Significant and widespread liquefaction occurred in İskenderun due to the 2023 moment magnitude (Mw) 7.8 mainshock of the Kahramanmaraş earthquake sequence. Liquefaction and building damage were concentrated in the Çay District and in other areas of reclaimed land and near the historic shoreline. Liquefaction-induced building settlements from 0 to 740 mm were measured in İskenderun by the Geotechnical Extreme Events Reconnaissance (GEER) team.
Building settlement affected the patterns of ground deformation observed adjacent to many structures. Often liquefaction-induced building settlement appeared to depress the ground surrounding the building downwards over several meters in a convex deformation pattern called hogging. Sagging was sometimes observed over wide buildings supported on shallow foundations. This paper summarizes observations of liquefaction-induced building and ground settlements in İskenderun, including liquefaction effects on building and interactions between nearby buildings. Before discussing these observations, the geologic setting of İskenderun and seismic demand from key events of the 2023 Kahramanmaraş earthquake sequence are discussed.
The areas of İskenderun that were significantly affected by liquefaction are composed of Quaternary alluvium from coastal and fluvial processes, and fill for reclaimed land (Brehme et al., 2011). Figure 1 provides an overview of the areas of İskenderun addressed in this paper, including the areas of reclaimed land (shown by the hatched pattern). Based on three geophysical surveys conducted in the İskenderun region by Ozdemir et al. (2019), the thickness of Quaternary alluvium is estimated to be approximately 35-40 m. Beneath this alluvium, rock formations, including claystone, sandstone, and limestone, are encountered. Soil descriptions and standard penetration test (SPT) blow counts (N) from three borings in İskenderun are illustrated in Figure 2; the boring locations are included in Figure 1a. The profiles suggest that very loose to loose silty sand with some gravel underlies İskenderun to depths of at least 30 m.
The SPT blow counts, which are uncorrected for hammer energy because this information is not available, ranged from 2 to 12 in the sand and are consistent with the very loose to loose description of the soil. Given the borings are not located on reclaimed land, the soils in the profiles are likely Quaternary alluvium deposits. The water level in these three borings is between 1.5 and 3 m below the ground surface, suggesting a shallow water table in the areas where major liquefaction was observed. İskenderun is a historic port city that was developed primarily in the 20 th century. The land around İskenderun is described in historical accounts as "swampy" and poorly drained; development of the city included large efforts to drain and fill the low-lying swampy areas (Nalça 2018). Development in the 20 th century also included building out the shoreline with reclaimed land. The areas of reclaimed land in Figure 1a are interpreted from shorelines in historical maps dated before and in 1916; the historical maps are compiled in Nalça (2018).
Atatürk Boulevard generally marks the boundary between the original land and reclaimed land.
An example of the historic shoreline and current Atatürk Boulevard is shown in Figure 3. The same building (Building P in Figure 1c), formerly a community center and currently a civil registry, is pictured pre-1916 in Figure 3a and in May 2020 in Figure 3b. Pre-1916 the building was located along the shoreline; it is now located along Atatürk Boulevard and is 150 m from the current shoreline. Satellite imagery of İskenderun in 1969 presented in Taftsoglou et al. (2023) shows that some of the present-day reclaimed land was built out from Atatürk Boulevard at the time, including part of the Çay District. The historical maps and 1969 image suggest that the Çay District fill was placed between 1932 and 1969. İskenderun has been historically subjected to extensive damage and liquefaction from earthquake events, as documented in Ambraseys (1989). The August 13, 1822 earthquake, called the Aafrine or Aleppo earthquake, had an estimated Mw 7.5. Shaking from this earthquake event destroyed several houses and caused liquefaction along the coast. There are also accounts that groundwater levels rose above low lying ground for an extended period, turning cultivated land into marshy areas. These water level changes could be due to liquefaction subsidence or potentially tectonic subsidence. Strong shaking and damage in İskenderun is also reported from the April 3, 1872 Mw 7.2 Amik Gölü earthquake (Ambraseys, 1989).
