Displacement analysis of the October 30, 2020 (M_w = 6.9), Samos (Aegean Sea) earthquake

1 Introduction

A great number of destructive earthquakes have occurred in the Aegean Sea from the historical period to the present. The Samos Island earthquake occurred on October 30, 2020 (M _w = 6.9) on the Aegean Sea (offshore of Seferihisar district (Izmir), North of Samos Island) is one of the most destructive earthquakes and affected Izmir city with many losses of lives and property. Samos Fault, which was the source of the October 30, 2020 Samos earthquake, is located in a highly deformed back-arc region with an extensional component of the Aegean microplate. In this region, the deformation has been caused by the effects of the westward motion of the Anatolian microplate along the North Anatolian Fault Zone and the subduction of the Eastern Mediterranean plate under the Aegean microplate. This active motion of the Anatolian microplate is transferred to the western part of Turkish coasts and Greek Islands (Sboras et al., 2011, 2014; Chatzipetros et al., 2013).

Aegean micro plate consists of horst-graben structures bordered by mainly normal faults as a result of the extensional tectonic regime. In the Aegean Sea and Western Anatolia, between the years 2012 and 2022, 420 earthquakes occurred in the range of 4 < M < 5, 37 earthquakes in the range of 5 < M < 6, and 3 earthquakes with M > 6 (http://deprem.afad.gov.tr, Disaster and Emergency Management Presidency of Turkey (AFAD)). The focal mechanisms of the most important earthquakes (M > 4.5) are presented in Table 1 and Figure 1. Due to the active tectonic regime, many destructive earthquakes have occurred and affected many lives in the region. Finally, the last major earthquake occurred in the Aegean Sea is October 30, 2020 Samos Island earthquake (Kandilli Observatory and Earthquake Research Institute (KOERI), 2020), and its epicenter location is shown in Figure 1.

Figure 1

(a) The main tectonic structures and the focal mechanisms of five earthquakes occurred in the study area (http://deprem.afad.gov.tr). Number 4 represents the October 30, 2020 earthquake. GF: Gülbahçe Fault, OFZ: Orhanlı Fault Zone (For Land Faults; http://yerbilimleri.mta.gov.tr/, for marine faults; DAUM 2020). (b) The focused map of Samos Island and its surroundings.

According to various research centres, the magnitudes and the focal depths of the earthquake are given as follows: M _w = 6.9 h = 10 km; M _w = 6.6 h = 16 km; M _w = 7.0 h = 11.5; and M _w = 7.0 h = 15 km by KOERI, AFAD, USGS, and GFZ (GeoForschungs Zentrum, Deutsches Helmholtz-Zentrum Postdam), respectively. The earthquake was felt in a wide region, especially in Izmir and its surroundings. The orientation and mechanism of the October 30, 2020 Samos earthquake (M _w = 6.9) is consistent with the back-arc extension in the Aegean region. In addition, the earthquakes occurred in the range of 3 < M < 7 between October 29, 2020, and October 20, 2023 in Samos Island, and surroundings are analyzed and 3D epicenter distributions are presented in Figure 2.

Figure 2

3D epicenter distributions of the earthquakes occurred in Samos Island and its surroundings (3 < M < 7, 29.10.2020-20.10.2023) (http://www.koeri.boun.edu.tr/sismo/zeqdb) (Samos Fault digitized from Pavlides et al., 2009) (Green, red, and blue axes represent; North/South, East/West, and Z direction, respectively.).

Samos Island and its surroundings are controlled by normal faults and strike-slip faults which are important in the transtensional tectonic setting behind the Hellenic subduction zone (Dewey and Şengör, 1979; Jackson and McKenzie, 1984; Le Pichon et al., 1995). Historical earthquake records show that this area has been under the influence of destructive earthquakes since 200 BC (Guidoboni et al., 1994). The 1751 AD, 1865 AD, and 1890 AD earthquakes strongly affected the settlement surrounded by Kuşadası Bay and Samos Island (Ergin, 1967; Guidoboni et al., 2007; Ambraseys, 2009). The 1904 earthquake (M _w = 6.8) caused huge damage around Samos Island (Tan et al., 2014). However, due to the fact that a significant part of the seismic sources is under the sea, the information about the faults causing the earthquakes is very limited, except for the Samos fault. The Samos fault is a normal fault (Sboras et al., 2011) and shapes the submarine collapse at deeper than 1 km depth to the north of Samos Island (Chatzipetros et al., 2013; Gurcay and Cifci, 2021).

