1 Introduction

Türkiye is one of the most earthquake-prone countries in the world. On average, a strong earthquake with a magnitude between 6.0 to 7.0 occurs every two years, and a major one with a magnitude of 7.0 or above occurs every three years (Mertol et al. 2021a). Almost all the population lives in areas where earthquakes are a threat. Historically, the country has suffered greatly from earthquakes. Since the establishment of the Turkish Republic in 1923, close to 130,000 people have died in earthquakes. The earthquake-related direct damage is estimated to be around $150 billion (Sahin et al. 2019; Aral and Tunc 2021).

Geologically, Türkiye has numerous major fault lines that have the potential to generate strong earthquakes. The two most important ones are the North Anatolian Fault (NAF) and the East Anatolian Fault (EAF) Zones (Fig. 1). The NAF runs from east to west across northern Türkiye and is about 1,500 km long. It has been responsible for some of the most destructive earthquakes in the country’s history. The EAF follows the shape of an arc from İskenderun Bay to Hakkari in the eastern and southeastern parts of the country. The EAF and the eastern part of the half arch, which is sometimes called the Bitlis-Tatvan-Zagros zone, together define the complete EAF zone. The western part of the country also has individual fault zones that generate strong earthquakes. These fault zones are collectively known as the West Anatolian Fault (WAF) Zone. The WAF encircles an area of about 45,000 square kilometers or about 6% of the country’s total land mass.

Fig. 1
figure 1

Active major fault lines and epicenters of four earthquakes (adapted from MTA 2023)

Full size image

This article investigates the four most recent earthquakes in Türkiye, namely the 2020 Sivrice, the 2020 Aegean Sea, and the dual 2023 Kahramanmaraş earthquakes. Figure 1 shows the epicenters of these four earthquakes. The first, Sivrice-Elazığ, struck the eastern part of Türkiye on January 24, 2020, with a moment magnitude of Mw = 6.8, causing a maximum peak ground acceleration (PGA) of 0.298 g (AFAD 2020a; Cheloni and Akinci 2020; Cetin et al. 2021; Mertol et al. 2021a and 2021b; Bayık et al. 2022; Dedeoğlu et al. 2023). The earthquake’s focal depth was 8.1 km. A total of 41 people died and over 1,600 were injured. The direct damage of the earthquake was around $350 million and nearly 21.5% of the buildings in the earthquake zone either collapsed or experienced major damage (AFAD 2020c, Çağlar et al. 2023). There were many studies conducted on this earthquake, describing its characteristics, causes, and damage to the existing building stock (Isik et al. 2020; Bayrak et al. 2021; Demir et al. 2021; Temür et al. 2021; Yurdakul et al. 2021; Çağlar et al. 2023). The outcome of these studies is compiled by the authors’ own work and their details are discussed in Sect. 3.

The second earthquake, in the Aegean Sea, hit the western part of the country on October 30, 2020, with an Mw of 6.6, resulting a maximum PGA of 0.183 g (Akinci et al. 2021; Chousianitis and Konca 2021; Kiratzi et al. 2022; Onat et al. 2022; Mertol et al. 2023a). The focal depth of this earthquake was 16.5 km, and it caused extensive structural damage to the buildings on the Greek island of Samos and in the İzmir province of Türkiye. In this study, only the structural damage to RC buildings in Türkiye is examined. The earthquake caused direct damage totaling approximately $400 million and resulted in 117 deaths and 32,000 injuries. Nearly 0.5% of the buildings in the earthquake-affected region experienced collapse and major damage. The structural causes of collapse and damage were previously studied by many researchers (Cetin et al. 2022; Ziotopoulou et al. 2022; Gurbuz et al. 2023). In this study, the outcome of previous research is combined with the authors’ site inspections and is provided in detail in Sect. 3 (Mertol et al. 2023a).

The third and fourth earthquakes are the February 6, 2023, dual Kahramanmaraş earthquakes with Mws of 7.7 and 7.6, which struck the eastern part of Türkiye approximately 9 h apart. These dual earthquakes caused extensive damage in a total of 11 provinces, located mostly in the eastern and southeastern parts of the country. The focal depths of these earthquakes were 8.6 and 7.0 km, respectively, and generated maximum PGAs of 1.399 and 0.648 g (Karabacak et al. 2023; Mertol et al. 2023b; Cetin et al. 2024; Hu et al. 2024). The earthquakes caused direct damage of approximately $81.6 billion, destroyed a total of nearly 50,000, and severely damaged 165,000 buildings (Binici et al. 2023 and TPS 2023). The total collapsed and severely damaged buildings constituted nearly 15% of the existing building stock in these 11 provinces. These earthquakes are historically among the deadliest in the region and consequently, have become the subject of many studies (Binici et al. 2023; Ozkula et al. 2023; Ozturk et al. 2023; Avcil et al. 2024; Vuran et al. 2024; Yilmaz et al. 2024). The authors spent almost a week in the damaged areas and inspected as many buildings as they could. The outcomes of the authors’ site inspections are in line with others who visited the sites numerous times. These outcomes are discussed in detail in Sect. 3.

In summary, the four earthquakes that are the subject of this study have been examined by many researchers, including the authors of this manuscript (Mertol et al. 2021a, 2021b, 2023a, 2023b). The authors visited the earthquake-affected sites immediately and spent an extended amount of time assessing structural damage and related issues. As a result, the authors not only witnessed, firsthand, their impacts on RC buildings but also personally observed common structural damage patterns. Although these four earthquakes’ locations vary, the purpose of this study is to determine common issues related to the RC buildings’ design and construction stage and to propose methods and solutions that can diminish the deadly effects of future earthquakes. The scope of this work also involves bringing together as many interpretations as possible so that common issues associated with each type of structural damage can be well established. In order to achieve this goal, the authors will propose a series of recommendations which can be utilized to construct a framework that can eventually diminish these common issues before another deadly earthquake.

