A Comparative Study on the V S 30 and N 30 Based Seismic Site Classification in Kahramanmaras, Turkey

Assessment of seismic site classiﬁcation (SSC) using either the average shear wave velocity ( V S 30 ) or the average SPT-N values ( N 30 ) for upper 30 m in soils is the simplest method to carry out various studies including site response and soil-structure interactions. Either the V S 30 - or the N 30 -based SSC maps designed according to the National Earthquake Hazards Reduction Program (NEHRP) classiﬁcation system are eﬀectively used to predict possible locations for future seismic events. The main goal of this study is to generate maps using the Geographic Information System (GIS) for the SSC in Kahramanmaras city, inﬂuenced by both East Anatolian Fault and Dead Sea Fault Zones, using both V S 30 and N 30 values. The study also presents a series of GIS maps produced using the shear wave velocity ( V S ) and SPT-N values at the depths of 5m, 10 m, 15 m, 20m, and 25m. Fur-thermore, the study estimates the bed rock level and generates the SSC maps for the average V S values through overburden soils by using the NEHRP system. The V S30 maps categorize the study area mainly under class C and limited number of areas under classes B and D, whereas the N 30 maps classify the study area mainly under class D. Both maps indicate that the soil classes in the study area are diﬀerent to a high extent. Eventually, the GIS maps complied for the purpose of urban development may be utilized eﬀectively by engineers in the ﬁeld.


Introduction
Seismic site classification (SSC), which defines engineering properties of the soils by means of V S or SPT-N values, is the simplest method to consider the site effects for numerous purposes including engineering projects and microzonation studies [1]. Many classification systems utilize the SPT-N and V S values measured in upper 30 m to provide an SSC assessment [2][3][4].
e National Earthquake Hazard Reduction Program (NEHRP), which is one of the widely used systems for the SSC applications, classified the soils by using their V S values in designing building codes and characterizing site response [5,6]. e objective of this study is to make an SSC assessment in the city of Kahramanmaras located in a seismically active region in southern-central Turkey. e region is influenced by the Eastern Anatolian Fault (EAF) and the Dead Sea Fault (DSF) [7]. No studies using SPT-N values have been performed in the city and vicinity area, although little efforts have been made using V S30 values obtained in the city centre [8]. Variations in the soil classification have been investigated at the depths of 5 m, 10 m, 15 m, 20 m, and 25 m by producing GIS maps for both the V S and SPT-N values obtained in the study area. e subsoil exploration has been conducted by means of seismic surface wave techniques (Multichannel Analysis of Surface Waves, MASW, and Microtremor Array Measurements, MAM) at 287 boreholes in 98 sites. e GIS maps of both weathered and engineering bed rock levels and average V S in overburden soils have also been produced for further studies. e present study is thought to be the first investigation carried out to estimate the SSC using both V S30 and N 30 values in the study area.