Previous studies (e.g. Aktuğ et al., 2021; Bulut et al., 2021; Ganas et al., 2021; Sakkas, 2021) have contributed valuable insights about the coseismic behaviour of the October 30, 2020 Samos (Aegean Sea) earthquake by using geodetic data. In the study by Aktuğ et al. (2021), it was found that the mainshock did not cause on-land surface rupture, but the uniform slip model showed a 43.1 km long and 16 km wide rupture with 1.42 m slip along a north dipping normal fault. Ganas et al. (2021) pointed out that the earthquake caused the permanent uplift of the island (up to 10 cm), except for a coastal strip along part of the northern shore that subsided 2–6 cm. The earthquake also caused various geological effects such as liquefaction, rock falls, road cracks, and landslides due to strong ground motion and mobilization of soil cover and sediments. Sakkas (2021) identified the existence of a NNE–SSW extensional system with normal faults along an E–W direction according to the preseismic analysis and the horizontal and vertical displacements of approximately 350 and 90 mm, respectively, were found in the epicentral extensional region by coseismic analysis.

In this study, we presented the continuity of seismic activity in Samos Island and its surroundings by tracking the earthquake distributions up to October 2023. We considered the Global Navigation Satellite System (GNSS), Interferometric Synthetic Aperture Radar (InSAR), the Peak Ground Displacement (PGD), and horizontal displacement analysis results, for determining the displacements with the effect of the Samos earthquake. The GNSS data for the months between 9 months before and 8 months after the earthquake (the data between February 2020 and June 2021) were processed for interpreting the response of the ground to the earthquake. Then, the horizontal displacement due to this earthquake was calculated by using the focal mechanism values of KOERI by using Coulomb failure criteria. In the InSAR application part of the study, the interferograms were evaluated, and the line of sight (LOS) velocity was calculated by using the Sentinel-1A/B data. As for the seismological data, the PGD was obtained from 104 Strong Motion stations data recording the earthquake. Finally, the displacements calculated from the different data set were evaluated together.

2 Applications

In this section, we explain the analysis of GNSS data, horizontal displacement, InSAR data, and PGD for presenting the effects of the earthquake occurred on October 30, 2020 (M _w = 6.9) on Aegean Sea (offshore of Seferihisar district (Izmir), North of Samos Island).

2.1 GNSS data analysis

For investigating the displacements caused by the Samos earthquake, the GNSS data of 19 continuous stations and 4 campaign type sites were used. These are as follows: six continuous GNSS stations belong to Continuously Operating Reference Station-Turkey (CORS-TR) stations; AYVL (Ayvalık, Balıkesir), CESM (Çeşme, Izmir), DATC (Datça, Muğla), DIDI (Didim, Muğla), IZMI (Konak, Izmir), KIKA (Kırkağaç, Manisa), one continuous station of Continuously Operating Reference Station (CORS) station; MNTS (Urla, Izmir), two temporary (campaign type) GNSS stations of General Directorate of Mapping; YAM2 (Karşıyaka, Izmir), SFRH (Seferihisar, Izmir), 11 continuous GNSS stations of METRICA S.A. HxGN SmartNet of Greece commercial network company located on Greek Islands; PRKV, LESV (Lesvos), KALY, KAL1 (Kalimnos), IKAR, IKA1 (Ikaria), SAMO (Samos), CHIO (Chios), MYKN (Mykonos), NAXO (Naxos), ASTY (Astypalaia), and 1 continuous GNSS station of Dokuz Eylul University station; DEUG (Buca, Izmir) were used and also two temporary stations (DU12 (Urla, Izmir) and DU20 (Ozdere, Izmir) were observed under the rapid financial support of The Scientific and Technology Research Council of Turkey (TUBITAK), and the locations of the stations are given in Figures 3 and 4.

Figure 3

Coseismic displacements of GNSS stations with the effects of October 30, 2020 Samos (Aegen Sea) earthquake. Beach ball represents the epicenter location.

Figure 4

The displacements of GNSS stations based on the coseismic and postseismic effects. Beach ball represents the epicenter location.

As the first step of this application, the coseismic displacements due to the October 30, 2020 earthquake were calculated for 19 continuous GNSS stations: AYVL, CESM, DATC, DIDI, IZMI, KIKA, MNTS, DEUG, PRKV, LESV, KALY, KAL1, IKAR, IKA1, SAMO, CHIO, MYKN, NAXO, ASTY, shown in Figure 3 and Table 2. The GNSS data were processed by using Gamit/Globk software v.10.71 (Herring et al., 2015), and the GNSS data analysis strategy is given in Table 3.