In the next sections, the seismicity and tectonic activities of the country will be briefly discussed. Then, the current earthquake hazard map will be presented along with the strong ground motion characteristics of the four earthquakes. A detailed assessment of the hazard map will also be provided using strong ground motion data collected during these earthquakes. Subsequently, the earthquake performance of RC buildings will be examined, from their design to construction stages. Finally, the findings from these evaluations will be used to propose general and specific solutions to commonly seen design and construction issues. It is important to note that the observations made in Sects. 3.2 and 3.3, along with the conclusions and recommendations in Sect. 4, are directly derived from the outcomes of the authors’ site visits. These outcomes are also verified against other researchers’ site observations for the four earthquakes as well as other major earthquake sites in the country (see Sect. 3.3).

2 Seismicity and tectonics

2.1 Earthquake hazard map of Türkiye

Plate movements in Türkiye are the direct result of continental drift that began during the Neotectonic period (roughly 12 million years ago, Fig. 2). These plate movements are the cause of all severe earthquakes, specifically when the larger Arabian plate collides with the smaller Anatolian one in the north-to-south direction (Okay et al. 2010).

Fig. 2
figure 2

Identification of plates and their movement in Türkiye and epicenters of four earthquakes (adapted from Okay et al. 2010)

Full size image

In 1940, the country prepared itself against the deadly results of these plate movements by initiating an earthquake-affected code following the 1939 Erzincan earthquake, with an estimated moment magnitude of 7.8, making it one of the deadliest earthquake in the history of the country (Tunc and Tunc 2022). Following this event, earthquake design codes were revised 9 times in a span of 84 years. Similar to the earthquake codes, earthquake hazard maps were also prepared starting from 1945 and were revised 6 times. The current earthquake design code was prepared based on the historical data assuming return periods of 2475, 475, 72, and 43 years. Figure 3 shows the Earthquake Hazard Map of Türkiye for a return period of 475 years (or a 10% exceedance probability in 50 years) (AFAD 2023a; AFAD is the Turkish Disaster and Emergency Management Authority). As illustrated in Fig. 3, the high risk areas in dark red simply follow the major fault zones of NAF, EAF, and WAF.

Fig. 3
figure 3

Earthquake hazard map for a return period of 475 years and epicenters of four earthquakes (adapted from AFAD 2023a)

Full size image

2.2 Sivrice, Aegean Sea and Kahramanmaraş earthquakes

In this study, four earthquakes are examined in detail. The first one, the Sivrice earthquake, with Mw = 6.8 and a focal depth of 8.1 km, struck the Pütürge segment of the East Anatolian Fault (EAF) Zone on January 24, 2020, at 20:55:11 local time (Table 1). The epicenter was the village of Çevrimtaş located in southern Elazığ, about 37 km away from the city (AFAD 2020a; Mertol et al. 2021a). The earthquake was a left lateral strike-slip event that caused extensive damage in the affected region.

Table 1 Characteristics of four earthquakes (data extracted from AFAD 2020a; Chousianitis and Konca 2021; Bayık et al. 2022; Kiratzi et al. 2022; Mertol et al. 2021a, 2023a, 2023b; Cetin et al. 2024)
Full size table

The second earthquake was the Aegean Sea, with Mw = 6.6. It occurred on October 30, 2020, at 14:51:23 local time, off the coast of Samos Island, Greece (Table 1). The epicenter was located 35 km away from the city of İzmir in the south-west direction (AFAD 2020b; Mertol et al. 2023a). The village of Payamlı, in Seferihisar, a town in İzmir, was the closest Turkish town to the epicenter. This earthquake’s focal depth was 16.5 km and caused significant damage to buildings located both on the Greek island of Samos and in the Bayraklı District of İzmir.

The third and fourth earthquakes were recorded on February 6, 2023, with Mws of 7.7 and 7.6 in Kahramanmaraş province. The first earthquake occurred at 04:17:34 local time and the second at 13:24:47 (Table 1). The epicenter of the first event was Pazarcık while the second was Elbistan, two districts of Kahramanmaraş which are approximately 90 km away from each other. The first and second earthquakes’ focal depths were 8.6 and 7 km, respectively. These earthquakes caused substantial damage to the infrastructure and buildings of 11 provinces (Adana, Adıyaman, Diyarbakır, Elazığ, Gaziantep, Hatay, Kahramanmaraş, Kilis, Malatya, Osmaniye, and Şanlıurfa). In terms of damage and the number of deaths, the dual earthquakes are the costliest and deadliest in the history of Türkiye. The death toll from these two earthquakes is estimated to be over 50,000. Almost 215,000 buildings were completely destroyed or severely damaged (AFAD 2023b; Mertol et al. 2023b).

Table 1 lists the maximum peak ground acceleration (PGA) values of these four earthquakes in g, the station numbers along with their distances to the epicenters, and their durations in seconds in a range from minimum to maximum value. The table also lists the fault types associated with each earthquake and the subsurface ruptures in kilometers to emphasize their impacts. The largest subsurface rupture was obtained from the Pazarcık earthquake as 300 km followed by the Elbistan earthquake with 160 km, and then the Aegean Sea and Sivrice earthquakes with nearly the same value of 30 to 40 km.

2.3 Evaluation of strong ground motion data

In previous research studies, the strong ground motion data was assessed in-depth for these four earthquakes by the same authors (Mertol et al. 2021a, 2023a, 2023b). In this article, only the most critical stations were selected in order to compare and contrast the outcome of these earthquakes. Figure 4a shows the closest ground motion stations to the epicenter of the first earthquake in Sivrice. The maximum PGAs recorded at the Sivrice station, station number 2308, were 0.243 and 0.298 g for the north–south and east–west directions, respectively. Figure 5a illustrates the acceleration values recorded at the same station for these two directions. The acceleration values are also plotted, along with the spectral acceleration curves in these two horizontal directions, based on a 5% constant damping ratio (Fig. 5a). As the data indicates, the PGAs at station 2308 revealed a slight increase, specifically in the range of 1.5 to 2 s. This increase, which is directly linked to the source-to-site effect of the Sivrice earthquake, was more pronounced in the north–south direction. In order to verify the design spectrum curves of the current Turkish Building Earthquake Code (TBEC 2018), the spectral acceleration data was plotted together. The code uses six local soil classes, denoted by letters A through F, where A is used for a soil type of hard rock and F is used for an extremely loose soil type that requires both soil testing and site evaluation. In this study, the first five soil types (A through E) were examined to evaluate the north–south and east–west components of the earthquake. Figure 5a exhibits the design spectrum and spectral acceleration curves plotted in these directions.