The Kahramanmaraş earthquake sequence had two primary events: a Mw 7.8 mainshock that occurred on February 6, 2023 at 4:17 am local time, and a Mw 7.5 aftershock that occurred nine hours later. The Hatay region and İskenderun were also affected by a Mw 6.3 aftershock that occurred on February 20, 2023. over the top 30 m (Vs30) of 233 m/s, 424 m/s, and 870 m/s for the TK-3112, TK3115, and TK-3116 stations, respectively. (a) (b) Figure 4. (a) Map showing strong motion stations within 25 km (red) and 50 km (black) of İskenderun; (b) shear wave velocity (Vs) profiles for stations TK-3112, TK-3115, and TK-3116 (source: METU EERC).
Figure 5 shows the 5%-damped acceleration response spectra of the East-West horizontal component of the recorded motions of the Mw 7.8 mainshock and Mw 6.3 aftershock at the TK-3112, TK-3115, and TK-3116 stations. The acceleration, velocity, and displacement time histories are also presented. TK-3112 had an early stoppage when recording the Mw 7.8 event;
the stoppage appears to have occurred before the strongest shaking; hence, response spectra are not shown for this station. The Vs profile of station TK-3116 (Vs values higher than 800 m/s) suggests that the recorded motions can be considered as rock-like motions. There is a noticeable amplification of the acceleration response spectra of TK-3115 and TK-3112 when compared to that of TK-3116 for the Mw 6.3 aftershock, suggesting seismic site response effects in İskenderun and the Çay District. Some intensity measures for the three events are summarized in Table 1; given the recording stoppage at TK-3112, intensity measures are not available from TK-3112 for the Mw 7.8 and Mw 7.7 events. However, the horizontal peak ground acceleration (PGA) in İskenderun can be estimated using the ground motion model presented in Buckreis et al. (2024, this collection) for the mainshock. The borehole profiles in Figure 2 indicate that there are loose silty sands to at least 30 m depth in areas of İskenderun, suggesting that the Vs30 throughout north-central İskenderun is closest to that of TK-3112 (233 m/s) than the other strong motion stations. Assuming Vs30 = 233 m/s and the event-corrected ground motion model (i.e., Kriging interpolation on within-event residuals), the horizontal PGA in the surveyed areas of İskenderun would be near 0.32 g, which is notably higher than the PGA values of 0.17g at TK-the north-south and east-west ground motion components). This level of amplification is consistent with the amplification observed in the Mw 6.3 aftershock. The cumulative absolute velocity (CAV) is known as an efficient intensity measure to assess liquefaction-induced damage (Bray and Macedo, 2017;Bullock et al., 2018). The recordings from the TK-3115 and TK-3116 stations show CAVs in the range of 1.5-3.0 g•s, which are significant. For context, the highest CAV during the Canterbury earthquake sequence was lower than 3.0 g•s (Bray and Macedo, 2017). With potential amplification effects, the CAVs in the liquefied areas of İskenderun could be even higher than 3.0 g•s; however, ground motions in some areas may have been damped after liquefaction was triggered. The peak ground velocity (PGV) values from the Mw 7.8 are significant (and higher compared to the other events), suggesting potential near-fault region effects. The mean period (Tm) follows an expected trend, i.e., it increases as Vs30 decreases; the estimated values also might be affected by potential near-fault effects.
Recorded velocity pulses in the Mw 7.8 event suggest forward directivity effects in the region, which may have exacerbated the liquefaction consequences in İskenderun. However, further investigation is needed into the near-fault and directivity effects at the İskenderun sites to understand the ground motion effects on these case histories. Ground motions and intensity measures within the areas of İskenderun impacted by liquefaction will be studied further using site investigation data and site response analysis to estimate and refine seismic demand at the impacted sites while accounting for local site conditions. However, these initial estimates and ground motion recordings provide context for the seismic demand that affected the areas investigated by the GEER team.
İSKENDERUN Evidence of liquefaction, including large amounts of ejecta and lateral spreading, was observed in İskenderun following the 2023 Kahramanmaraş earthquake, as documented by Turkish-US GEER and Earthquake Engineering Research Institute (EERI) teams (e.g., Çetin et al. 2023, Moug et al. 2023). This evidence of liquefaction is attributed to the closer, larger magnitude of the February 6, 2023 Mw 7.8 mainshock within the Kahramanmaraş sequence. A detailed description of liquefaction manifestations, lateral spreading, and regional subsidence in İskenderun is presented in Bassal et al. (2024, this collection). Figure 1a summarizes the evidence of liquefaction in the Çay District and north-central İskenderun observed by GEER teams in February and March 2023. Green triangle symbols in Figure 1a represent locations where liquefaction ejecta were observed by GEER teams, in photos uploaded to SiteEye (2023), or by residents. White circles represent locations where no liquefaction ejecta was observed. These locations of ejecta and no ejecta are provided as the locations captured by the photographs as opposed to the geotag of the camera location. The ejecta observations are limited to February 2023 since it was assumed that much of the ejecta was either degraded by flooding, rain, or wind or cleaned up before March 2023. Reconnaissance also confirmed that major liquefaction occurred at the İskenderun port (Çetin et al. 2023), located east of the area captured in Figure 1.