Table 2

Displacements of the GNSS stations based on coseismic effects

Site	Longitude (°)	Latitude (°)	Displacement (mm)
			East	North
ASTY	26.353322	36.545131	−2.9	−6.0
AYVL	26.6861	39.3114	−0.8	5.1
CESM	26.372570	38.303815	−11.5	51.0
CHIO	26.127168	38.367912	−7.5	20.8
DATC	27.691836	36.708570	−0.2	−3.5
DEUG	27.194238	38.375095	12.4	28.5
DIDI	27.268661	37.372130	2.3	−15.8
IKAR	26.224234	37.628204	−8.0	−33.2
IKA1	26.273341	37.605420	−15.1	−46.6
IZMI	27.081822	38.394808	12.7	31.3
KALY	26.976152	36.955801	0.9	−12.6
KAL1	26.961662	36.962370	3.0	−14.6
KIKA	27.672206	39.105991	0.9	3.8
LESV	26.553789	39.100083	−1.5	8.6
MNTS	26.717429	38.426585	5.6	44.6
MYKN	25.329069	37.441645	−0.1	−2.4
NAXO	25.381169	37.09819	−3.2	−1.6
PRKV	26.265000	39.245702	−0.8	5.4
SAMO	26.705335	37.792767	−59.9	−374.7

Table 3

GNSS data processing strategy

Software	Gamit/Globk Version 10.71
Sampling of the GPS data	30 seconds/24 hours daily data
Cutting angle	10°
Ephemeris information	IGS final orbits and IGS ERP files
Antenna Phase Centre information	Weighted phase centre model related to the height angle (PCV-antmod.dat)
Troposphere parameter	VMF1 (Vienna Mapping Function) were used. Zenith delay parameters were calculated for each 2 h
International Terrestrial Reference System	ITRF 2014
Fixed stations	13 IGS stations; BUCU, DYNG, GLSV, ISBA, MATE, NICO, MIKL, NOT1, PENC, RAMO, SOFI, ZECK, WTZR
Final coordinate computation	GNSS data were combined by GLOBK

In the second step, the GNSS data between January 1, 2020, and June 30, 2021 (9 months before and 8 months after the Samos earthquake) were collected, and the continuous GNSS stations of CORS-Tr, DEU, and METRICA S.A. HxGN SmartNet of Greece network were processed for investigating the displacements due to the coseismic and postseismic effects. Moreover, after the Samos earthquake, the GNSS measurements were performed at YAM2, SFRH, DU12, and DU20 in 2020 (after the mainshock) and 2021. These observations were evaluated together with the previous data of these stations (YAM2: 2014-2016-2017-2018-2021; SFRH: 2014-2016-2018-2020-2021; DU12: 2009-2010-2011-2020-2021; DU20: 2009-2010-2011-2018-2020-2021) and are shown in Figure 4. In addition, time series of the GNSS stations were created for investigating the displacements on the vertical and horizontal components. The time series of Greek Islands’ stations (CHIO, IKAR, KALY, LESV, PRKV, and SAMO) belonging to the days between January 1, 2020, and June 30, 2021, is shown in Figure 5a–f. In addition, the time series of AYVL, CESM, DATC, DIDI, IZMI, KIKA, and MNTS belonging to the days between January 1, 2020, and January 1, 2022, and the time series of DEUG for the days between January 1, 2020, and June 30, 2022, are presented in Figure 6a–h.

Figure 5

Time series of the Greek Islands’ Stations: (a) CHIO, (b) IKAR, (c) KALY, (d) LESV, (e) PRKV, and (f) SAMO. The date of Samos earthquake (October 30, 2020) shown at 304th day of 2020 at horizontal axes of time series.

Figure 6

Time series of CORS-Tr: (a) AYVL, (b) CESM, (c) DATC, (d) DIDI, (e) IZMI, (f) KIKA), (g) DEUG (the continuous station of Dokuz Eylul University), and (h) MNTS (CORS continuous station).

As the last step of GNSS application, postseismic displacements were estimated via specification of an exponential fit model to the time of the earthquake by using equation (1) (Herring et al., 2015).

The exponential model is expressed as follows:

(1) ∑ i = 1 n e A i e 1 − e − t − t i e τ i e ,

where τ i e is relaxation time (days) of the ith exponential term; t i e is the time of the earthquake corresponding to the ith exponential term; A i e is the amplitude of the ith exponential term; and n e is the number of exponential terms of the model. The stations SAMO, IKAR, MNTS, IZMI, CESM, CHIO, and LESV which represented postseismic displacements are shown in Figure 7a. The total post seismic displacement of SAMO, which is the most affected station, is shown in Figure 7b.

