1 Introduction

On February 6, 2023, two earthquakes, with moment magnitudes M7.8 and M7.6, occurred in Kahramanmaraş-Türkiye on the East Anatolian Fault zone, at local times 04:17 and 13:24, respectively. After the mainshocks, more than ten thousand earthquakes were recorded in the period of February 6 to March 1, within 200 km radii from the event epicenters. Over 400 of these aftershocks have M ≥ 5.0. Figure 1 shows the focal mechanisms and spatial distribution of mainshocks and aftershocks along with their induced fault rupture patterns. The epicenter of the first event, which has a focal depth of 8.6 km, is located at 37.288°N and 37.043°E (AFAD), close to Kahramanmaraş-Pazarcık. The second event has a focal depth of 7.0 km with its epicenter in Kahramanmaraş-Elbistan-Ekinözü, at 38.089°N, 37.239°E.

Fig. 1
figure 1

The spatial distribution of mainshocks and aftershocks (as of March 1, 2023) along with the induced fault rupture patterns of Türkiye-Kahramanmaraş-Pazarcık and Ekinözü-Elbistan Earthquake sequence

Full size image

The moment tensor solution suggested purely left-lateral strike-slip during both events. The fault rupture of the first event was triggered on Narlı Fault, at the northern end of the Dead Sea Fault zone. The rupture stepped over to the East Anatolian Fault zone, and continued along Pazarcık, Erkenek, and Amanos segments, propagating bilaterally in the north- east and south-west directions. The total fault rupture length exceeded 300 km, with a maximum offset of 4 m (Cetin et al. 2023b, a; Cetin and Ilgac 2023). After the Ekinözü-Elbistan event, a 160 km long fault rupture is mapped with a maximum offset of 6 m. The second event was initiated on Çardak Fault and propagated along Doğanşehir Fault Zone.

During the first event, the maximum peak ground acceleration (PGA) levels were recorded in Hatay-Antakya, where most of the damage was reported. The PGA levels at SGMS # 3126 were recorded as 1.23 g and 1.04 g in the north–south (NS) and east-east (EW) directions, respectively. Similarly, 244 SGMS recorded the second event. The maximum PGA was recorded as 0.65 and 0.53 g in the NS and EW directions, respectively, at Kahramanmaraş-Göksun Station 4612. The Ekinözü-Elbistan Mw = 7.6 earthquake affected the northern provinces of Kahramanmaraş, Malatya, Adıyaman, and Kayseri, the most.

After the earthquakes, reconnaissance field investigations were performed to collect perishable data and document the extent of damage to the natural and built environment. As part of these reconnaissance studies, the sites with surface manifestations of seismic soil liquefaction in the form of soil ejecta, excessive foundation and ground deformations were identified and documented. The deformations were mapped and samples from ejecta were retrieved. Within the confines of this manuscript, the liquefied soil sites are introduced first. The discussions are then followed by the presentation of the results of soil classification test, which are performed on the retrieved soil ejecta at liquefied sites. The soil classification laboratory test results are comparatively shown on the plasticity chart summarizing the Wang (1979) soil ejecta database. Similarly, liquefaction susceptibility chart recommendation by Seed et al. (2003) is used to comparatively present and assess the resulting soil ejecta database.

2 Liquefied soil sites

Reconnaissance studies to identify liquefied soil sites started on the third day after the earthquakes and continued for longer than 2 months. Various research groups contributed to the reconnaissance studies, and the resulting information is documented on a digital platform named SiteEye (Saha Gözü, in Turkish), which is accessible at www.sahagozu.com. A total of 428 liquefaction case history sites were identified as reported in various documents including but not limited to Cetin and Ilgac (2023), Cetin et al. (2023a), Cetin et al. (2023b), and Moug et al. (2023). A map showing the locations of liquefied sites is presented in Fig. 2.