Lateral spreading was surveyed along the İskenderun shoreline through measured transects.
The locations of these transects are included in Figure 1a. Details on the magnitude and extent of lateral spreading in İskenderun are provided by Bassal et al. (2024, this collection). The three transects in the Çay District revealed extensional ground cracks that extended from the shoreline to areas with buildings. It is possible that these ground cracks were indications of large-scale lateral spreading capable of influencing the observed building settlement. The building settlements reported in this study do not differentiate between different settlement mechanisms (e.g., shear, volumetric, ejecta, and lateral displacement), as there are insufficient data to be able to do so at this time. Follow-up studies may be able to collect sufficient data to understand better the contribution of different liquefaction-induced building settlement mechanisms (e.g., Bray and Macedo 2017).
Zones of "major liquefaction," "possible or marginal liquefaction," and "no evidence of liquefaction" are shown on Figure 1a. The zones are interpreted in areas where observations were made by GEER teams or SiteEye teams; this study does not provide an assessment of whether liquefaction did or did not occur outside of these zones. Representative GPS tracks of the GEER teams are shown in Figure 1a to provide the context of where the ground reconnaissance was performed in İskenderun. The zones of no evidence of liquefaction are areas where the teams deliberately looked for evidence of liquefaction (e.g., ejecta, settlement, rolled curbs), and none were observed. These zones include the southwestern part of Figure 1a, away from the shoreline. The zones of possible or marginal liquefaction are those where building settlement less than 30 mm was observed, and there was no surficial evidence of liquefaction. A possible or marginal liquefaction area is the underpass that runs approximately east-to-west through İskenderun. Flooding and sand in the underpass were evident in aerial photos on Google Earth in the days following the earthquake, which was confirmed by a local taxicab driver. However, the only possible surface evidence of liquefaction was nearby buildings that appeared to have settled less than 30 mm. The flooding and sand may be evidence of subsurface buildup of excess porewater pressures that were relieved through the storm drains at the base of the underpass.
The zones of major liquefaction are associated with land between the historic shoreline and the current shoreline (i.e., reclaimed land) and near the historic shoreline. Liquefaction ejecta were observed south of the Çay District, about 200 m from the historic shoreline, and further from the historic shoreline than other observations of liquefaction in İskenderun. This occurrence could be associated with historic infill of low-lying swampy areas, historic depositional environment, or some other factor.
Building settlements were predominately estimated through laser-level hand surveys and lidar scans at select buildings. The surveyed buildings in the paper are referred to by reference letters, as shown in Figures 1b and 1c, and by their location latitude and longitude. Laser-level hand surveys were performed by measuring settlement relative to a reference datum that was judged to have been minimally affected by earthquake shaking and liquefaction. An example of a laser-level hand survey at Building E (36.5902N 36.1777E) on Bahçeli Sahil Evler Street is shown in Figure 6a. Lidar scans used a terrestrial 3D scanner (FARO Focus3D X 330 2014) to capture point clouds of the ground and buildings at 2 to 3 mm resolution. An example of a lidar scan being performed at Building K (36.5906N 36.1780E) on Atatürk Boulevard is shown in Figure 6b. Lidar scans at multiple locations around a building were registered together to create a three-dimensional (3D) model of a building and the surrounding ground. The lidar data were processed and registered in FARO Scene software (version 2023.0.1) and analyzed in the CloudCompare software (CloudCompare 2023). The analysis included estimating an original reference plane of the ground surface (i.e., the pre-earthquake ground surface) from ground points that were judged to be minimally affected by liquefaction and earthquake shaking and then calculating the vertical distance from the reference plane to the surveyed ground surface to create a ground settlement map. Building settlement was estimated from elevations at the bottom of the building (assumed to be in contact with the ground pre-earthquake) to the reference plane. A challenge of lidar analysis is shadowed zones where data points were not obtained. These data shadows primarily occurred near buildings where the ground was deformed downward by the building settling into foundation soils. In these cases of data shadows near buildings, the building settlements were estimated by referencing known heights or elevations of building features, including façade tile or balconies. Buildings surveyed in the Çay District had settlements ranging from 0 mm to 740 mm. The significant amount of ejecta observed in the Çay District indicates these are liquefactioninduced building settlements. Figure 1b shows the surveyed buildings and their average settlement. They were largely reinforced concrete frame with infill walls (RCF-IW) structures.