Fig. 4
figure 4

Locations of strong ground motion stations for: a Sivrice earthquake, b Aegean Sea earthquake, and c Kahramanmaraş earthquakes with Mw = 7.7 and Mw = 7.6. (adapted from Mertol et al. 2021a, 2023a, 2023b)

Full size image
Fig. 5
figure 5

PGA and spectral acceleration versus design spectrum curves for: a Sivrice earthquake, b Aegean Sea earthquake, and c through f Kahramanmaraş earthquakes with Mw = 7.7 and Mw = 7.6 (adapted from Mertol et al. 2021a, 2023a, 2023b)

Full size image

In the Aegean Sea earthquake, the results from station number 3513, located in Bayraklı, near the İzmir Bay area, were used since the town sustained the most damage (Fig. 4b). Figure 5b illustrates the PGA values of this station in the north–south and east–west directions. Based on the data, the largest recorded PGA for this earthquake was 0.183 g and occurred at station number 0905, located in Kuşadası, a town in İzmir (20 km west of the epicenter, or 100 km south of İzmir). However, the damage and collapse occurred in another town, Bayraklı located further away from Kuşadası or 80 km north-east of the epicenter, since the soil was alluvial and amplified the impact of the earthquake. The maximum horizontal PGAs in Bayraklı were recorded at station number 3513, located near the town. They were measured as 0.108 and 0.097 g in the north–south and east–west directions, respectively. Figure 5b compares the design spectrum curves of TBEC 2018 against the measured spectral acceleration curves. As illustrated by the data, the spectral accelerations exceeded the design values both in the north–south and east–west directions in the case of local soil classes ZA and ZB (hard rock to rock) for periods between 1.0 to 2.0 s. This exceedance was also observed for buildings located in local soil class ZC (very dense soil) with a fundamental period of 1.5 s.

For the dual Kahramanmaraş earthquakes, emphasis was given to the data extracted from the four stations where the largest PGAs were recorded. These stations were station numbers 3125, 3129, and 3135, located in Antakya, Defne, and Arsuz, all towns of Hatay, for the first earthquake (Pazarcık earthquake with Mw = 7.7) and station number 4612, located in Göksun, Kahramanmaraş, for the second event (Elbistan earthquake with Mw = 7.6) (Fig. 4c). Figures 5c through f display the PGAs recorded at these four stations in the north–south and east–west directions. Based on the PGAs of the Mw = 7.7 earthquake, the largest one in the north–south direction was at station number 3129 as 1.379 g (Fig. 5d). The largest in the east–west direction was measured at station number 3135 as 1.399 g (Fig. 5e). The maximum vertical component of this earthquake exceeded the unit value of g and therefore was decided to be cited in this study. The largest vertical value was measured at station number 3125 as 1.174 g (Table 1). For the second earthquake (Mw = 7.6), the largest recorded PGA was at station number 4612 and was measured as 0.648 g. The PGA values were once again plotted using a constant damping ratio of 5% and compared to the design spectrum acceleration values of TBEC 2018. Figures 5c through f display the PGAs and response spectrum curves plotted in both the north–south and east–west directions. As the data indicates, the measured spectral accelerations exceeded the spectrum curves of TBEC 2018. At station numbers 3129 and 3135, this exceedance was a lot more prominent specifically between periods of 0.2 to 1.0 s. At station number 4612, this range of period exceedance shifted to 1.0 to 2.0 s.

In summary, based on the spectral acceleration values plotted in Figs. 5a through f, the measured values were estimated more correctly for the Sivrice and Aegean Sea earthquakes. However, this was not the case for the dual Kahramanmaraş earthquakes since measurements exceeded design values by as much as three times.

2.4 Aftershocks

The aftershocks of the four earthquakes were investigated for a 10-day period following the main events, as illustrated by the data from previous research conducted by the authors (Mertol et al. 2021a, 2021b, 2023b). Figure 6 displays the number of aftershocks recorded daily. In order to understand the characteristics of these aftershocks, they are categorized in terms of their magnitudes, as shown in the same figure. Based on this data, the number of aftershocks for the dual Kahramanmaraş earthquakes was much greater than those for the other two earthquakes. The magnitudes of the Kahramanmaraş aftershocks were also significantly higher than those for the other two events, specifically for magnitudes of 4.0 or more. These high magnitudes are believed to have contributed to the extensive damage that resulted from the dual Kahramanmaraş earthquakes.

Fig. 6
figure 6

Number of aftershocks sorted based on their magnitudes for Sivrice, Aegean Sea, and Kahramanmaraş earthquakes with Mw = 7.7 and Mw = 7.6 (data extracted from Mertol et al. 2021a, 2021b, 2023b)

Full size image

3 Earthquake performance of reinforced concrete buildings

3.1 Earthquake specifications

The design of earthquake-resistant RC buildings has been regulated by earthquake specifications in Türkiye since the 1940s. The first specification was prepared after a major earthquake struck Erzincan, a city in the eastern part of the country, with a magnitude of Mw = 7.8 in 1939. This specification was very similar to the Italian Earthquake Specifications of the period and was revised 9 times up to date.

The four earthquakes of this study and the corresponding earthquake codes of 1975, 1998, 2007, and 2018 are the main subject of this manuscript since the majority of the RC buildings that collapsed or severely damaged were constructed beyond 1975. All of these codes used a percentage of the seismic weight (a combination of dead loads and some part of live loads) to be applied as the earthquake loads on the buildings. Turkish Earthquake Code 1975 (TEC 1975) used other factors such as four earthquake zones, two building importance factors, four soil types, and two ductility levels (ductile or non-ductile). Based on TEC 1975, the calculated total base shear for residential buildings was approximately 8 to 15% of the seismic weight. The minimum specified concrete compressive strength for the buildings having an importance factor of more than 1 in earthquake zones 1 and 2 was 22.5 MPa. However, no limit was specified for concrete compressive strength for the buildings located in earthquake zones 3 and 4. There was also no specified minimum yield strength for reinforcing steel. The requirements related to the detailing of the reinforcement such as 135° bent stirrups and confinement zones at the ends of the beams and columns existed in TEC 1975.