Fig. 2
figure 2

Map of liquefaction case history sites

Full size image

The liquefaction case history sites were grouped in 6 geographical regions, namely L1 through L6, as presented in the same figure. A higher resolution map of these liquefaction regions is given in Figs. 3, 4 and 5. For each liquefaction region, a set of sample pictures documenting the surface manifestations of liquefaction triggering is provided in the same figures.

Fig. 3
figure 3

a) A higher resolution map of liquefaction regions L1 and L2, b) sample pictures from L1 region, and c) sample pictures from L2 region

Full size image
Fig. 4
figure 4

a) A higher resolution map of liquefaction regions L3 and L4, b) sample pictures from L3 region, and c) sample pictures from L4 region

Full size image
Fig. 5
figure 5

a) A higher resolution map of liquefaction regions L5 and L6, b) sample pictures from L5 region, and c) sample pictures from L6 region

Full size image

A sample summary of these liquefaction sites is presented in Table 1, whereas a complete presentation of them is provided in the electronic supplements. As part of the summary tables information regarding the coordinates of the site, and type of liquefaction surface manifestations in the form of soil ejecta (SE), lateral spreading (LS), excessive ground settlement (EGS), excessive foundation displacements (EFD) is provided.

Table 1 A sample summary of coordinates and liquefaction manifestation types of liquefaction sites
Full size table

3 Characteristics of liquefied soil ejecta

A total of 81 samples were retrieved from seismic soil liquefaction sites, shown in Fig. 6. The regions where these samples were retrieved are labeled as A through F. These liquefied soil ejecta samples were tested at Middle East Technical University (METU) Soil Mechanics laboratory to assess their grain size and distribution characteristics along with their consistency limits.

Fig. 6
figure 6

The locations of sites where ejecta samples were collected

Full size image

4 Sampling and testing procedures

Consistent with the extent of the area shaken by the earthquake sequence, samples were retrieved from also a wide range of locations, including but not limited to beaches, fishermen’s wharf, ports, factories, schools, and residential buildings (collapsed and non-collapsed), Hatay airport, bridge abutments, dams, farmlands, etc. Similar to the soil liquefaction sites, the regions where ejecta samples were collected were clustered geographically and labeled A through F. In Fig. 7, sample pictures are presented, which show the ejecta materials retrieved from regions A through F. Each sample is assigned a sample identification number, and their grain size, distribution and consistency limit characteristics are assessed, which are summarized in the next section.

Fig. 7
figure 7

A sample picture showing liquefied soil ejecta retrieved from six regions A through F

Full size image

When retrieving surface soil ejecta from liquefaction sites, special attention was given to collecting representative soil samples. To achieve this, a continuous sampling approach was employed by pushing vertically a tube into the ejecta cone (or volcano). This approach was adopted to eliminate concerns related to possible segregation or layering within the ejecta material. Moreover, efforts were made to trace the travel path of the surface ejecta and identify its origin, thereby confirming that the surface ejecta accurately represented the liquefied soil layer. This confirmation was essential to ensure that the material had not eroded to the ground surface due to excess pore water pressure-induced flow. Where available, existing borelogs were utilized to support this validation process. However, in instances where site-specific borelog information was absent, attributing the characteristics of the ejecta soil sample directly to potential liquefied soil layers became challenging. To address this potential ambiguity, additional site investigations, laboratory testing and liquefaction triggering evaluations were needed for more accurate conclusions, further discussions of which is beyond the scope of this preliminary (reconnaissance) assessment manuscript.

Ejecta samples after labeling were brought to METU Soil Mechanics Laboratory. They were oven dried for 24 h. After drying, a representative amount of sample portion was taken and put in sodium hexametaphosphate solution (5 ml/1000 g) overnight to ensure separation of soil particles. Following that, wet sieving is conducted to separate and collect the finer portion (< 0.074 mm) of the specimen. Fines portions is accumulated in large bottles for further grain size assessments. Coarser portions after the wet sieving are oven dried again, then tested for dry sieving. These abovementioned procedures are performed in conformance with ASTM International (2017) D6913, ASTM International (2016) D7928. The Atterberg limits of the retrieved samples are estimated as per ASTM D4318 (2010). Laboratory test results will be presented next, separately for each ejecta region.

