values recorded during the 2010-2011 Canterbury earthquake sequence was below 3.0 gÁs (Bray and Macedo, 2017). The peak ground velocity (PGV) values from the M w 7.8 event are also significant (i.e. 35-50 cm/s).
Figure 4 shows CPT data near the station TK-3112; there is a thin layer of very stiff material to 1.2 m depth but given that it is thin and close to the surface, it is likely construction debris. A stratified silty-clayey layer, with a normalized friction ratio (F r ) that fluctuates between 1% and 10%, extends to a depth of 2.8 m. This is followed by a medium dense sand, with a normalized cone tip resistance (Q tn ) (Robertson, 2009) of 130 to 150, that extends to a depth of about 4 m. A stiff layer (Q tn ø 400; F r around 0.5%) of gravelly sand to sand is encountered from 4 to 5 m, after which loose to medium dense sand and silty sand (Q tn between 50 and 100) extend down to about 14.5 m. From 14.5 m until the end of the sounding at 30 m, clay (I c .3) with Q tn <5 and F r ;2.5% exists. The GWL is estimated to be at a depth of 1.5 m at the time of the earthquake, as discussed later.
The V s measurements (shown in Figure 3a and 4) indicate V s <200 m/s above 25 m in depth, below which the V s begins to increase to about 250 m/s. The V s values measured in this study are about 20% lower than the METU values (Figure 3a). The V s30 based on the SCPT-measured V s profile in this study is 186 m/s, instead of 233 m/s as reported by METU based on data collected using ReMi and MASW tests.
Ground motion records from Buckreis et al. (2024) were used to estimate intensity measures for the M w 7.8 earthquake at the study sites. The BSSA14 Ground Motion Model (GMM, Boore et al., 2014) and the algorithm proposed by Jayaram and Baker (2010) were employed to estimate the within-event residuals of the M w 7.8 Kahramanmaras x earthquake. An empirical semi-variogram was estimated based on these within-event residuals and fitted using an exponential variogram model. This fitted variogram model was used to interpolate the within-event residuals. The interpolated within-event residuals, combined with the between-event residuals and the median intensity measures estimated by the BSSA14 GMM, were used to generate maps of intensity measures. This approach, known as kriging, is consistent with the methodologies used by Jayaram and Baker (2009) and Bradley and Hughes (2012).
The interpolated maps of intensity measures and their residuals are presented in Figure 5 for PGA, Spectral Acceleration at 1 second (Sa(1s)), and CAV, respectively. The PGA, Sa(1s), and CAV values at the study sites were extracted from these maps. The study sites are grouped into five areas for building settlement investigations: Areas 1 through 5 and seven lateral spread transects: LS1 through LS7, as shown in Figure 6. Considering representative V s30 values of 226, 201, 193, 193, 193, and 185 for Areas 1, 2, 3, 4, 5 and seismic station TK 3112, respectively, and V s30 values from Okay and O ¨zacar (2023) for other locations. The estimated median PGA, Sa(1s), and CAV values (with 16% and 84% values) are reported in Table 1. These values are used for triggering liquefaction assessments and settlement estimations in later sections of this paper. Table 1. Earthquake ground motion parameters at key areas (median and 16-84% percentiles in brackets).
Post-earthquake reconnaissance (e.g. Cetin et al., 2023;Moug et al., 2023) found most of the liquefaction-induced damage was located within the reclaimed areas along the current shoreline (Figure 6). Significant volumes of ejecta were observed covering Atatu¨rk Boulevard (Figure 7a), which is the main roadway near the shoreline and generally aligns with the historic shoreline and boundary of reclaimed land; little ejecta was observed south of Atatu¨rk Boulevard, except for the C xay District where land reclamation extended further inland. Moug et al. (2024a) and Bassal et al. (2024) reported ranges of liquefaction-induced settlements from less than 1 to 74 cm and lateral spreading up to 147 cm within the reclaimed areas along the shoreline. Settlement was captured using terrestrial lidar scans and laserlevel hand surveys. Moug et al. (2024a) observed the building-ground interaction mechanism of hogging in many of these areas. Lateral spread displacements, which mostly occurred along the seawall within the reclaimed areas (e.g. Figure 7b), were measured by Bassal et al. (2024) via direct measurement of crack widths. Frequent floods affected the shoreline area during high tides and storm events in the months after the earthquakes, which have been attributed to liquefaction-induced free-field settlement and lateral spreading, damage to flood protection systems, and ongoing regional subsidence (e.g. Bassal et al., 2024;O ¨ztu¨rk et al., 2024).