Seismic Activity and Geological Setting of Kahramanmaras Area
Kahramanmaras basin developed by tectonic-origin surface deformations in the Quaternary and Holocene is located in south of the Bitlis-Poturge Massif area nearby the triple junction of the Arabian, African, and Anatolian plates [9,10]. Due to the collision of Arabian and Eurasian plates over the Bitlis Suture, the basin was eventually filled out by heavy alluvial sediments and dense turbiditic flysch sequences [11][12][13][14]. In the basin, Alacik formation overlies the older units with unconformable contact starting with base conglomerate [15]. In the north, the formation shows different lithological properties that contain marl, clayey limestone, coal band alternations, and claystone. e quaternary age alluviums, dominant materials available in the basin, are composed of gray/light gray, gravel, sand, and silt, which are loose textured and cementless deposited in horizontal and vertical directions. e alluviums thickness in the region was found to increase up to 300 m, which can increase the earthquake intensity by 2-3 degrees [7] ( Figure 1). e collision along Arabian and Anatolian Plates formed the Kahramanmaras territory. e available structural schemes set the triple junction of Arabian, Anatolian, and African plates at the northern Karasu valley, nearby the study area [17]. e study area is surrounded by active faults of the EAFZ (i.e., Surgu Segment, Savrun Segment, Cardak Segment, Toprakkale Segment, Cokak Segment, Amanos Segment, and Golbasi Segment), Kahramanmaras Fault Zone, Engizek Fault Zone, and Narli segment of the Dead Sea Fault Zone (DSFZ) [18]. e Engizek Fault Zone with a length of 66 km locates in north of the study area [19,20].
e Kahramanmaras Fault Zone (KFZ) begins from east of the study area and passes through the northern part of Kahramanmaras city and extends until the Kilavuzlu Dam [19,20]. Yilmaz et al. [16] revealed that the Golbasi-Turkoglu segment has suffered from numerous devastating earthquakes (Table 1). An earthquake with a magnitude of 6.8 occurred in 1544 on Cardak Fault [21]. Nalbant et al. [22] studied the KM zone (the area between Kahramanmaras and Malatya), which has gathered significant amount of stress during the last 200 years; thus the magnitude of a possible earthquake along the segment ruptured by 1114 (Mw > 7.8) and 1513 (Mw > 7.4) earthquakes is expected to be larger than 7.3. Palutoglu and Sasmaz [18] pointed out that another destructive earthquake in the vicinity of study area took place in 1795 with a magnitude of 7.0 ( Figure 2). Accordingly, Golbasi-Turkoglu Segment, the KFZ, the Engizek Fault, and the Cardak Fault come to the forefront. erefore, Kahramanmaras and its vicinity area are defined as an area with a seismic gap, and large earthquakes are expected to occur in a near future.

Methodology
One of the most widely used surface wave analysis methods to calculate V S30 and site response analysis is the Multichannel Analysis of Surface Wave (MASW) method [23,24]. e MASW is a seismic survey employed to record Rayleigh waves on a multichannel record [25]. Subsurface wave velocities at 88 sites were measured using MASW in the study area by the municipality of Kahramanmaras in order to complete the geophysical studies. e test equipment consists of a 12-channel geometrics seismograph with 4.5 Hz frequency vertical geophones spaced 3 m apart from each other and a sledge hammer with 9 kg weight. Locations, where the 88 MASW tests were carried out, are shown in Figure 3.
Microtremor Array Measurements (MAM) carried out by Ozmen et al. [8] at 10 different locations in the study area were used to estimate the shallow S-wave velocity profiles. e Spatial Auto Correlation (SPAC) method was performed to recover Rayleigh wave phase velocity in a 1-30 Hz frequency range. Microtremor data were recorded simultaneously by means of a 24-bit A/D portable recorder on sensors mounted in the field over various durations. e MAM tests were carried out on the clastic and carbonate rocks except sites 4624 and 4625 lying in a quaternary soil. Ozmen et al. [8] revealed that most of the profiles in the study area have top layers with less than 10 m thickness. V S of the top layers were found to be high at many sites, apart from three sites (4617, 4624, and 4626) with less than 200 m/s in the study area. Figure 4 shows an inverted 1D S-wave velocity profile up to 100 m depth by means of a hybrid heuristic.

Seismic Site Classification for the Kahramanmaras Area
Site classification using NEHRP system is determined based on the average properties of the soil within upper 30 m in the ground, including V S30 , N 30 , and undrained shear strength (s u30 ) [26]. Table 2 lists the five different site classes in terms of V S30 and N 30 proposed by the NEHRP system. e present investigation proposes a series of SSC maps prepared using a GIS software in order to classify the study area based on V S30 for 76 locations and N 30 for 287 boreholes.