4.1 Region A- Dörtyol, Hatay

Soil ejecta from 8 different locations were collected. 3 of them were sampled from industrial zones, whereas the rest were sampled from a shoreline and a port. The locations of these sites are shown in Fig. 8a. Similarly, Table 2 summarizes the coordinates of these sites along with the color of the ejecta and major soil classification test results. The samples from the shoreline were classified as clean sands with fines percent < 5–6%. The sample taken from the metallurgical facility with ID #40, is classified as non-plastic, and its fines and silt contents are estimated as 27.2 and 25%, respectively. The grain size distribution curves of soil ejecta are comparatively shown in Fig. 8b, along with liquefaction susceptibility bounds of Tschuida (1970). They are observed to consistently fall within the suggested susceptible soils range. A more comprehensive presentation of liquefaction triggering assessments for the soil ejecta sites, accompanied by the documentation of available borelogs and laboratory test results, can be found elsewhere (Cakir and Cetin 2024; Sahin and Cetin 2024). These details will not be reiterated herein.

Fig. 8
figure 8

a) Retrieved samples from Region A, b) the grain size distribution curves of the ejecta

Full size image
Table 2 A summary of the site coordinates and soil classification characteristic of the ejecta samples from Region A
Full size table

4.2 Region B—Çay Neighborhood, İskenderun

In this region, a total of 13 soil ejecta samples were retrieved. 6 samples were collected from ports, while the remaining 7 were collected from the İskenderun-Çay neighborhood. The locations of these sites are shown in Fig. 9a. Coordinates of these sites, color of the soil ejecta as well as major soil classification test results are summarized in Table 3. All samples were classified as non-plastic. The sample retrieved from the Çay neighborhood with ID #38 has fines and silt contents of 18.4 and 15%, respectively. Figure 9b shows that except for the specimen ID #45, which has a gravel content of 45%, all grain size distribution curves of the ejecta material fall within the bounds of “susceptible to liquefaction region” defined by Tsuchida (1970). A more comprehensive presentation of liquefaction triggering assessments for the soil ejecta sites, accompanied by the documentation of available borelogs and laboratory test results, can be found elsewhere (Ozener et al. 2024; Bol et al. 2024). These details will not be reiterated herein.

Fig. 9
figure 9

a) Retrieved samples from Region B, b) the grain size distribution curves of the ejecta

Full size image
Table 3 A summary of site coordinates and soil classification of the samples from Region B
Full size table

4.3 Region C—Hatay Airport, Demirköprü Bridge, Yarseli Dam, Tepehan Village

In this relatively wide region, a total of 17 samples were retrieved: 5 samples were collected from Hatay Airport (#4,6,7,11,13), 6 samples were from a site in the proximity of Demirköprü Bridge, which was severely damaged due to soil liquefaction (#58,59,60,61,62,68), 2 samples were taken from Yarseli Dam site (#51,57), and remaining 4 were sampled from a farmland (#1,2,5,10) in Tepehan village. For the Demirköprü Bridge site, the retrieved non plastic ejecta sample has 25% fines contents, in the average. Ejecta samples collected from the farmland are also non plastic. On the contrary, samples retrieved from Hatay airport have plasticity index values varying in between 21 to 37. In the same region, one of the samples from Yarseli Dam (#57) has 58% fines content with plasticity index of 28. The geographical positions of these sites are shown in Fig. 10a, and grain size distribution of the ejecta are shown in Fig. 10b. Further details, including the coordinates of these sites, the color of the soil ejecta, and the major soil classification test results are summarized in Table 4. A more comprehensive presentation of liquefaction triggering assessments for the soil ejecta sites, accompanied by the documentation of available borelogs and laboratory test results, can be found elsewhere (Cetin et al. 2024a, 2024b; Ocak and Cetin 2024; Bol et al. 2024). These details will not be reiterated herein.