Figure 7

(a) Postseismic displacements of GNSS stations with the effect of October 30, 2020 Samos earthquake. Beach ball represents the epicenter location. (b) Total postseismic displacement of SAMO station for North and East components. Since there is no displacement in Up component, it is not shown here.

2.2 Horizontal displacement analysis by using earthquake sources and Coulomb parameters

An earthquake deforms the crust and causes the stress changing on faults by affecting their locations, geometries, and slips (Toda et al., 2011). In previous studies (e.g. King et al., 1994; Ganas et al., 2006; Toda et al., 2011; Çırmık et al., 2017; Yildiz et al., 2021; Chousianitis and Konca, 2021; Kiratzi et al., 2021; Över et al., 2021; Sboras et al., 2021, Utku, 2022, Lentas et al., 2022), the stress changes caused by earthquakes were analyzed by using the Coulomb failure criterion with Coulomb software v3.4 (Toda et al., 2011). Coulomb software is specifically crafted to examine alterations in Coulomb stress concerning the mapped faults and nodal planes of the earthquake. The static displacements (on various surfaces or at GNSS stations), strains, and stresses induced by activities such as fault slip, magmatic intrusion, or dike expansion can be computed by this software. Coulomb addresses inquiries related to the impact of earthquakes on nearby faults, the role they play in either promoting or inhibiting failure, and the compression of a neighbouring magma chamber resulting from fault slip or dike expansion. The software is relevant to understanding geologic deformations associated with strike-slip faults, normal faults, or fault-bend folds. All calculations are conducted within an elastic halfspace, assuming uniform isotropic elastic properties as outlined by Okada (1992) (Toda et al., 2011).

In this section, the horizontal displacement occurred with the effect of the October 30, 2020 earthquake (M _w = 6.9) was calculated and presented in Figure 8. In the calculation, the nodal planes (Strike/Dip/Rake) of the mentioned earthquake presented by KOERI (2020) and top/bottom depths of the source fault (Samos fault) used as source parameters and presented in Table 4. In addition, the effective frictional coefficient, shear modulus (μ), and Poisson ratio were used as 0.4, 3 × 10¹⁰ Pa, and 0.25, respectively, in the calculation.

Figure 8

The horizontal displacement calculated by using Coulomb v3.4 software (Toda et al., 2011).

Table 4

Source parameters used as input for the Coulomb v3.4 software

The name of the institution	Fault top/bottom (km)	M w	Strike (°)	Dip (°)	Rake (°)
KOERI	0/15	6.9	272	55	−93

2.3 InSAR analysis

In the present day, radar techniques find applications in diverse fields such as civil, geodetic, and military uses. The synthetic aperture principle, developed in the 1950s and 1960s, opened up new possibilities, and with recent technological progress, interferometry with synthetic aperture radar (SAR) has become a powerful remote sensing tool with high quality and broad spatial coverage (Calvet et al., 2023). The advancements in satellite technology allow for precise identification of changes in the Earth’s structure over both spatial and temporal dimensions, owing to dynamic processes. InSAR is a microwave imaging technique and relies on analyzing the phase disparity between two intricate Synthetic Aperture Radar scans captured from identical geographical location but at distinct moments. This method furnishes valuable distance-related insights into the topography of the Earth (Nigussie and Eshagh, 2023).

InSAR survey was conducted on the Western Anatolia region in Turkiye using Sentinel-1 satellite images. Time-series were generated for the period January to November 2020.

InSAR is an effective method during the day and night to determine ground deformation using radar images. The Sentinel-1 provides data with a repeat interval of 6–12 days in Turkiye. Sentinel-1 interferograms were produced by the COMET LiCSAR automated processing system (https://comet.nerc.ac.uk/COMET-LiCS-portal/) (González et al., 2016; Lazecký et al., 2020). Atmospheric corrections which are a weather-based model and phase-elevation correlation were systematically applied to interferograms over the western of Anatolia. Generic Atmospheric Correction Online Service for InSAR (GACOS) (http://www.gacos.net/) was used for the weather-based model corrections. The method is established on the high-resolution European Centre for Medium-Range Weather Forecasts weather model and the Digital Elevation Model (SRTM).