Sixteen buildings were surveyed in the Çay District. Twelve of the buildings were five-or sixstory ("mid-rise") buildings with courtyards and side yards between each building, three were high-rise structures over eight stories high, and one had a single story. Settlements at the five and six-story buildings ranged from 240 mm to 740 mm, with all but one settling over 300 mm.
The three high-rise buildings (Buildings X, Y and Z; 36.5904N 36.1762E), which were located adjacent to one another, had no discernable settlement (i.e., 0 mm of settlement). These buildings are assumed to be on pile foundations, given the height of the buildings founded on reclaimed land and their good foundation performance during the earthquake. The single-story building (Building L; 36.4806N 36.1778E) appeared to settle differentially due to influences from nearby buildings, with an average settlement of 220 mm; this building is presented with additional details below.
Building plans from two five-story buildings in the Çay District are shown Figure 7. These two building plans represent typical foundation types for mid-rise buildings in the Çay District. (a) (b) Figure 7. Building plans for two buildings in the Cay District: (a) cross-section view of Building D (36.5902N 36.1776E) on Bahçeli Sahil Evler Street showing the mat foundation; (b) plan view of foundation at Building C (36.5902N 36.1774E) on Bahçeli Sahil Evler Street showing the strip footings that orthogonally intersect.
Nine buildings were surveyed in the Yenişehir neighborhood, with average settlements ranging from 0 mm to 520 mm (Figure 1c). The largest building settlements were measured near the historic shoreline and then decreased with distance inland from the historic shoreline.
For instance, settlements up to 520 mm were measured at buildings surveyed along Atatürk Boulevard. Settlements less than 100 mm were apparent at Building P (36.5916N 36.1702E, pictured in Figure 3), which had an average 70 mm of settlement, and Building W (36.5925N 36.1660E) on Mareşal Fevzi Çakmak Street, one block south of Atatürk Boulevard, which had an average settlement of 60 mm (discussed below). As noted above, the areas of minor or marginal liquefaction in Figure 1a include those where < 30 mm of building settlement were observed, further supporting that liquefaction-induced building settlements decreased inland from the historic shoreline.
Building B on Bahçeli Sahil Evler Street (36.5902N 36.1772E) was surveyed with laser level hand survey, lidar scans, and photogrammetry that was subsequently combined with lidar data. The street view of the building is shown in Figure 9a. Building B is a six-story residential RCF-IW structure. There was no apparent basement, however, there was a sub-ground level room for the electrical panel. A summary of the building settlement estimates by hand survey and lidar scan data is provided in Figure 9b. The average building settlement by lidar scan analysis is 400 mm. There are discrepancies between the hand survey and lidar analysis estimates of settlement at the west side of this building. This likely reflects difficulties in Lidar scans were performed at the four corners of Building B. A reference ground plane that slopes 1 o from north (rear of the property) to south (front of the property) was fit to the ground elevation data; this plane represents the assumed pre-earthquake ground elevation. The ground contours in Figure 9c represent the difference between the scanned ground elevation and the reference plane (i.e., contours of ground settlement due to the earthquake event). The ground deformation settlement around the building extends over 2 m around the building footprint. As evident in Figure 9c, there were data shadows in the lidar scans, particularly around the base of the building. At the building's northeast corner, the ground adjacent to the building was largely shadowed, not allowing the ground elevation to be analyzed in this area.
Combining the lidar data with photogrammetry can fill in these data shadows. Photogrammetry was performed by stitching many high-resolution photos obtained from the scanner into a 3D
model. The result of combining the lidar and 3D photogrammetry is shown in Figure 9d,
showing that photogrammetry was able to compensate for several of the areas shadowed in the lidar data, including the ground near the east side of the building.