Fig. 10
figure 10

a) Retrieved samples from Region C, b) the grain size distribution curves of the ejecta

Full size image
Table 4 A summary of site coordinates and soil classification of the samples from Region C
Full size table

4.4 Region D—Gaziantep

Four of the samples in this region were retrieved from a farmland, and the remaining two were sampled from the Arıklıkaş Dam site. Figure 11a displays the site locations. The coordinates of these sites, the color of the ejecta, and the main soil classification test results are summarized in Table 5. Only one specimen (ID #49) is categorized as low plasticity clay (CL), while the others are classed as silty sand (SM) or clayey sand (SC). Figure 11b illustrates the grain size distribution curves of the soil ejecta, which fall within the Tsuchida (1970) liquefaction susceptibility bounds.

Fig. 11
figure 11

a) Retrieved samples from Region D, b) the grain size distribution curves of the ejecta

Full size image
Table 5 A summary of site coordinates and soil classification of the samples from Region D
Full size table

4.5 Region E—Gölbaşı, Adıyaman

Due to the high number of residential buildings where foundation failures due to liquefaction induced bearing capacity failures, excessive settlements and lateral spreading, the highest number of specimens (31 samples) were collected from Gölbaşı-Adıyaman as shown in Fig. 12a. All samples were brown in color. Out of 20 samples, for which Atterberg limit tests were performed, 19 of them have plasticity index values varying in the range of 16 to 23%, except for sample #76, which has a plasticity index of 34%. These soil classification test results will be further elaborated as part of susceptibility of fines containing soil mixtures discussions. Figure 12b illustrates the grain size distribution curves of the soil ejecta. Large number of ejecta grain size distribution curves fall outside the liquefaction susceptibility limits defined by Tsuchida (1970) due to their high fines content. Out of 31 specimens, hydrometer tests were performed on 6 to estimate their clay contents. Clay contents of these 6 samples were estimated to vary in the range of 13–20%. Further details, including the coordinates of these sites, the color of the soil ejecta, and the major soil classification test results are summarized in Table 6. A more comprehensive presentation of liquefaction triggering assessments for the soil ejecta sites, accompanied by the documentation of available borelogs and laboratory test results, can be found elsewhere (Cetin et al. 2024c). These details will not be reiterated herein.

Fig. 12
figure 12

a) Retrieved samples from Region E, b) the grain size distribution curves of the ejecta

Full size image
Table 6 A summary of site coordinates and soil classification of the samples from Region E
Full size table

4.6 Region F—Doğanşehir, Malatya

6 samples were collected from this region. Two of them (ID #48 and 67) were sampled from Sultansuyu Dam, located on Sultansuyu River in Malatya. The other 4 ejecta samples (#44,56,79,81) were retrieved along the benches of Goksu Stream, next to a liquefaction-induced retaining wall failure. The geographic positions of these ejecta are presented in Fig. 13a. Table 7 provides a summary of site coordinates, the color of the soil ejecta, and major soil classification test results. Out of 4 samples from the failed retaining wall, 2 of them were classified as non-plastic. The grain size distribution curves of the soil ejecta are comparatively shown in Fig. 13b, which consistently fall within the liquefaction susceptibility bounds of Tsuchida (1970) except for the sample #67, which is retrieved from Sultansuyu Dam site, and classified as silty, clayey gravel (GM or GC). A more comprehensive presentation of liquefaction triggering assessments for the soil ejecta sites, accompanied by the documentation of available borelogs and laboratory test results, can be found elsewhere (Cetin et al. 2024d). These details will not be reiterated herein.