The earthquake specifications were improved in 1998 (TEC 1998) since they included strong column-weak beam criteria. In order to meet this criteria, the bending moment capacity of a connecting column was increased by 20% compared to that of a connecting beam. Other technical details related to building irregularities, mode superposition and time-history analysis methods were also specified in the same earthquake code. In TEC 2007, more information was provided with respect to strengthening methods for existing buildings with different construction materials in addition to other updates for material and soil properties including revised ductility ratios. In the current earthquake code of TBEC 2018, performance based design approach was introduced. The earthquake hazard map was updated to include mapped spectral acceleration values of SS and S1 based on a location’s GPS coordinates. Vertical design acceleration spectrum was added to calculate the vertical effects of earthquakes.

Using a few critical design parameters, a comparison of the last four specifications is given in Table 2. These parameters are the concrete compressive strength, the steel yield strength, the 135-degree-bent required for stirrups, confinement zones for the beams and columns, and earthquake zones and their associated ground acceleration coefficients. According to the data in Table 2, as years passed, the minimum concrete compressive strength increased steadily from 20 to 25 MPa. In the case of the minimum steel yield strength, a minimum value of 420 MPa was assigned following TBEC 2018. In previous codes, 220 MPa of yield strength was permitted. In fact, in TEC 2007 code, 220 MPa of yield strength was even permitted to be used both in stirrups and in slab reinforcements. However, the 135-degree requirement for the ends of the bent stirrups and the confinement zones for beams and columns remained the same.

Table 2 Comparison of last four earthquake specifications in Türkiye (adapted from Mertol et al. 2021a)
Full size table

The evolution of seismic zones and their associated ground acceleration coefficients are also examined based on the earthquake codes of 1975, 1998, 2007, and 2018. As listed in Table 2, until TBEC 2018, there were four seismic zones. In TEC 1998 and 2007, these zones generated the same PGA values of 0.1, 0.2, 0.3, and 0.4 g. In TEC 1975, once again there were four seismic zones, but with lower ground acceleration coefficients of 0.03, 0.06, 0.08, and 0.10 g. In the current earthquake code of TBEC 2018, the concept of the seismic zone was abolished. Instead, mapped spectral acceleration values of SS and S1 were introduced.

The PGA values of the four earthquakes are also calculated based on TEC 1975, 1998, 2007, and TBEC 2018 to further compare them to those of the maximum horizontal on-site measured values (Table 3). According to TEC 1998 and 2007, the same equation was used for PGAs, while TEC 1975 used a different one, leading to much smaller accelerations. In the current code of TBEC 2018, the PGA calculations included new soil related parameters of SS and S1 by incorporating their impacts through six local soil classes (ZA to ZF). As discussed before, in this study, only the first five soil types (ZA through ZE) are examined. As displayed in Table 3, regardless of soil class, the PGAs of the Sivrice earthquake were much larger in TBEC 2018 than those of the previous three codes. However, in the case of the Aegean Sea and Elbistan earthquakes, the PGAs of TBEC 2018 were close to those of TEC 1998 and 2007. For the Pazarcık earthquake, the PGAs of TBEC 2018 were much less than those of TEC 2007.

Table 3 Comparison of PGAs in g from last four earthquake specifications in Türkiye to site measured values
Full size table

The ultimate goal in evaluating PGAs is to determine the proximity of code-based PGAs to site-measured ones. All the earthquake codes (TEC 1975) generated much safer (or conservative) values when the PGAs of the Sivrice and Aegean Sea earthquakes were considered. However, this was not the case for the Pazarcık and Elbistan earthquakes, since their measured values well exceeded the code-based ones by as much as 7.5 times when a local soil class of ZA was considered in TBEC 2018. The underestimated PGAs are believed to be one of the primary sources of the extensive damage that occurred in the dual Kahramanmaraş earthquakes. It is therefore recommended to revise all mapped spectral accelerations of the current TBEC 2018 specifically at locations near major fault lines.

3.2 Damage summary of four earthquakes

Table 4 lists the direct costs, for buildings and infrastructure, of the four earthquakes. The data for the dual Kahramanmaraş earthquakes were listed together due to their time proximity (9 h apart), which made it impossible to assess them individually. In the same table, the number of deaths, injuries, damaged, and undamaged buildings that were investigated are listed.

Table 4 Cost and damage summary of four earthquakes (data extracted from AFAD 2020c; Binici et al. 2023; Çağlar et al. 2023; Mertol et al. 2023a and 2023b; TPS 2023)
Full size table

The Sivrice earthquake caused direct damage of approximately $350 million with 41 deaths and 1,607 injuries. As the data in Table 4 indicates, total or partially collapsed buildings constituted 1.2% of the total buildings in the earthquake-affected area. If buildings with major damage are also added to the collapsed list, then building damage increases to 21.5%, underscoring the severity of this earthquake. A similar evaluation can be made for the Aegean Sea earthquake. Its direct damage in Türkiye was around $400 million, somewhere comparable to the damage of the Sivrice earthquake. The sum of the collapsed and severely damaged buildings constituted only 0.45% of the total building stock in the earthquake-affected zone. Although the number of severely damaged buildings was much lower compared to the Sivrice earthquake, the total number of deaths reached 117, and the number of injuries was almost 32,000 due to taller and much more congested buildings.

According to the data in Table 4, the direct damage resulting from the dual Kahramanmaraş earthquakes was about 110 times more than the combined total damage of the Sivrice and Aegean Sea earthquakes. In terms of the number of deaths, Kahramanmaraş surpassed all the Turkish earthquakes of the twentieth century, including the December 27, 1939, Erzincan Earthquake which had an estimated loss of 33,000 lives, and the August 17, 1999 İzmit (Kocaeli) Earthquake, with a total loss of 17,200 lives (Sezen et al. 2003; Tunc and Tunc 2022). In comparison, the extent of the structural damage for the Kahramanmaraş earthquakes was beyond comprehension. As a result, the government established a joint task force, comprised of universities, nongovernmental organizations, and public and private engineering firms, departments, and branches, to assess the damage. Based on government data, approximately 7,100 people were actively involved in this collective effort (Milliyet 2023).