V S30 -Based Seismic Site Classification. V S30
, a parameter used for site classification in building codes, can be estimated by dividing 30 m with the required time of the wave travelling over an upper 30 m depth [2,27]. e V S30 values required for the SSC could be estimated by using the following equation [26]: where d i is the thickness of the i th layer, V Si is V S of the i th layer, and n is the number of the strata within the upper 30 m. Boore [28] proposed four different methods to estimate the values of V S30 for the profiles which do not have a 30 m depth. e method employed in the present study was the correlation between shallow velocities and V S30 , where the key parameter is time-averaged velocity to a certain depth d (V Sd ). V Sd can be calculated using the following equation: 2 Advances in Civil Engineering where tt (d) is travel time to a depth d and it is described as follows: e datasets of 15 MASW and 10 MAM V S sites with actual depths of measured shear wave velocity models reaching to 30 m were employed to extend the depths with less than 30 m. Actually, V S30 for shallow models were estimated by correlating V S30 using these readily available 25 sites against their V Sd for the depths (d) assumed. e depths (d) were found to be ranging from 10 m to 24 m in such approach over the study area. As recommended by Boore [28], the authors have ignored the sites with less than a 10 m depth and produced straight regression lines for various depths from 10 m to 24 m ( Figure 5). It seems that the results obtained are consistent with those obtained by Boore [28]. As can be seen from Figure 5, coefficient of correlations (R 2 ) is higher for the larger depths likely because of the increase in effective stress resulting in decrease in void ratio [29][30][31][32]. e form of the equations for these regression lines was found to be as follows: where V S30 is the average shear wave velocity, V Sd is the timeaveraged velocity at a certain depth, a and b are the coefficients. e authors have estimated the V S30 values for the 76 sites by employing equations (1)-(4) and plotted their findings using ArcGIS software on Figure 6. e V s estimates made by the authors were found to be consistent with the studies by Biricik and Korkmaz [7], Ozmen et al. [8], and Perincek and Kozlu [12] carried out on the nearby area. It can be seen that the study area was primarily classified as C with a 78% of the entire area, while classes B and D were seen in much narrow regions through the study area. Figure 7 has been plotted in order to characterize the variation of soil types at the depths of 5 m, 10 m, 15 m, 20 m, and 25 m. A 2.9% of the soils at 5 m depth was classified as E (red regions close to the airport), an 11.6% of those was classified as B, a 39% of those was classified as D, and 46.5% was classified as B (the northern regions) (Figure 7(a)). Similarly, variation of the soil types at a depth of 25 m was found to be about 6.1%, 28.6%, and 65.3% for classes B, D, and C, respectively (Figure 7(e)).

N 30 -Based Seismic Site Classification.
e SPT-N values, which were corrected for hammer energy variation, show the presence of dense to very dense sands across the entire study area apart from Yavuz Selim region, a recently settled industrial area in the southern region, where loose sands with clay and silt inclusions were found. e SPT-N values up to 30 m depth can be estimated by using the equation proposed by Kumar et al. [33] as follows: where N 30 is the average SPT-N through 30 m, d i is thickness of the i th layer, n is number of the layers in 30 m depth, and N i is the value of SPT-N for layer i. e authors had employed equation (5) in order to generate the SPT-N based SSC maps through the 287 boreholes across the study area. Following N 30 estimations, the SSC maps over the study area using the NEHRP classification system have been produced (Figure 8) [3]. As can be seen from the Figure 8, 57.5% of the soils were classified as D, while the rest (42.5%) were classified as C. Average SPT-N maps for the depths of 5 m, 10 m, 15 m, 20 m, and 25 m have been plotted in Figure 9. For example, types of the soils at 5 m depth were found to be class E (dark brown colour in the central, southern, and eastern parts), class C (northern parts), and class D with the percentages of 8.4, 11.1, and 80.5, respectively (Figure 9(a)). In general, class D was found to be the dominant soil type at any depth in the study   Advances in Civil Engineering area. Soft soil deposits were mainly observed at relatively shallow depths in the study, and their amounts were decreased by increasing depth.