With regard to calculating the quality assessment of interferograms, the quality factor (F) is calculated using the following formula:

(2) F = std i − std f std i × 100 ,

where std i is before correction and std f is the standard deviation of each interferogram after correction (Albino et al., 2020). The time-series of LOS ground displacements were obtained by carrying out a least-squares inversion method (Berardino et al., 2002; Usai, 2003; Schmidt and Bürgmann, 2003).

(3) D = ( A T A ) − 1 A T L ,

where D is the ground displacements and A is the temporal distribution of SAR scenes. S and H are 1 × N vector which includes the dates of SAR scenes and 2 × M matrix includes the beginning and ending dates for each interferogram, respectively. Also, L is the InSAR data containing the pixel values of each interferogram. Moreover, the LOS velocity was calculated using a weighted averaging method by the following formula (Le Mouélic et al., 2005):

(4) V = ∑ i = 0 N − 1 ∆ t i × ∆ c i ∑ i = 0 N − 1 ∆ t i 2 ,

where ∆ t i is the time scale of interferograms and ∆ c i is the corrected phase values. In addition, east ( d e ) and up ( d u ) displacement components into the LOS direction are calculated using the following equations (Hanssen, 2001):

(5) d LOS = d e × cos ( θ inc ) − d A × sin ( θ inc ) ,

(6) d A = d e × cos ( ∅ h ) − d n × sin ( ∅ h ) ,

where d LOS is the LOS displacement, d A is the projection of the multiple horizontal components on the azimuth look direction, θ inc is the incidence angle (between 30^o and 46^o), and ∅ h is the azimuth angle (ascending: 350^o, descending: 190^o).

Figures 9 and 10 show the atmospheric correction result of the ascending interferogram on September 18, 2020, to October 18, 2020, and the descending interferogram on April 9–21, 2020, respectively. The quality assessment of the atmospheric corrections method was calculated using equation (2) and shown in Figure 11. LOS velocity and cumulative LOS displacement maps of ascending and descending interferograms were considered from the study by Le Mouélic et al. (2005) using equations (3) and (4) and presented in Figure 12. In addition, the east and up components shown in Figure 13 are calculated using equations (5) and (6) (Hanssen, 2021).

Figure 9

Atmospheric correction result of 18 September–18 October 2020 ascending interferogram: (a) uncorrected interferogram showing a strong negative LOS change over the Northern west part of study area, (b) atmospheric phase delay maps based on GACOS weather-model, (c) residual phase delay map after GACOS correction, (d) atmospheric phase delay maps based on phase-elevation correlation, (e) residual phase delay map after phase-elevation correction, (f) combination delay maps of GACOS, phase elevation method and detrend method to eliminate linear ramp, and (g) residual phase delay map after combined method.

Figure 10

Atmospheric correction result of 9–21 April 2020 descending interferogram: (a) uncorrected interferogram showing a strong positive LOS change over the North part of study area, (b) atmospheric phase delay maps based on GACOS weather model, (c) residual phase delay map after GACOS correction, (d) atmospheric phase delay maps based on phase-elevation correlation, (e) residual phase delay map after phase-elevation correction, (f) combination delay maps of GACOS, phase elevation method, and detrend method to eliminate linear ramp, and (g) residual phase delay map after combined method.

Figure 11

Quality assessment of the atmospheric corrections method: the reduction of the standard deviations after corrections (in percent) as a function of the initial standard deviations using: (a) only GACOS weather-model, (b) phase-elevation linear model, and (c) combined method (ascending (blue) and descending (red) interferograms with mean quality values of ascending and descending interferograms).

Figure 12

The LOS velocity maps of ascending and descending interferograms calculated from Le Mouelic et al. (2005): (a) ascending LOS velocity map, (b) descending LOS velocity map, (c) ascending Cumulative LOS displacement map, and (d) descending cumulative LOS displacement map (A, B, and C points, which are selected for comparison and the cross corresponds to the reference point).

Figure 13

East and Up components calculated from the study by Hanssen (2001): (a) East component and (b) up component (A, B, and C points, which are selected for comparison and the cross corresponds to the reference point).

2.4 PGD analysis

In this application, 104 Strong Motion stations data recordings on the October 30, 2020 (M _w = 6.9) earthquake records were accessed via http://tadas.afad.gov.tr. The acceleration records were converted into displacement records by applying the double integration over the horizontal components of three-component (NS-EW-UD) acceleration records. The PGD was determined from the horizontal component (NS-EW) of the displacement records for each station, and the maximum horizontal displacements, recorded during the October 30, 2020 Samos earthquake, were given in the NS and EW directions for the study area and given in Figure 14.

Figure 14

(a) PGD Map of NS direction and (b) PGD map of EW direction during 30 October 2020 M _w = 6.9 Samos Earthquake.