3.3 Common causes of structural damage to RC members

Earthquakes cause minor to substantial amounts of damage to buildings. Damage patterns and their associated issues have been thoroughly studied globally following major earthquakes (Fintel 1995; Rosetto and Elnashai 2003; Rosetto and Peiris 2009; Achour et al. 2011; Kam et al. 2011; Ricci et al. 2011; Wallace et al. 2012; Massone 2013; Suppasri et al. 2013; Barbosa et al. 2017). It is important to note that the findings of these studies mostly aligned with each other, even though earthquake locations varied worldwide. Similar studies were also conducted in Türkiye following major earthquakes such as the 1992 Erzincan, 1998 Adana, 1999 Kocaeli, and 2011 Van earthquakes (Saatcioglu and Bruneau 1993; Adalier 2001; Sezen et al. 2003; Cagatay 2005; Arslan 2007; Korkmaz 2015).

In this study, four recent large-scale earthquakes in Türkiye were investigated. As listed in Sect. 1, many researchers conducted studies on buildings following these earthquakes (Isik et al. 2020; Bayrak et al. 2021; Cetin et al. 2021; Demir et al. 2021; Temür et al. 2021; Yurdakul et al. 2021; Ziotopoulou et al. 2022; Binici et al. 2023; Çağlar et al. 2023; Gurbuz et al. 2023; Ozkula et al. 2023; Ozturk et al. 2023; Avcil et al. 2024; Vuran et al. 2024; Yilmaz et al. 2024;). The authors of this article visited the earthquake-affected sites as well as others and decided to compile their findings by comparing and contrasting them with those of others. The result of this effort is a key chart listing all the common issues related to structural damage (Baran et al. 2014; Mertol et al. 2021a, 2023a, 2023b). Figure 7 outlines the common causes of the damage observed at these four earthquake sites. Each one of these causes was identified, in person, during all site visits, without exception, indicating the overall problem in the building sector. In the following sections, these common issues and their solutions will be discussed.

Fig. 7
figure 7

Common causes of structural damage

Full size image

3.3.1 Materials

During the four site investigations, the inspection teams observed that the quality of concrete in most of the collapsed or heavily damaged buildings did not follow the requirements of the specifications valid at that time. For example, in some of the collapsed or heavily damaged buildings, the color of the concrete was light brown, suggesting the existence of fine aggregates mixed with clay (Fig. 8a). This may result in a lower concrete compressive strength than expected by design. Furthermore, round-shaped natural aggregates were commonly seen in the damaged concrete pieces. These aggregates, which do not comply with the requirements and regulations, had an adverse impact on the bonding properties of concrete (Fig. 8b). The site investigations revealed another problem related to the aggregates’ diameters. Pieces as large as 100 mm were present in the damaged concrete (Fig. 8c). These deficiencies may cause damage and collapse due to easy pull-out of reinforcement and detachment of connecting structural elements.

Fig. 8
figure 8

Problems related to concrete quality and aggregates: a Light brown concrete at Sivrice earthquake, b Round-shaped aggregates at Aegean Sea earthquake, and c Large coarse aggregate size at the Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes

Full size image

The aggregate grading was also very poor due to the existence of large size aggregates. In addition, improper compaction and placement of concrete were found in many structural members. Figure 9 shows some of the examples related to low concrete quality. This type of deficiency may result in poor integrity of structural elements and eventually causing failure.

Fig. 9
figure 9

Poor concrete compaction: a Sivrice earthquake, b Aegean Sea earthquake (Onat et al. 2022), and (c) Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes

Full size image

The longitudinal reinforcement in some of the collapsed or heavily damaged buildings experienced heavy corrosion (Fig. 10a, c). This corrosion was so significant in some of the buildings that it reduced the effective area of reinforcement by as much as 50%. Also, the longitudinal reinforcement in some of the RC pieces was found to be closely spaced that there was almost no adequate space for concrete to infiltrate in between. Clearly, faulty rebar layouts severely impaired the bonding between the rebar and the surrounding concrete (Fig. 10b).

Fig. 10
figure 10

Problems related to reinforcement: a Heavy corrosion in Sivrice earthquake, b Closely spaced longitudinal reinforcement in Aegean Sea earthquake, and c Heavy corrosion in Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes

Full size image

In order to achieve an effective load transfer from a beam to a column, the beam’s longitudinal reinforcement should be adequately anchored to the connecting column. The beam’s top and bottom longitudinal reinforcement should be extended into the column with adequate development length or bent 90° to provide this anchorage. The beams not having this type of reinforcement detail may be separated from the columns causing a total collapse of a building. In many RC buildings inadequate development lengths and lap splices were observed (Fig. 11a, b, c). These issues caused severe damage and failure in buildings since the internal forces resulting from lateral and gravity loads were not transferred effectively from one member to another.

Fig. 11
figure 11

Problems related to development length and lap splices: a Sivrice earthquake, b Aegean Sea earthquake, and c Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes

Full size image

In RC buildings constructed before 2000, the use of smooth rebars was observed in columns, beams, and structural walls, as well as smooth/ribbed combinations (Fig. 12a, b, c). Smooth rebars caused serious bonding issues with the surrounding concrete, eventually leading to structural failure. It is because of this issue that their use as longitudinal rebars in RC columns, beams, and structural walls was banned by TEC 1998. However, it took a number of years for it to be completely phased out.

Fig. 12
figure 12

Problems related to use of smooth rebars: a Sivrice earthquake, b Aegean Sea earthquake, and c Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes

Full size image

When the RC structural elements are subjected to high stresses due to bending moments, shear forces, and axial loads, the concrete outside the stirrups (cover concrete) spalls. If the ends of stirrups are not bent into the core concrete area using an angle of 135°, the stirrup will open up by following a cover spalling. Therefore, the use of 90° bent stirrups is not recommended under earthquake loads since it will result in severe damage and collapse. Based on the site investigations, it was found that the ends of the stirrups were not bent 135° towards the core concrete in most of the collapsed and heavily damaged buildings. This clearly resulted in the opening of the stirrups, which loosened the stability of core concrete (Fig. 13a, b, c).