The 47 CPT soundings, seven of which were SCPT, were conducted per ASTM D5778-07 using a 10 cm 2 A.P. van den Berg cone and a truck-mounted CPT rig. Only the results from the soundings relevant to the selected study areas are presented in this paper. In addition, three soil borings were performed for soil sample collection for use in future laboratory tests. The CPT investigations considered the five building settlement areas and seven lateral spread transects shown in Figure 6. At times, the locations of the CPTs had to be moved away from buildings or lateral spreads due to underground utilities and access restrictions. Most of the CPTs had to be pre-drilled to depths of 1.5 to 4.5 m due to the presence of the upper gravelly fills in the investigation areas. The difficulty in advancing the CPT and to drill through the gravelly fills indicated they were dense. SPTs by O ¨zener et al. ( 2024) indicate a median overburden corrected blow count N 1 ð Þ 60 of 30 in the gravelly fill after estimating energy by comparing SPT blow counts in sand with the CPT data collected in this study. The CPTs generally extended to a depth of 30 m below the ground surface, while a few CPTs extended deeper to a maximum depth of 43.5 m; 1,509 m of CPT soundings were performed in this investigation.
The depth of the GWL was measured in the pre-drilled holes. Porewater dissipation tests were also performed at selected intervals in most soundings to cross-check the groundwater levels and to obtain porewater dissipation data. The GWL was measured to be close to 1 m in most areas, but considering post-earthquake ground settlement, the preevent GWL depth is estimated to be about 1.5 m, which aligns with values in Denge Mu¨hendislik Ltd. Sti. (2011). This value is used in subsequent analyses.
The subsurface conditions along the _ Iskenderun shoreline are similar with some important variations. Figure 8 provides representative CPTs from several areas highlighted in Figure 6 to display the characteristics of CPT readings along the _ Iskenderun shoreline area. For instance, Area 1 (CPT29), LS1 (CPT15), and LS4 (CPT47) are all within the reclaimed shoreline but are in the east, west, and center of the study area, respectively. TK-3112 (SCPT63), Area 3 (SCPT39), and Area 4 (CPT42) represent the ground characteristics within the historic land boundary and are in the west, center, and center of the study area, respectively. As shown in Figure 8, CPTs conducted within the reclaimed land typically required significant pre-drilling to bypass gravelly fill that extended to a depth of 3 m on average. CPTs conducted outside of the reclaimed shoreline (i.e. within the historic land boundary) did not require significant pre-drilling.
Iskenderun indicates a predominantly sandy layer, identified by I c values between 1.3 and 2.2, which is fairly consistent within the reclaimed land but noticeably denser in those areas within the historic land boundary (e.g. Area 3, SMS 3112, and Area 4). Within the eastern portion of reclaimed land (including Areas 1 and 2, and LS 5, 6, 7) the sand layer is underlain by an interbedded zone of sandy silts to clayey silts which ultimately transitions toward more homogeneous silty clay with some thin seams of sand and silt. To the west within the reclaimed land (i.e. LS 1, 2, 3 and 4), the upper sand layer becomes thicker, extending to approximately 12 m, and the interbedded zone becomes sandier, in some cases extending up to 18 m in depth. CPTs soundings within the historic land boundary generally indicate denser sandy silt to sand in the upper 3 m, followed by sand to silty sand to a depth of 15 m, which is typically underlain by silts and clays.
Area 1 is located within the reclaimed shoreline in the C xay District of northeast _ Iskenderun, as shown in Figure 6. It includes four buildings: Buildings G, H, I and J. The layout of the buildings and CPT explorations in this area is shown in Figure 9, with CPT profiles shown in Figure 10. The building plans indicate the four 6-story buildings are constructed from reinforced concrete (RC) frames with infill walls (RCF-IW) and feature a salon-type first floor (having an open ceiling to the second floor on one side). The buildings are depicted as having 2.5 m deep basements based on their architectural drawings (as shown in Figure 11). This is a common construction method in _ Iskenderun for buildings of this size (Moug et al., 2024a). The building foundations consist of two design types. Buildings J and G apparently have a 50-cm thick RC mat overlain with 1 m deep intersecting beams that form a grid, wherein the cells are filled with compacted earth and a concrete cap slab. In contrast, Buildings I and H apparently have only 50 cm slabs with no crossbeams.