Building E on Bahçeli Sahil Evler Street (36.5902N 36.1777E) provided an example of the influence of the building foundation type on building-ground interactions. Building E is a fivestory residential RCF-IW structure; the building's street view is shown in Figure 10a. The building owner reported that the structure has a 2.8 m deep basement with a 40 cm thick reinforced concrete mat foundation that extends out 1.5 m around the footprint of the building.
The average building settlement was estimated to be 740 mm from combined laser-level hand surveys and lidar scan data. Figure 10b shows the effect of the building foundation on ground deformations, where a depression that coincided with the foundation mat was observed around the building. The settlement at this building was the largest of the buildings surveyed in the Çay District. The foundation type appeared to be unique among the buildings surveyed in the Çay District: few buildings had full basements without pile foundations, and there were no other similar ground deformation patterns that could be attributed to a foundation mat that extended a similar distance around the building footprint. The large liquefaction-induced settlements at Building E likely indicate poor ground conditions. Buildings on pile foundations generally performed well, with little liquefaction-induced settlement observed. In addition to the three high-rise buildings with no noticeable settlement described in the previous section, Building F on Bahçeli Sahil Evler Street (36.5903N 36.1781E; shown in Figure 11a) settled less than the other buildings without deep foundations surveyed in the Çay District. The building owner reported that the foundation has a full basement with concrete piles. A settlement of 240 mm was estimated at Building F from a laser-level hand survey and from a lidar scan at the northwest corner. Building settlement also deformed the ground around the building; Figure 11b shows the rear courtyard (north side of the building), where there is an apparent slope in the ground from the partition wall towards the building. The ground pavers around the partition wall in Figure 11b show additional ground settlement, which may be attributed to poor fill compaction close to the wall, ground extension from nearby lateral spreading, ground extension from building settlement-induced ground deformations, or some combination of these factors. The contoured lidar ground data at the rear courtyard in Figure 11c further show these building settlement-induced ground deformations, where lidar data indicate ground settlement extends over 2 m from the building.
The ground to the west of Building F and on the right of Figure 11c had a depression that did not appear to be related to liquefaction settlement, and was likely due to erosion during flooding or poor fill compaction around the basement.
Documented building-ground interactions in the Çay District and İskenderun indicate a distinct pattern of ground deformations induced by building settlement. It appeared that building settlement dragged down adjacent ground over a length of several meters from the building edges, with the ground settlement decreasing with increasing distance from the building. This resulted in convex ground surfaces near the buildings as depicted in Figure 12 and described herein as a "hogging" pattern of ground deformation. The hogging ground deformation was observed in the ground surrounding buildings that settled significantly, which differs from hogging due to differential building settlement with the exterior of a foundation settling more than the center. Most of the buildings in the Çay District and İskenderun had stiff foundations that appeared to settle without significant internal distortion. Free-field observations did not provide evidence of ground settlement or heave that would be an alternate explanation for the observed hogging ground deformation surrounding buildings that settled significantly. In addition to the building settlement associated with ground hogging, some punching settlement was observed where the building settled into the ground immediately adjacent to the building. However, in general, the punching settlement was notably less than the building settlement associated with ground hogging settlement throughout the Çay District and İskenderun. Three representative cases of hogging building-ground interaction were documented in detail. These three cases are for hogging deformation between two buildings and are distinguished by the spacing between buildings and how nearby buildings were affected by the building-induced ground deformation. The observed cases had access between the buildings, and the buildings were spaced far enough apart to allow unobstructed observations of the building-ground interactions. It is likely that similar ground deformations and building-ground interactions occurred throughout liquefied areas in İskenderun, including for the building cases previously described; however, building spacing, building collapse, features like fences and landscaping, and available time did not allow observations of similar building-ground interactions between two buildings.