Fig. 13
figure 13

Stirrup ends bent 90°: a Sivrice earthquake, b Aegean Sea earthquake, and c Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes

Full size image

The site investigations also revealed that confinement zones were not formed at the ends of beams and columns in most of the collapsed and heavily damaged buildings, as shown in Figs. 14a through c. This outcome was a result of the inadequate use of closely spaced stirrups at the ends of columns, beams, and walls.

Fig. 14
figure 14

No confinement zones at ends of beams and columns: a Sivrice earthquake, b Aegean Sea earthquake, and c Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes

Full size image

The deficiencies mentioned in this section are some, but not all, of the most significant causes of the widespread damage and collapse of the buildings experienced in the four earthquakes under examination. These deficiencies—the use of low-quality concrete, corrosion, the misconfiguration of reinforcements, 90° bent stirrup ends, and no confinement zones at the ends of structural elements—could have been avoided if the regulations and specifications valid at that time had been followed during construction.

3.3.2 Technological limitations

The other common problem was the technological limitations that existed during the construction of thousands of buildings. Over half of Türkiye's buildings are made of RC, and this percentage is over 90 for buildings constructed after 2000 (Gunes 2015). The quality of reinforcing steel has never been a concern due to its well-controlled fabrication and testing conditions. Concrete mixing, pouring, and handling, on the other hand, have always raised serious concerns. This began to change in the late 1980s, when concrete from ready-mixed concrete plants came into use. Therefore, in general, buildings constructed after 2000 have better concrete quality. This statement is exemplified by the damaged buildings inspected at the four earthquake sites. While some of the damaged buildings in the dual Kahramanmaraş earthquakes were built after 2000, the main reason behind the damage was not faulty concrete, but other causes, as seen in Fig. 7.

3.3.3 Framing system

Shear walls are quite effective in lateral resisting systems. Their existence in earthquakes is therefore critical, as they withstand horizontal forces and act as a redundant. Numerous publications in this area emphasize their benefits (Inel et al. 2008; Erdil 2016; Tunc and Al-Ageedi 2020; Şimşek et al. 2023). In some low-rise buildings (five to ten floors), there were no or very few shear walls (the ratio of the area of the shear walls to the floor area was less than 1.0%, (Burak and Comlekoglu 2013; Tunc and Al-Ageedi 2020; Özmen 2021)), indicating that their lateral resisting systems were almost entirely comprised of moment resisting RC frames. Thus, the absence of shear walls contributed to the damage and collapse of buildings (Mertol et al. 2021a; Onat et al. 2022; TPS 2023).

When there are few or no shear walls in a framing system, proper beam-to-beam or beam-to-column framing becomes crucial in transferring lateral loads effectively down to a building’s footing. If this is not properly considered during the design stage, it can cause damage and collapse. Poor design decisions with respect to these kinds of connections can propagate errors during the construction phase since more labor-intensive manufacturing will be required (Fig. 15a, b, c).

Fig. 15
figure 15

Framing issues: a Sivrice earthquake, b Aegean Sea earthquake, and c Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes

Full size image

In many residential buildings, balconies cantilever anywhere from 1.5 to 2.5 m. These cantilevering slabs create both bending and torsional moments mostly on their supporting beams. Also, due to the nature of earthquake loads, these cantilevering slabs experience vertical forces in both the downward and upward directions. However, the vertical component of an earthquake was not included in the design, since their impact was omitted from the specifications until the latest earthquake code, TBEC 2018. Moreover, as the team observed, in many cases their construction did not fully comply with design drawings due to the steel congestion created in the supporting beams. These cantilevering slabs, therefore, are believed to be another reason for damaging the supporting columns and the building itself (Fig. 16a, b, c).

Fig. 16
figure 16

Balconies: a Sivrice earthquake, b Aegean Sea earthquake, and c Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes

Full size image

In earthquake-resistant design, the columns that connect to a joint must be stronger than the beams that connect to the same joint, also known as the strong column-weak beam design philosophy (TEC 2007). However, during field inspections after the four earthquakes, it was observed that some buildings had beam capacities that were larger than their column capacities. This created a serious problem as it led to the damage or collapse of buildings during an earthquake (Fig. 17a, b, c).

Fig. 17
figure 17

Strong column-weak beam issue: a Sivrice earthquake, b Aegean Sea earthquake, and c Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes

Full size image

Soft and weak story phenomenon was another parameter that may have contributed to the building failures, especially at the ground floor level due to the existence of shops. In order to accommodate shops, story heights were often increased at the ground floor level and non-load-bearing partition walls present on other floors were removed. These arrangements unfortunately caused stiffness and strength weaknesses, mostly at the ground floor levels and eventually leading to building failures (Fig. 18).

Fig. 18
figure 18

© 2021 Google): a Sivrice earthquake, b Aegean Sea earthquake, and c Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes

Collapsed buildings having soft and weak stories before and after (map data

Full size image

Short columns were another reason for the damaged or collapsed buildings. Although the current and previous earthquake codes (TBEC 2018; TEC 1975, 1998, and 2007) allow the use of short columns, they must be designed accordingly, or transverse reinforcement should be applied along their lengths to resist larger shear forces (Fig. 19a, b, c).

Fig. 19
figure 19

Short columns: a Sivrice earthquake, b Aegean Sea earthquake, and c Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes (with permission from Earthquakedatabase, photo by Burak Ozsecer)

Full size image

3.3.4 Building inspection

In Türkiye, construction quality has always been an issue. In order to better understand this concept, existing buildings should be classified into two groups: government-owned and private. Until 2011, the inspections of government-owned buildings were carried out more effectively since rules and regulations existed for these structures. However, almost all privately owned buildings were excluded from such inspections, shifting responsibility to owners. Therefore, many privately owned buildings built before 2011 were not inspected during their design or construction stages (Erdik 2000; Sezen et al. 2000; Mertol et al. 2021a). In order to solve this matter, in 2001, the Construction Inspection Law No. 4708 was passed and enacted in 27 out of 81 Turkish cities. It is important to note that inspection was finally mandated in all cities in 2011.