3 Results and discussion

The October 30, 2020 Samos earthquake is recorded in the literature as a strong and huge impacted earthquake in recent years. In this study, for analyzing the effects of this earthquake, we presented results of the GNSS, horizontal displacement, InSAR, and strong motion studies. In addition, the seismic activity of the region from October 29, 2020, up to October 20, 2023, is presented in Figure 2 by giving the epicenter distributions of the earthquakes (3 < M < 7) occurred. According to the distributions of the earthquake, it is seen that the seismicity has continued for 3 years, and the energy of the earthquake has not unloaded yet.

In this study, firstly, the coseismic displacements obtained by GNSS analysis were used for representing the effects of the October 30, 2020 Samos earthquake (M _w = 6.9). Figure 3 shows that the largest displacement is observed at SAMO (located in Samos Island) which is the closest station to the epicenter. In Table 2, it is seen that the displacements are 374.7 and 59.9 mm towards S and W, respectively, and the direction of the resultant displacement is towards SW at SAMO. According to Figure 3, the direction of the coseismic displacement is towards SW at IKAR and IKA1 (Ikaria Island), and the directions of the displacements at KALY, KAL1 (Kalimnos Island), and DIDI (Didim, Mugla) are towards S. The directions of the coseismic displacements at IZMI (Izmir City Center) and DEUG (Buca, Izmir) are towards NE, those at CESM (Çeşme, Izmir) and CHIO (Chios Island) are towards NW, and those at MNTS (Urla, Izmir) are towards approximately N. The displacements at CESM and MNTS, which are located to the north of the epicenter, are larger than the other northern stations (IZMI, DEUG, CHIO). A minor northward displacement occurred at the LESV (Lesvos Island), while a smaller displacement was observed at the PRKV station located on the same island. At MYKN, NAXO, ASTY, and DATC, which are located to the south of the epicenter, the displacements are small and the directions are approximately southwards. At AYVL, which is located to the north of epicenter and on the Aegean Sea coast, the displacement is small and the direction is approximately northwards. At KIKA, which is located far from the Aegean Sea cost, the displacement is small and NE directed.

In the second step of GNSS analysis, the coseismic and postseismic displacements were calculated. According to Figure 4 and Table 5, the largest displacements are seen at SAMO, the displacements are 424.2 and 69.6 mm towards S and W, respectively, and the direction is towards SW. SFRH is the other station with large displacement located to the North of the epicenter (Seferihisar district, Izmir). The direction of the displacement is NE, and the displacements are 101.1 mm towards N and 27.7 mm towards E. The other high displacement is seen at DU20 (Özdere, Izmir) located to the NE of the epicenter. The displacements are 44.1 mm towards N and 52 mm towards E, and the direction is towards NE. The eastern component is more dominant at DU20 than the other northern stations. The directions of the displacements are towards NE at IZMI (Izmir City Center) and DEUG (Buca district, Izmir), and the directions of the displacements are towards NW at CESM (Çeşme, Izmir) and CHIO (Chios Island). The directions are approximately towards N at MNTS (Urla district, Izmir) and DU12 (Urla, Izmir). At YAM2 (Karşıyaka district, Izmir), the displacement is NE and its value is lower than surrounding stations. Among the eastern stations of the epicenter, while the SW-directed displacements are seen at IKAR and IKA1 (Ikaria Island), the directions of the displacements are approximately towards S-SE at KALY, KAL1 (Kalimson), and DIDI. In Figure 3, it is seen that the displacements are low at MYKN, NAXO, ASTY, and DATC. For LESV, the direction is approximately N, and for PRKV, AYVL, and KIKA, the values of total (coseismic and postseismic) displacements are shown in Figure 4 similar as their coseismic displacements shown in Figure 3.

Figure 5a–f represent the time series of Greek Islands’ stations (CHIO, IKAR, KALY, LESV, PRKV, and SAMO) belonging to the days between January 1, 2020, to June 30, 2021. Therefore, the effects of the October 30, 2020 Samos earthquake (304th Day of year (Doy)) are seen in the North (N), East (E), and Up (U) components of the time series. Figure 5a shows that displacements of CHIO are 20 mm towards N seen on the N component and 5–10 mm towards W on the E component. In Figure 5b, the displacements of IKAR are approximately 30 mm towards S and 5–10 mm towards E. Figure 5c shows that the displacement of KALY is approximately 30 mm towards S on N component, and any displacement is not seen clearly on the E component. Figure 5d and e show that the displacements are approximately 10 and 5 mm towards N on the N components of LESV and PRKV, respectively, and no displacement is seen clearly on the E components of both stations. Figure 5f shows that the displacements of SAMO are approximately 400 mm towards S on the N component, 60 mm towards W on the E component, and 100 mm towards up on the U component. The displacement on Up components for all stations except SAMO is not seen clearly. In addition, according to the time series presented in Figure 5a–f, it can be said that permanent deformation occurred in the Greek Islands’ stations with the effect of the earthquake.