In Area 1, 14 CPTs and 1 SCPT were performed. Material presumed to be fill due to the presence of debris and cobbles was encountered in the upper 3 to 4.5 m. This upper material could not be penetrated with the CPT due to its density and large particles and objects; therefore, pre-drilling was performed to depths ranging between 3 and 4.5 m in this area. As shown in Figure 11, this corresponds to the base of the foundation indicating that the subsequent estimations of LBS, S v , LSN, and L D , as outlined in the Liquefaction Assessment Procedures Section, are unaffected by the pre-drilling. Figure 10 shows that the fill is underlain by a predominately sandy layer that extends to about a depth of 10 m.
The sand unit is characterized by Q tn between 75 and 120 and an average I c of 1.8. Beneath the sand unit is an interbedded sandy silt and silty clay unit exhibiting lower Q tn (<35) and a fluctuating I c between 2.4 and 2.9, which transitions into a clay unit below 14.5 m. V s increases from about 200 m/s to 250 m/s over the depth of the sounding (Figure 10). Figure 11a and 11b show these stratigraphic units considering the North-South (Figure 11a) and East-West (Figure 11b) directions. The CPT profiles in this area are remarkably uniform; therefore, variable liquefiable unit thickness was not considered for the liquefaction assessments.
Simplified liquefaction assessments were carried out using the procedures described previously. Based on the subsurface conditions, the upper sand unit from 3 to 10 m is liquefiable (Figures 10 and 11). In addition, there are thinner layers of potentially liquefiable material in the looser interbedded zone from 10 to 15 m. As discussed previously, the preevent GWL depth is estimated to be about 1.5 m in the analyses. As shown in Figure 9 the observed building settlements range from 47 to 71 cm, which are largely from the shearinduced component of liquefaction-induced building settlement (Bray and Macedo, 2017). LBS is used as an index of performance of the shear-induced building settlement component. For the ejecta-induced building settlement component, LSN and LD are used as indices of ejecta potential. The free-field ground surrounding the buildings as well as the buildings also settled due to 1D post-liquefaction volumetric strain (S v ). The range of estimated S v values shown in Figure 10
Area 2 contains three buildings in the C xay District which are about 75 m west of Area 1. It is also located within the reclaimed shoreline. The group of buildings consists of two RCF-IW buildings (K and M) with a 1-story steel-frame building (L) on a concrete slab foundation between them (Figure 12). Building plans for Buildings K and M depict similar construction as Area 1 with salon-type first floors. Building M is 6 stories with a partial seventh floor as well as a 2.5-m deep basement on a 30-cm thick slab. Building K is a 5story structure with a 2-m deep partial basement that does not extend across the full building footprint.
Four CPTs and one SCPT were performed along Atatu¨rk Boulevard at the corners of buildings M and K, and one additional CPT was located at the rear corner of building M, as shown in Figure 12. Prior to the CPT soundings, 3 m of pre-drilling was performed to bypass the upper dense gravelly fill. The subsurface conditions shown in Figure 13 are like those of Area 1. The sand unit beneath the fill has similar Q tn (75 to 120) and I c (1.8) values as those of Area 1. The interbedded sandy silt and silty clay unit below the sand has fewer layers of liquefiable material compared to Area 1 due to its higher I c values; its Q tn (<15) is lower than the sand layer above it. The deeper clay unit is like the same unit in Area 1 and is characterized by Q tn <15 and I c ø 3. V s fluctuates between 180 and 220 m/s over the depth of the sounding. A geotechnical section across buildings M, L (not shown), and K (Figure 14) indicates the subsurface conditions are uniform.