Hogging ground deformation was observed at Building U (36.5930N 36.1657E) on Atatürk Boulevard in the Yenişehir district. The hogging ground deformation developed between two six-story RCF-IW structures: Building T (36.5930N 36.1661E; east of Building U in Figure 13a) and Building V (36.5932N 36.1655E; west of Building U in Figure 13c). Building U sits between these buildings in the background of Figure 13b. Building T settled an average of 450 mm, and Building V settled an average of 500 mm; at the building corners along Atatürk Boulevard and closest to Building U, Building T settled 420 mm, and Building V settled 450 mm. There was no discernible building settlement at Building U. All settlements at this site were performed with hand laser-level measurements. Building U was built before 1946, although the construction date is not known at this time. Building T was measured to be 16.8 m in width (i.e., parallel to Atatürk Boulevard), estimated to be 22.5 m in length (i.e., orthogonal to Atatürk Boulevard) by Google Earth, and 18.7 m in height. Building V was measured to be 36.4 m in width, estimated to be 25.3 m in length by Google Earth, and 21.8 m in height. The total distance between Buildings T and V is 28.8 m. The manager of Building U reported that Building T does not have a basement, and it is also likely Building V does not have a basement. The settlement of Buildings T and V appeared to drag down the surrounding ground in the same pattern as shown in Figure 12. There was a diminishing amount of depressed ground surface extending out from the buildings until it was no longer noticeable. Because there were contributions from both buildings on either side of Building U, the hogging ground deformation pattern in Figure 12 is horizontally mirrored. The central horizontal segment between the convex ground deformations, and where there was no apparent ground settlement, was about 9 m long. This central area appears to be free-field conditions, which is also supported by observations of negligible settlement and damage at Building U. Therefore, the zone of influence over which the hogging deformations extended is about 9.9 m from both Buildings T and V. Given the observations of widespread liquefaction, it is likely that the entire area also settled. However, this interpretation assumes that the area around these buildings is equally affected by this widespread settlement, which does not significantly affect these localized building-ground interactions.
Hogging ground deformation was also observed in the courtyard between two sets of two similar buildings in the Çay District. Figure 14 shows the courtyard between the buildings with the hogging ground settlement pattern between the buildings; the convex warped ground is also apparent through the spreading of the ground tiles and floodwater present only around the base of the buildings (photographs were taken on March 29, 2023 when there was significant flooding in the Çay District and İskenderun). The plan view of the set of four buildings, including dimensions and estimated settlements, is shown in Figure 15.
The two west structures, Building J (36.5907N 36.4789E) located on Atatürk Boulevard, and Building G (36.5905N 36.1790E) located on Bahçeli Sahil Evler Street, appear to have similar design, age, and construction. These two structures were built prior to September 2007, according to Google Earth imagery. Similarly, the two east structures, Building I (36.5908N 36.1791E) located on Atatürk Boulevard, and Building H (36.5906N 36.1792E) located on Bahçeli Sahil Evler Street, also appear to have similar design, age, and construction. These two structures were built between September 2007 and 2010 based on
Google Earth images. All the structures appear to be RCF-IW structures, six stories high with a half-seventh story. There were no apparent basements for the buildings; however, there were sub-ground level utility rooms on the east side of Buildings I and H and on the west side of Buildings J and G. Settlement at the corners of the buildings was estimated from a combination of hand laserlevel measurements, lidar scan data, and flood water depth during the March 29, 2023 flood event. The flood water was used to estimate settlements by measuring flood water depth from the building base and then referencing that depth to a location where settlement could be measured by hand survey or lidar. The flood water approach assumed consistent water surface elevation during measurements and was only used at locations that could not be surveyed otherwise. Settlement estimates were generally consistent between lidar and laser-level approaches, with differences being less than 40 mm or 10%. Average settlements were significant at all buildings, with 560 mm, 710 mm, 660 mm, and 470 mm estimated for Buildings H, G, J, and I, respectively. The settlements of Buildings H and I were similar, as were the settlements of Buildings G and J; however, the settlements of the pairs of buildings differed. It is unknown if the difference in the pair of building settlements is due to differing ground conditions, different design and construction details, or influence from adjacent buildings (discussed further below).
Lidar scans performed on March 29 and April 1, 2023 captured ground settlement between the four buildings, which displayed a hogging ground deformation pattern from all four buildings. Figure 16 provides contours of vertical distance from a horizontal reference plane.