Another related issue is funding (Mertol et al. 2023b). Between 2001 and 2019, general contractors were solely responsible for selecting and paying inspection companies. This led to conflicts of interest and troublesome practices. For example, if a contractor did not agree with the inspection report, they could have refused to pay the inspection company. This resulted in the hiring of another, more flexible, inspection, while construction continued. Once this practice was detected, a new system was created in 2019 that made the inspection process more reliable, transparent, and trustworthy. Inspection companies are now randomly assigned to construction projects, and their payments are made directly from a central pool to which all general contractors contribute. This has eliminated previous conflicts of interest and has encouraged proper inspections (Mertol et al. 2021a, 2023a).

3.3.5 Soil studies

The first geological site studies in Türkiye were conducted in 1986 as part of zoning plan requirements. However, some municipalities ignored and did not enforce this prerequisite; thus, many privately owned buildings were designed without proper site investigations. In 1987, the government stepped in and consequently, the Ministry of Public Works now mandates the approval of all geological site reports by either a university or the General Directorate of Natural Disasters. After the 1999 Kocaeli and Düzce earthquakes, a geological site report became a required element when issuing a building permit. In 1999, a comprehensive memorandum was released, which then became part of the design requirement. Therefore, in general, before 1999, soil-related design parameters were not a priority in building design. This negligence resulted in major damage and collapse in many privately owned buildings. Thus, the year 2000 can be considered a milestone in soil studies since it is when proper and more accurate soil parameters were introduced into building design (Mertol et al. 2021a, 2023a, and 2023b). This issue was another factor behind the damaged and collapsed buildings of the four earthquakes.

Soil liquefaction was another notable problem that led to serious building damage (Fig. 20a, b, c). Since 2000, when soil studies became mandatory for all buildings regardless of their location, this problem has diminished significantly. However, buildings constructed before 2000, especially in such problematic soil regions, require extensive soil investigation to prevent any further earthquake damage.

Fig. 20
figure 20

Soil Liquefaction: a Sivrice earthquake (METU 2020), b Aegean Sea earthquake (Ziotopoulou et al. 2022), and c Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes (with permission from Earthquakedatabase, photo by Hamdi Koser)

Full size image

3.3.6 Workmanship

Another cause of damaged or destroyed buildings is poor workmanship. This particular topic has been studied extensively by other researchers (Sezen et al. 2000; Inel 2008; Baran et al. 2014; Gunes 2015; Atmaca et al. 2020; Çağlar et al. 2023). During site investigations, poor workmanship was repeatedly observed (Fig. 21a, b, c, d, e, f).

Fig. 21
figure 21

Poor workmanship: a Sivrice earthquake, b Aegean Sea earthquake, c through f Mw = 7.7 and Mw = 7.6 Kahramanmaraş earthquakes (figures d, e, and f used with permission from Earthquakedatabase; photos by Esra Kevser Gül, Tayyip Karakoyun, and Evrim, respectively)

Full size image

It is clear that the other causes shown in Fig. 7 contribute to this category, which ultimately impacts all other factors. Figure 21a shows a poorly constructed staircase and its failure at the landing and column support locations. The landing is being supported by a non-load-bearing brick wall which collapsed during the Sivrice earthquake and led to the ultimate failure of the staircase. In Fig. 21b, the progressive collapse of a building in İzmir is shown. The collapse mechanism indicates a shear failure between the slabs and the columns and the slabs and the walls, most likely at roof level, due to inadequate rebar anchoring, which was a result of faulty field practice. Consequently, the failed roof slab collapsed on a lower floor slab and initiated a progressive collapse that eventually yielded the partial collapse of this building. In Fig. 21c, a very slender and structurally weak vertical member, which does not even fit the description of a structural column, was constructed as a corner column to support a balcony. Due to its poor workmanship and inadequate section details, it failed, causing a part of the structure to fail. Alarming workmanship is also evident in Figs. 21d through f. In Fig. 21d, failed concrete pieces in a shear wall resulting from improper mix details can be observed. Figure 21e shows extremely poor workmanship that resulted in multiple column failures at the base and led to the failure of the building. Finally, in Fig. 21f, shear wall failure due to inadequate rebar anchoring is shown. It resulted in progressive collapse and led to partial failure of the building.

During construction, overlooked or underestimated problems can be eliminated or solved. The responsibility lies with experienced and knowledgeable contractors who are, in the end, accountable for construction projects. Thus, it is important to work with skilled, qualified, and conscientious contractors. The unskilled, not experienced, and not eligible ones then should be entirely removed from the system so that clients would not even have an option of selecting them. It is believed that the critical decision is to prevent these types of ineligible firms from participating in any public or private bids so that their only driving force of “offering low-cost construction” for once and all can be eliminated. In order to do this, a collaborative effort from the private and public sides should be made without giving any consent to any of these ineligible firms.

3.3.7 Evaluation of common causes of structural damage

In the previous subsections, Sects. 3.3.1 through 3.3.6, structural damage resulting from the four earthquakes was examined in detail. The site investigations concluded that some of the damage types were more commonly seen when compared to others. Moreover, some of the design and construction issues shown in Fig. 7 most likely contributed more to the outcome than others. Therefore, the list in Fig. 7 has been modified based on the recurrence rate of these common issues (see Table 5). As listed in the table, issues related to workmanship were most commonly found, followed by building inspections and soil studies. Other issues include the reinforcement and concrete under materials, the framing system, and technological limitations. It is important to note the order of the issues from one to six does not necessarily represent the significance of their contribution to ultimate failure or damage, since each failure might have been the result of one or more issues.

Table 5 Common issues based on their rate of recurrence
Full size table

4 Conclusions and recommendations

In this article, lessons learned from the four recent earthquakes were discussed in detail. The first, the Sivrice earthquake, struck the eastern part of Türkiye on January 24, 2020, with Mw = 6.8. The second, the Aegean Sea earthquake, hit the western part of the country on October 30, 2020, with an Mw of 6.6. The third and fourth were on February 6, 2023, the Kahramanmaraş dual earthquakes, with Mws of 7.7 and 7.6 in the eastern part of Türkiye, approximately 9 h apart. The authors visited these sites immediately after each earthquake. They strongly believe in the importance of these visits, even though many challenges occur in the process. Such visits are crucial since rescue efforts might destroy evidence that might be useful in determining structural issues related to damage or collapse. Therefore, being on-site undeniably adds to the value of finding concrete evidence for structural damage.