The range of estimated S v values depicted in Figure 13 are between 16 and 22 cm. LBS, LSN, and L D have 16% to 84% ranges of 23 to 31, 27 to 30, and 55 to 110, respectively. These values are consistent with those estimated for these indices in Area 1. The observed settlements of 38 cm and 51 cm of the taller buildings K and M, respectively, in Area 2 (Moug et al., 2024a) are lower than those observed in Area 1 (i.e. 47 to 71 cm). The lower building settlement could be due in part to the underlying interbedded sandy silt and silty clay unit that exhibits a more fine-grained response (higher average I c ) as compared to the same unit in Area 1.
Area 3 is in the Yenisehir neighborhood, which is about 1 km west of the C xay District (Areas 1 and 2) and is on the historic shoreline (i.e. CPTs may or may not be on reclaimed land). It contains two 6-story RCF-IW buildings (V and T) that are adjacent to the 2-story middle Building U (Figure 15). Plans were not available for these buildings. It is assumed that Buildings V and T are similar in construction to the Area 1 and 2 buildings described previously. Based on field observations, Building V has a basement and Building T does not have a basement.
Three CPTs and one SCPT were performed at this location: two near Atatu¨rk Boulevard to the north of Building U and two to the south behind Building U (Figure 15). Northern CPTs required 3 m of pre-drilling, consistent with Areas 1 and 2, indicating they penetrate reclaimed land; however, the southern CPT did not require pre-drilling and the SCPT required only 1.5 m of pre-drilling. A dense sand unit (Q tn ø 100 to 300; I c ;1.3 to 2) is present near the ground surface, extending to 3 m, which then becomes less dense (Q tn <100; I c ;2) and continues to loosen (Q tn ;50) to a depth of roughly 15 m. Below 15 m, the response gradually transitions from silty sand to clayey silt materials to a depth of 18 m, which is characterized by Q tn <10 and I c ;2.9, after which the deeper clay unit exists with Q tn <5 and I c of 3.2. This soil profile differs from those in the C xay District (i.e. Areas 1 and 2), which had a more notable interbedded soil unit starting at a depth of 10 m. The GWL was measured at a depth of about 1 m. Figure 16 shows the measured CPT profiles as well as the resulting FS liq , which highlights the greater thickness of liquefiable material in Area 3 compared to Areas 1 and 2. The V s ranges between 150 and 200 m/s for the upper 15 m of material, beneath which it gradually increases from 200 m/s to 250 m/s at a depth of 30 m.
Buildings V and T have similar amounts of building settlement (Figure 15) as observed for the taller buildings in Areas 1 and 2 (Moug et al., 2024a). However, the range of estimated S v (24 to 32 cm) from the CPTs in Area 3 is higher than in the other areas, and the 16% to 84% ranges of LBS and LSN of 37 to 44 and 32 to 38, respectively, in Area 3 are also higher than in the other areas. The L D values of 28 to 65 are lower than those in Areas 1 and 2 but they also indicate severe ejecta with an estimated localized free-field ejectainduced settlement of 3 to 20 cm. The generally higher indices for Area 3 are largely due to the greater thickness of the upper sand unit, as shown in Figure 16, extending to 15 m as compared to 10 m depth in Areas 1 and 2. The observed building settlements in all three areas are similarly large, indicating similarly poor seismic performance of these buildings.
Area 4 is directly south of the Buildings in Area 3, as shown in Figure 6 and 17a, and lies within the historic shoreline (i.e. not on reclaimed land). Liquefaction settlement was observed at a single 5-story building with no basement on a 50-cm thick slab. There is an empty lot on the northwest side of Building W and an adjacent building on its southeast side. Building W in Area 4 is notable as it settled only 6 cm, which is significantly less than buildings of similar size in Areas 1, 2, and 3. Area 5 consists of six buildings (S, R, Q, P, O, and N) located along Atatu¨rk Blvd., between Areas 3 and 2, as shown in Figure 6 and 17b. Building plans were unavailable for Area 5. S, Q, and N are 7-story structures, while Buildings O and R are smaller, with 6 and 5 floors, respectively. Building P is a historic 2-story structure (Moug et al., 2024a).