The original ground elevation is estimated from the elevations at the sides of the courtyard, not directly between the buildings because the lidar data show the sides of the courtyard are at a higher elevation than the areas between the buildings. The settlement contours indicate the least ground settlement occurred in the middle of the courtyard between the buildings, and ground settlement increased towards the buildings. Six profiles of ground surface deformation in the courtyard between the four buildings were extracted from the lidar data and plotted in Figure 17; the locations of these profiles are shown by the red lines in Figure 16a. The ground settlement of these six profiles between the buildings is relative to the interpreted initial ground elevation. The ground deformations show distinct warping of the ground from individual buildings, consistent with the interpreted pattern in Figure 12. The average building settlements measured on the courtyard side of the buildings are plotted as the triangle symbols for Buildings H and I and the circle symbols for Buildings G and J in Figure 17. As plotted in Figure 17, the building settlement measurements fall into a similar pattern as the ground deformation between the buildings, with some variation due to measurement variability and other potential settlement mechanisms (e.g., punching shear failure, global seaward lateral spreading); however, the supporting evidence suggests the ground deformation patterns of the ground adjacent to the buildings relate strongly to building settlements. Color scale represents the ground distance from the initial reference plan (in m). Note that data shadows prevent lidar data from extending to the base of the buildings in most cases.
Figure 17. Hogging ground deformation pattern between the four buildings with measured settlements at other sites where hogging ground deformation was observed. Settlement of the K, L, and M buildings are also shown.
Elevation differences between the middle of the courtyard and the reference planes indicate that free-field conditions did not exist between the buildings, and there were likely interactions between the ground deformations induced by the settlements of each building. The area between Buildings G and J is 60 mm lower than at the west edge of the courtyard (where it is assumed that the ground did not settle), and the area between the Buildings H and I is 130 mm lower than at the east edge of the courtyard (where it is also assumed that the ground did not settle). A schematic of the interpreted building-ground interaction between these buildings is illustrated in Figure 18, including the overlapping of the zones of influence between the buildings. There were likely similar interactions between adjacent buildings (e.g., between Buildings J and I and between Buildings H and G) where there is about 6 m separation between buildings. For instance, Figure 19 shows ground deformations and bending to the west of Building G, where there is an empty lot to its west. (36.5904N 36.1790E;15 FEB 2023).
The influence of settlement at adjacent buildings can possibly explain observed differential ground settlements or varying performance between similar buildings. Although the buildings to the east and west of the group of four buildings were not surveyed, some discussion of their influence is merited. There was an empty lot to the east of Building I; the settlement measured on the east side of Building I is 100 mm to 140 mm less than the west side, where the settlement of Building J may have dragged down the west side and the east side was not affected by an adjacent building. There was a building to the east of Building H that collapsed after the earthquake event. Settlement of Building H is relatively uniform from the east to west sides of the building, possibly attributable to buildings to both the east and west influencing ground deformations in addition to settlement at Building H itself. However, there was no building to the west of Building G, and there are no distinct differential settlements at Building G. This is possibly due to large settlements from Building G relative to Building H obscuring influence from the adjacent building. There is a building to the west of Building J settlements on the west side of Building J were 50 to 80 mm larger than on the east side. If settlement at the adjacent building was significantly larger than at Building J, this may account for the differential settlements.
These observations indicate that adjacent building settlements possibly affected settlement at some buildings; however, further investigation is warranted to examine this issue comprehensively. Building L Restaurant, Çay District A three-building group of Building K (36.5906N 36.1780E), Building L (36.5905N 36.1778E; a restaurant), and Building M (36.5905N 36.1777E) on Atatürk Boulevard in the Çay District were not taken as it was unsafe to do so. The measured settlements at Building L and the adjacent buildings are plotted with the diamond symbols in Figure 17; note that the error bars on the symbol for settlement at Building K represent uncertainty in the distance between it and Building L. The convex point of the ground deformation is assumed to be the middle of the distance between Buildings K and M.
The pattern of ground deformation is similar to the deformations measured in the group of four buildings discussed previously. The consistency of deformation suggests a characteristic pattern for building-ground interactions in this area. Further investigation is required to understand the role of the buildings and subsurface conditions on these ground deformation patterns. The consistency of measured ground deformation at the Building L restaurant site suggests that the presence of the Building L structure had a minimal influence on the ground deformation between the two heavier buildings.
In contrast to hogging ground deformation, sagging ground deformation is defined by a concave ground curvature. Sagging across a building results in larger settlement beneath the middle of a building relative to the ends of the building.