The time series of AYVL, CESM, DATC, DIDI, IZMI, KIKA, MNTS, and DEUG are shown in Figure 6a–h and Figure 6a, and approximately 5 mm displacement is seen towards N on the N component and any displacement is not seen clearly on the E component at AYVL. In Figure 6b, the displacements of CESM are approximately 50 mm towards N and 10 mm towards W. Figure 6c shows that some displacement is seen on the N component, while any displacement is not clearly seen on the E component at DATC. In Figure 6d, for DIDI, the displacements are approximately 15 mm towards S and 3–5 mm towards W. Figure 6e shows that the displacements are approximately 30 mm towards N and 10–15 mm towards E on the N and E components, respectively, at IZMI. In Figure 6f, the displacement is approximately 3–5 mm towards N on the N component, while no displacement is evident on the E component at KIKA. In Figure 6g, the displacements are approximately 30 mm towards N and 10–15 mm towards E at DEUG. In Figure 6h, the displacements are approximately 40–45 mm towards N and 5 mm towards E at MNTS. The displacement is not clearly seen on the Up components for all stations. Therefore, according to the time series presented in Figure 6a–h, it can be said that permanent deformation occurred in the stations located in Izmir and its surroundings with the effect of the earthquake.

In the last step of GNSS analysis, for investigating the behaviours of the study region after the earthquake, the postseismic displacements were obtained at seven GNSS stations: CESM, CHIO, IKAR, IZMI, LESV, MNTS, and SAMO. Figure 7 shows that the largest postseismic displacement is seen at SAMO and its direction is approximately towards SW and the displacement is approximately towards S at IKAR and its amplitude is lower than SAMO. The directions of the displacement at CESM, CHIO, IZMI, MNTS, and LESV located north of the earthquake epicenter are towards approximately N and NE.

The direction of the postseismic displacements is towards NE at MNTS and IZMI shown in Figure 7. Although the directions of the postseismic and coseismic displacements of these stations are similar, the amplitudes of the postseismic displacement are smaller than the coseismic displacements shown in Figure 3. Figure 7 shows that the postseismic displacement is not seen at DEUG, while IZMI has displacement, even though the locations of IZMI and DEUG are close to each other. The absence of deformation at DEUG can be related with the location type of the station because DEUG is located on the ground (on the bedrock) but IZMI is located on the roof of a state building. In Figure 7, the postseismic displacement is NE oriented at CESM, but the amplitude has decreased with respect to the coseismic displacement shown in Figure 3. While the displacement of CESM is NW oriented after removing the coseismic effects, it is seen that the E component becomes dominant in postseismic displacement. Similar features are seen at CHIO (Chios) and LESV (Lesvos). It is remarkable that CESM, CHIO, and LESV show displacements in the same direction. Although the amplitudes of the displacements are different for Samos Island and Ikaria Island, it can be seen that they show similar directions of the displacements during and after the earthquake, as shown in Figures 3 and 7. In Figure 3, it can be seen that DIDI, located on the shore of Aegean Sea in Didim district of Muğla province and KALY, located at Kalimnios Island and SW of DIDI, presents coseismic displacements with similar directions and amplitudes. The directions and the amplitudes of the coseismic displacements of IZMI (Izmir City Center) and DEUG (Buca, Izmir) are almost the same as shown in Figure 3. In Figure 4 which includes the coseismic and postseismic displacements, the directions of the displacements are towards NE at GNSS stations located in Izmir and its surroundings; MNTS, DU12, SFRH, IZMI, DEUG, YAM2, and DU20 except CESM. At CESM which is located in the western part of Izmir, there is NW-oriented and similar directional displacement with CHIO (Chios Island). We think that the Gülbahçe fault shown in Figure 1 can be the reason for different oriented displacements at CESM with respect to the eastern stations. At SFRH (Seferihisar) and DU20 (Özdere), although the directions of the displacements are NE, the East component is more dominant at DU20, as shown in Figure 4. The reason for the directional differences between SFRH and DU20 is thought to be related with the existence of the Orhanlı Fault zone, as shown in Figure 1. Moreover, if the displacement directions of SAMO, IKAR and KALY, DIDI, shown in Figures 3 and 4, are examined, the submarine faults may be the reason of the directional differences between Ikaria and Kalimnos Islands. Besides, the differences in the direction of the displacements at PRKV and LESV which are located on Lesvos Island may be explained by the existence of a fault between these stations.