Drawing solid conclusions requires full access to all documents, drawings, and details. However, in this study, there were notable limitations with respect to conducting a thorough inspection work on buildings located in earthquake-damaged areas. These limitations were mainly as follows: (a) the authors were not permitted to examine some damaged buildings due to strict government regulations, (b) the permit drawings of buildings that were of interest were almost always unavailable due to ongoing lawsuits and legal issues, (c) the authors could not take core samples from damaged buildings since the sole authority on this matter is AFAD, and (d) building inspection reports and other types of construction documents and material test reports were almost always inaccessible again due to ongoing lawsuits and permit issues. Although these limitations clearly prevent the exposure of all the details related to the collapsed and damaged buildings’ construction and design phases, additional efforts were made to compensate for these deficiencies during site visits by collecting data and information from all responsible parties that were involved in the design and construction.

The following list includes the conclusions and recommendations made based on the site observations. This list was prepared in two clusters. The first involves the most critical and urgent issues (see items 1 through 8). In the second cluster, other types of issues, such as advanced design and construction details, are addressed (items 9 through 14).

  1. 1.

    The content of the four earthquake design codes that were examined, 1975, 1998, 2007, and 2018, were all found to be adequate. As a result of this outcome, it is fair to say that the design phase is not the primary suspect of the building damage observed on these sites. However, the construction phase was found to be the most problematic area due to a lack of site inspection. Although the current inspection law seems to be adequate, many damaged and collapsed buildings were constructed at a time when site inspection laws were either entirely missing (before 2000) or they existed with major problems related to their financial structuring and establishment (before 2019). Therefore, the authors strongly believe in the need for a more stringent site inspection law to be enforced immediately for all buildings of Türkiye.

  2. 2.

    During site visits, the common practice of allocating the first floors of buildings to shops was found to be very troublesome since their existence adversely affected the structural integrity of buildings and eventually led to their failure. Therefore, it is highly recommended that this common practice is abolished entirely.

  3. 3.

    Another common problem dealt with buildings that have smaller footprints on their ground floors and expand as they progress to their upper floors. This practice led to unnecessary cantilevering overhang slabs and jeopardized healthy structural load transfer from upper floors to foundations. Therefore, it is highly recommended to abolish this practice and maintain a uniform footprint across the height of a building.

  4. 4.

    Shear wall presence on buildings is found to be very crucial since it plays a significant role in resisting earthquake loads. In many collapsed or severely damaged buildings, the structural framing either lacked shear walls or included them in very small quantities, leading to severe damage or collapse since lateral loads were not effectively carried by the columns. Therefore, buildings with no shear walls (or frame structures) should not be permitted at all in earthquake risk areas. The authors strongly believe that the next earthquake design code should be revised to reflect this change. More specifically, buildings should be designed with shear walls that occupy at least 1.0 to 1.5% of the total gross area per floor. These walls should be oriented as equally as possible in both principal directions.

  5. 5.

    Another common issue observed during site visits was related to illegal alterations on columns, beams, and shear walls, which were performed to create unobstructed areas in shops and parking lots often found on ground and basement floors. These alterations are believed to have been made either by tenants or owners, without official permission. Based on the site investigations, some buildings are believed to have experienced either collapse or sustained heavy damage because of this severe negligence. Therefore, it is highly recommended to inspect buildings regularly during their occupancy to deter such practices.

  6. 6.

    In many collapsed or severely damaged buildings, the existence of balconies is found to be a critical parameter that adversely affects the expected flow of a lateral load transfer. The existence of these elements caused the formation of inappropriate framing, which eventually affected the expected lateral load transfer. Therefore, in earthquake risk areas, balconies should be eliminated as much as possible. Otherwise, their framing should include the direct support of columns and beams spanning between them.

  7. 7.

    The results shown in Figs. 5a through f emphasize the fact that the severe destruction caused by these earthquakes is somewhat related to the soil amplification phenomenon due to the existence of alluvial soil in all these regions. Therefore, it is important to conduct thorough geotechnical site surveys before starting any construction. In areas near major fault lines, site-specific soil investigations should be conducted.

  8. 8.

    During the dual Kahramanmaraş earthquakes, it was observed that the 5% damped spectral acceleration values in stations near the epicenter and those specifically in Hatay well exceeded the design acceleration values of TBEC 2018 calculated in 5 different soil types (A through E). Therefore, it is important to reevaluate short and 1-s spectral acceleration values in areas near all major fault lines.

  9. 9.

    Earthquake duration is considered to be an important parameter in assessing structural damage. In this study, the durations of the dual Kahramanmaraş earthquakes, approximately three to four times longer than those of the Sivrice and the Aegean Sea earthquakes, are believed to have played a significant role in causing the collapse and extensive damage to buildings.

  10. 10.

    The number of aftershocks and their magnitudes resulting from the dual Kahramanmaraş earthquakes were quite significant compared to the numbers recorded from the other two earthquakes. These high numbers and magnitudes are strongly believed to have contributed to the extent of severe damage from the Kahramanmaraş ones.

  11. 11.

    Deficiencies such as the use of low-quality concrete, corrosion, the misconfiguration of reinforcement, 90° bent stirrup ends, and no confinement zones at the ends of structural elements were observed in most of the heavily damaged or collapsed buildings in the four earthquake zones. If the regulations and specifications valid at the time of construction were followed, loss of life and property damage might have been mitigated.

  12. 12.

    A comprehensive new design and construction approach should be developed for severe earthquake regions surrounded by major fault lines. This new approach would help compensate for the common causes of earthquake damage, as outlined in Fig. 7. The quality of design and, more importantly, construction should be prioritized over any other factor. A simple yet effective earthquake resistant design approach, based on redundancy, should be promoted in these areas so that safer building construction can occur.

  13. 13.

    In order to minimize any further loss of life and property in future earthquakes, an earthquake hazard assessment should be completed near all major fault lines.

  14. 14.

    Buildings located near major fault lines (a two-kilometer strip along the both sides of the fault line), regardless of their use, should be designated as critical. Their design should also reflect their special status. For example, a constant importance factor of I = 1.5 should be used for all these buildings.