Two CPTs were performed to characterize the subsurface conditions in Area 4 as shown in Figure 17a: one on the southwest corner and the other offset about 15 m from the southeast corner due to access. Both CPTs required 1.5 m of pre-drilling, which is consistent with the reduced thickness of fill seen in other areas outside of the reclaimed shoreline. Four CPTs were performed to characterize the subsurface conditions in Area 5 as shown in Figure 17b. CPTs 43, 44, and 46 were pre-drilled to a depth of 1.5 m, while CPT-12 required 3 m of pre-drilling. The CPT responses shown in Figures 18 and 19 are consistent with those of Area 3; however, with a slightly reduced thickness in the liquefiable sandy layer. The liquefiable sandy layer extends to roughly 15.5 m in Area 3, while it ends at around 14 m in Areas 4 and 5.
The values of S v (23 and 27 cm), LBS (34 and 39), and LSN (28 and 34) calculated for the two CPTs in Area 4 are relatively consistent with those from the areas discussed previously. Despite these results, the observed settlement experienced by Building W in Area 4 is significantly lower (6 cm), which is an important aspect that warrants further study. The values of L D (42 and 88) for Area 4 indicate severe ejecta potential, which is consistent with the other areas. The overall ranges of S v (25 to 32 cm), LBS (40 to 62), LSN (33 to 53), and L D (60 to 90) in Area 5 are more similar to Area 3 in that they are higher than those in Areas 1, 2, and 4. However, the buildings in Area 5 experienced a wider range of settlements (between 7 and 37 cm). Interestingly, Buildings Q, O, and N experienced significantly less settlement than Buildings S and R, despite CPT44 (which is the CPT closest to Buildings Q, O, and N) indicating larger S v , LBS, and LSN values. This discrepancy could be due to various factors, including potential design differences (e.g. foundations and basements), which are currently unknown, but should be investigated in future studies.
Subsurface characterization of lateral spread sites Bassal et al. (2024) describe the observations of lateral spreading captured from postearthquake surveys. Seven lateral spread transects were measured in three areas of _ Iskenderun, as shown in Figure 20. The location of the transects is also depicted in Figure 6. The seven lateral spread transects are in the reclaimed land along the current shoreline, therefore, the stratigraphy is likely represented by fill overlying alluvial deposits. All but three (i.e. LS5, LS6, and LS7) are directly behind the sea wall.
The subsurface conditions at LS1 differ from those of the other lateral spread transects (Figure 21). After 4.5 m of pre-drilling, the CPT indicates a mostly sand unit alternating with several silt layers (i.e. with I c .2.6) extending to a depth of 18 m, which is characterized by Q tn ranging from 25 to 100. This unit is underlain by a predominately clay unit characterized by Q tn <5 and I c of 3.4. Despite these differences, the LDI values at LS1 (;420 cm) are similar to the other LS sites along the seawall, which range between 330 and 480 cm.
The LS2, LS3, and LS4 groups are within a park behind the seawall, just northeast of building settlement Area 3. Due to their proximity, LS2 and LS3 were characterized jointly by three CPTs (Figure 22), and LS4 was characterized by two CPTs and one SCPT (Figure 23). Both sets show an upper sand unit extending to a depth of 12 m, characterized by Q tn of 70 and I c ;1.9. The CPT sounding transitions into a sandy silt unit with a significantly lower Q tn of 20. This is underlain by a predominately clay unit characterized by a Q tn of 4 and I c of 3.2. The resulting LDI values for the CPTs representing LS2 and LS3 (Figure 22) show consistent results within 360 to 400 cm.
In the case of LS4 (Figure 23), there is a larger variation in LDI values between CPTs of 330 cm for CPT48 to over 450 cm for CPT47. As shown in Figure 20, these CPTs had greater spacing with CPT48 being 25 m behind the most inland ground crack that was observed. This may suggest a difference in subsurface behavior that controlled the extent of lateral spreading. Despite this, the LDI results agree with estimates from other LS sites along the seawall.
The LS5, LS6, and LS7 group is within the C xay district, with LS6 and LS7 being within building settlement Areas 1 and 2. LS5 is 100 m west of LS6, but the sites show consistency in their subsurface conditions. These CPT profiles are also consistent with the nearby CPTs in Areas 1 and 2. The subsurface conditions of LS7 are captured well by the crosssection shown in Figure 11 and CPT profiles shown in Figure 10, which crosses through Buildings I and H. Due to the uniformity of the CPTs in Area 2, LS5 and LS6 can be described using a common set of CPTs which span a similar distance from north to south and are located between the two lateral spreads, as shown in Figure 24. The resulting LDI (;230 cm) shown in Figure 24 is reduced from estimates of those LS sites along the seawall. This contributed to the reduced thickness of the liquefiable sandy layer. As noted by Bassal et al. (2024) the mechanisms of extensional ground cracking at the LS 5,6, and 7 transects are complex as they were measured between rows of buildings that settled. The ground cracks may have been induced entirely or partially by ground settlement rather than seaward spreading.