Minor sagging was observed over the footprint of Building W (36.5925N 36.1660E) on Mareşal Fevzi Çakmak Street, which had an average of 60 mm of settlement. The five-story RCF-IW structure is 26.55 m wide and located one block south of Atatürk Boulevard. The middle of the structure had about 10 mm more settlement than both corners. The street view of Building W is pictured in Figure 21a; the building settlement relative to the sidewalk in front of the building is pictured in Figure 21b.
Sagging building-ground interactions were also observed across Buildings S, R and Q, as presented previously in Figure 8. Although sagging was not measured for any of these buildings individually, the overall settlement pattern showed larger settlements for the buildings and building corners closest to the center of the city block than for the building corners at the edges of the city block.
Liquefaction-induced building settlements in İskenderun were strongly associated with areas of reclaimed land and areas near the historic shoreline with a wide range of settlements measured. Settlements were particularly large in the Çay District, which is entirely founded on reclaimed land. Significant building settlements were documented along the historic shoreline (i.e., Atatürk Boulevard); however, overall building settlements were generally lower than in the Çay District, particularly for five and six-story buildings. While there were observations of ejecta along Atatürk Boulevard outside of the Çay District, ejecta occurrence and amounts were significantly less than those observed in the Çay District, indicating liquefaction was not as prominent along the historic shoreline as in areas of reclaimed land. Historical accounts of the city note that the shoreline was extended beyond Atatürk Boulevard after 1916; satellite imagery and historical maps indicate that reclaimed land in the Çay District was built between 1932 and 1969. Reconnaissance in areas inland from the historic shoreline yielded evidence of moderate or marginal liquefaction or no liquefaction. However, there is some evidence of liquefaction that occurred at depth, including underpass flooding and small building settlements, which did not produce surface evidence of liquefaction.
In the areas of major liquefaction, rigid body tilt of buildings with robust foundations at liquefied sites was observed, as well as one case of a sagging of building foundations for a structure with a more flexible foundation, which is consistent with observations made after other earthquakes (e.g., Bray & Sancio 2009, Bray et al. 2014). Three examples of distinct building-ground interaction were observed. In these cases, the ground appeared to be dragged down by building settlements within a zone of influence around the building; the maximum observed zone of influence was 9.9 m wide. This ground deformation pattern is described as hogging. In one case, a hogging deformation pattern distorted a one-story building (i.e., the Building L restaurant) between two larger nearby buildings.
The ground deformations in the Çay District (group of four buildings and the Building L restaurant) had similar relationships between vertical deformation and distance from the building, as shown in Figure 16. However, the Building T and V settlements and ground deformations do not strictly follow this pattern. The settlements at Buildings T and V were similar to the Çay District mid-rise buildings with the smallest settlements (i.e., Building K and I), however, the distance from Buildings T and V over which the hogging was observed (9.9 m) was larger than at Building K (6.7 m) and Building I (6.25 m). Possible explanations for the inconsistency include: (a) Buildings T and V are spaced considerably further apart than the buildings where hogging ground deformation was observed in the Çay district, which led to no building-ground-building interaction at Building U, (b) different soil conditions at these sites, and (c) different structural systems and building weights. Further data gathering and research are required to understand the manifestations of hogging ground deformation at liquefaction sites in İskenderun. Understanding the building-ground and building-groundbuilding interactions is important to identify the conditions that may cause building settlement to induce ground settlement that damages adjacent structures, such as in the Building L restaurant case described in this paper. This is an important consideration in the evaluation of nearby and adjacent buildings with shallow foundations in liquefiable ground.
Significant liquefaction-induced building settlements were observed in İskenderun, with the most severe liquefaction impacts to buildings occurring in areas of reclaimed land and near historic shorelines. There were cases with negligible settlement (i.e., 0 mm), minor settlement (i.e., < 100 mm), and large settlement (greater than 100 mm and up to 740 mm) documented in these areas. Building-ground interactions were observed where liquefaction-induced building settlement deformed the ground in a notable zone of influence around the building.
The observed ground deformation was a dragging down of the ground in a convex pattern called hogging. Three representative cases of this hogging ground deformation between buildings were documented. In one case, free-field conditions existed between the zones of influence from the two buildings; in another case, the zones of influence of building settlement overlapped; and in the third case, the building-induced hogging ground deformations damaged a lightweight building between two heavier buildings.