Fig. 13
figure 13

a) Retrieved samples from Region F, b) grain size distribution curves of the ejecta

Full size image
Table 7 A summary of site coordinates and soil classification of the samples from Region F
Full size table

5 Liquefaction susceptibility of fines containing soil mixtures

Laboratory test results, performed on liquefied soil ejecta, are shown on the plasticity chart shown in Fig. 14. The liquefied ejecta samples are mostly classified as low plasticity clays (CL). Most of this CL type soil ejecta are retrieved from Gölbaşı-Adıyaman region. The liquid limits of these samples vary in between 32 and 38%. Their plasticity index values are estimated to be in the range of 16 to 23%. Unexpectedly, two ejecta samples with plasticity indices higher than 30% are retrieved from Hatay airport, one of which is classified as high plasticity clay (CH).

Fig. 14
figure 14

Fine-grained ejecta shown on the plasticity chart

Full size image

Laboratory test results are also comparatively shown on the Seed et al. (2003) liquefaction susceptibility chart, given in Fig. 15. The majority of the ejecta samples fall on “Zone B: Testing Recommended” region of the chart. However, inconsistently, 12 out of 74 samples fall outside the susceptibility region of the Seed et al. chart.

Fig. 15
figure 15

Liquefaction Susceptibility Assessment of Fines Containing Soil Ejecta by Seed et al. (2003)

Full size image

Finally, laboratory test results are jointly presented along with the Wang (1979) data, which were compiled from samples of liquefied “silty soil” layers at 7 sites in China. Figure 16 presents comparatively the overall database on the plasticity chart, which clearly suggests that clayey soils are liquefiable and capable of producing ejecta when subjected to strong cyclic loading. Detailed site investigation and laboratory testing programs are ongoing to investigate the liquefaction (cyclic) responses of fine-grained soils liquefied during Türkiye-Kahramanmaraş earthquake sequence. Their results are believed to contribute to current state of knowledge and practice on assessing cyclic response of fine-grained soils. Until the findings of these studies become available, the liquefaction susceptibility of silty-clayey soils mixtures is recommended to be assessed cautiously and conservatively.

Fig. 16
figure 16

Assessment of liquefiable soil types by Wang (1979) (after Boulanger 2004)

Full size image

6 Summary and conclusions

Seismically induced soil liquefaction was listed as one of the major sources of damage observed in the natural and built environment during the 2023 Kahramanmaraş earthquake sequence. After the earthquakes, reconnaissance field investigations were performed to collect perishable data and document the extent of damage. As part of these reconnaissance studies, the sites with surface manifestations of seismic soil liquefaction in the form of soil ejecta, excessive foundation and ground deformations were identified and documented. The deformations were mapped and samples from ejecta, if available, were retrieved. Within the scope of this manuscript, the liquefied soil sites were introduced. The results of soil classification tests, performed on the retrieved soil ejecta at liquefied sites, were comparatively shown on the liquefaction susceptibility chart of Seed et al. (2003). Additionally, the soil liquefaction database of Wang (1979), composed of fine grained soils, is also used in the comparisons.

The majority of the ejecta samples were classified as sands with varying degrees of fines. Laboratory test results revealed that fines-containing liquefied ejecta samples were mostly classified as low plasticity clays (CL). Most of these CL type soil ejecta samples were retrieved from Gölbaşı-Adıyaman region. The liquid limits of these samples varied in between 32 to 38%. Their plasticity index values were estimated to be in the range of 16 to 23%. Surprisingly, two ejecta samples with plasticity indices higher than 30% were retrieved from Hatay airport, one of which was classified as high plasticity clay (CH). The majority of the fine-grained ejecta samples fall on “Zone B: Testing Recommended” region of the Seed et al. (2003) susceptibility chart. Unexpectedly, 12 out of 74 samples fall outside the susceptible limits of Seed et al. The laboratory testing of soil ejecta retrieved after Türkiye-Kahramanmaraş earthquake sequence prematurely suggested that clayey soils are capable of producing ejecta when subjected to cyclic loading. Detailed site investigation and laboratory testing programs are ongoing to further investigate this rather unexpected response. The results of these ongoing research studies are believed to provide further input to seismically induced liquefaction response of fine-grained soils. Until they become available, the liquefaction susceptibility of silty-clayey soils mixtures is recommended to be assessed conservatively with caution.