Spatial Variability of Rock Depth
Classification of rock based on the shear wave velocity (V S ) in geotechnical earthquake engineering is still in practice. According to the NEHRP classification system, soft rock and very dense soil at upper 30 m have the average velocities of 360 m/s to 760 m/s. e 30 m average V S of a value more than 760 m/s is classified as a hard rock [34]. For the purpose of seismic microzonation, Nath [35] defined the seismic bed rock with an average V S value of 3000 m/s and higher and from 400 m/s to 700 m/s for the engineering bed rock. Miller et al. [36] had assigned the level of bed rock through considering V S as 244 m/s and above by using MASW survey. For the present study, velocity of shear wave of 330 ± 30 m/s has been specified for the weathered rock (WR) and 760 ± 60 m/s has been appointed for the engineering rock (ER). Among the 98 V S locations for the MASW and MAM surveys, only 83 locations were able to be utilized to assign the upper level of the stratum corresponding to V S of 330 ± 30 m/s. e estimated depths were tabulated (Table 3) and employed to generate the map for weathered rock level up to 14 m depth via ArcGIS Software V 10.4.1 (Figure 10). About 67% of the region has weathered rock level at the entire ground surface except the central part and a small area at the eastern side. e weathered rock level between 0 m and 5 m was estimated about 17.7% of the region. About 12.7% of the region was observed with weathered rock from 5 m to 10 m depth (dark orange colour). Around 2.6% of the region has weathered rock level from 10 m to 14 m (around the airport). In general, more than half of the study area was found to be a very dense soil or soft rock at the surface. Engineering bed rock strata were observed at 52 different locations, where V S values correspond to 760 ± 60 m/s. Data in those locations have been employed to generate the map of engineering bed rock level (Figure 11). Engineering bed rock level in the study area was found to be at the ground level in northern region (Ahir Dagi mountain) with a percentage of 9.7% of the boreholes. Depth of the engineering rock between 0 m and 10 m depth in the northern-central and western regions was about 17.3% of the whole boreholes. Almost half of the study area (48%) has an engineering depth ranging from 10 m to 20 m (dark green color). e engineering rock level was found to be between 20 m and 32 m for 23% of the entire study area (light green colour). For a depth greater than 100 m, the engineering rock level was found to be about 1.92% of the whole data (site 4624 shown in brown colour). Figures 12 and 13 present the overburden soil depth to the weathered rock and engineering bed rock levels, respectively. e equivalent V S for the depth of soil is indicated as V H and computed as follows [37]:      Furthermore, average V S value for the dense overburden soil and/or soft rock from 5 m to 32 m depth, about 85% of the study area, was found to be more than 360 m/s.

Conclusions
e purpose of this study is to generate an intensive series of Geographic Information System (GIS) maps for the seismic site classification (SSC) in Kahramanmaras city, influenced by both west part of East Anatolian Fault and northern part of Dead Sea Fault Zones with a capability of producing the strongest future earthquakes in southern-central Turkey.
e study presents the GIS maps for average SPT-N values (N 30 ), average shear wave velocity values (V S30 ), variation in depths to weathered and engineering rock levels, and average V S of overburden soil. e V S30 based SSC maps indicated that majority of the soils in the study area were found to be class C, although N 30 based SSC maps indicated that more than 50% of the study area was class D. e differences among the findings by two different approaches could be explained by the accuracy of both geophysical tests of MASW and MAM at the field because of some difficulties in opening 287 boreholes and recording their SPT-N drops. Another reason could be the method employed for extending the V S profiles up to 30 m. is approach increases the queries on consistency of NEHRP classification system. Furthermore, another series of raster maps generated for average V S and SPT-N values at 5 m, 10 m, 15 m, 20 m, and 25 m depths revealed that 85% of the study area, according to the weathered bed rock level, was classified as soil type D with depths ranging from 3.9 m to 14 m, while 84% of the study area, according to the engineering bed rock level, was classified as C from 5 m to 32 m depth.