The _ Iskenderun level ground case histories are well-constrained by the seismic demand of the M w 7.8 Kahramanmaras x earthquake utilizing ground motion recordings, subsurface data through several CPTs, and observations of ground failure (i.e. ejecta). The critical layer for each level ground site case history is selected utilizing the approach of Dhakal et al. (2020), which is similar to that of Green et al. (2014). The critical layer is the soil layer that would have most likely liquefied early during earthquake shaking and produced manifestations of liquefaction if it had occurred. Thick sand to silty sand units likely liquefied at these sites in _ Iskenderun as indicated in Figure 10, 13, 16, 18, and 19. The weakest, shallowest soil layer within these soil units which was sufficiently thick to produce ejecta is selected as the critical layer at these sites. Profiles of Q tn , I c , FS liq , and S v (e.g. see The key properties of the critical layers selected in this study are summarized in Table 2. For areas where there was a transition in the subsurface conditions within the area, as indicated by the CPT profiles, two sites are used. In Area 3, CPT10 indicated a homogeneous sandy layer from 3 to 11 m. However, CPT39 to the south of CPT10 had a 1.5-m thick clay layer at a depth of 5 m. Based on these different profiles and the transition profiles depicted by intermediate CPTs 38 and 40, Area 3 was divided into sites 3A (CPT10) and 3B (CPT39), with each site having different critical layers. For Area 5, the eastern most CPT (44) lacks a dense upper sand layer (Q tn ø 200) that is present in the three western CPTs (46,12,and 43). Thus Area 5 is split into sites 5aA (CPTs 46, 12 and 43) and 5B (CPT44). The information in Table 2 can be used in future efforts such as assessing liquefaction triggering procedures, site response analyses, and effective stress soil-structure interaction analyses at the building sites discussed in this study.
The 2023 Kahramanmaras x earthquake caused extensive damage in the port city of _ Iskenderun due to liquefaction. A comprehensive field investigation using 40 CPTs and 7 SCPTs was conducted to characterize the subsurface conditions at several building settlement and lateral spread sites and a seismic station surveyed after the earthquake. The site investigations and subsequent liquefaction assessments reveal the reclaimed shoreline area has a consistent subsurface profile with a dense gravelly fill layer overlying a thick layer of liquefiable sand, followed by interbedded silt and sand, which is followed by a thick clay deposit. Recorded ground motions and a kriging approach provided the basis for estimating intensity measures such as PGA, CAV, and Sa(1s) that are used for liquefaction assessments. By documenting these subsurface conditions and providing estimates of ground motion intensity measures, this study provides the missing components of the Moug et al. (2024a) and Bassal et al. (2024) studies that captured ground and building deformation data to realize high-quality liquefaction field case histories in _ Iskenderun.
The reclaimed shoreline area, which exhibited the most significant liquefaction effects, is characterized by a thick deposit of saturated medium dense sand and silty sand layers prone to liquefaction. It is not surprising these sandy layers liquefied during the 2023 Kahramanmaras x earthquake. However, there were significant differences in the liquefaction-induced building settlements in the areas studied, ranging from less than 1 to 74 cm. The large range of settlements highlights the importance of adjacent building interactions during earthquakes and its potential impact on ground deformations. Hence, these field case histories of liquefaction-induced building settlement provide the opportunity to evaluate and improve liquefaction assessment procedures. The well-documented lateral spreading field case histories also provide important data to refine empirical procedures to estimate lateral spread ground displacements and to develop new analytical procedures to model displacement patterns. Finally, critical layer interpretations and liquefaction manifestations are summarized for updating liquefaction triggering procedures. located field-testing data reported in this study are available at DesignSafe DOI: https://doi.org/ 10.17603/ds2-6473-fs88 (Macedo et al., 2025).