In the second application, the horizontal displacements were calculated by using the focal mechanism solutions given by KOERI and Coulomb parameters in Coulomb v3.4 software and shown in Figure 8. Figure 8 shows that the horizontal displacements are through N–S; therefore, it is said that the propagation direction of earthquake energy is in the N–S direction. Moreover, substantial horizontal displacements are observed in areas near the earthquake’s epicenter, with the amplitudes of the displacements diminishing as the distance from the epicenter increases. When contrasting the displacements derived from this analysis and the GNSS data presented in Figure 3, it is evident that directions of the displacements align consistently.

In the InSAR application of the study, 77 ascending and 139 descending interferograms were used and GACOS, phase elevation, and linear trend atmospheric effects corrections were applied to the interferograms. Figure 9a shows the ascending interferogram from September 18, 2020, October 18, 2020, which has an overall std of 4.46 cm. Figure 9b shown that the atmospheric signal consists of a negative LOS change of −20 rad over the Gulf of Izmir and its surroundings and running from north east to south west of the interferogram. The corresponding model created utilizing GACOS is shown in Figure 9b, and the residual after correction is shown in Figure 9c. The F value is 52%, and the standard deviation after correction is 2.13 cm. The model generated using the phase-elevation approach is shown in Figure 9d, and the residual after correction is shown in Figure 9e. In this case, the F value is 1%, and the standard deviation after correction is 4.42 cm. We also test the F value of a combined method by applying first the GACOS correction, linear detrending, and the phase-elevation method as suggested by Albino et al. (2020) (Figure 9f and g). As a result, the standard error is 1.72 cm, and the quality factor is 61%.

Figure 10a shows the descending interferogram April 9–21, 2020, which has an overall standard deviation of 3.38 cm. Figure 10b shows that the atmospheric signal consists of high positive to negative LOS change from north to south of the interferogram. The corresponding model created using GACOS is shown in Figure 10b, and the residual after correction is shown in Figure 10c. The F value is 59%, and the standard deviation after correction is 1.39 cm. The model obtained using the phase elevation method is shown in Figure 10d, and the residual after correction is shown in Figure 10e. In this case, the F value is 0.1% and the standard deviation after correction is 3.37 cm. We also test the quality of a combined method performing first the GACOS corrections, linear detrending, and the phase-elevation method, as suggested by Albino et al. (2020) (Figure 10f and g). As a result, the standard error was 1.16 cm, and the quality factor was 66%.

The general efficiency of the atmospheric corrections is computed for all the processed interferograms of western Anatolia, as shown in Figure 11a–c. Figure 11a shows the results of implementing the GACOS weather-based model method to the interferograms. In total, 26% of the descending and 40% of the ascending interferograms were overcorrected. In addition, 47% of the ascending and 51% of the descending interferograms had 0 < F < 0.25; 20% of the ascending and 21% of the descending interferograms had 0.25 < F < 0.50; and 4% of the ascending and 2% of the descending interferograms had 0.5 < F < 0.75. In Figure 11b, 89% of the ascending and 94% of the descending interferograms had 0 < F < 0.10. Eleven percent of the ascending and 6% of the descending interferograms had 10 < F < 0.25. Figure 11c shows the results of applying the combined method; 25% of the ascending and 15% of the descending interferograms were overcorrected; 39% of the ascending and 49% of the descending interferograms had 0 < F < 0.25; 31% of the ascending and 32% of the descending interferograms had 0.25 < F < 0.50; and 5% of the ascending and descending interferograms had 0.50 < F < 0.75.

The interferograms between January 4 and November 11 were utilized, and the LOS velocity values were calculated for 312 days. In Figure 12a, high LOS velocity values were observed over Samos Island, the little islands located to the SE of Samos Island, and some parts of the coast in the study area. Low LOS velocity values (≥−10 cm) were observed along the Karaburun Belt and its eastern part in the results of ascending interferograms. It can be said that the results of the descending LOS velocity map appear less noisy than ascending results although high LOS values were observed over Samos Island (Figure 12b). In the cumulative LOS displacement maps, high displacement values were recognizable after the earthquake (Figure 12c), but no significant displacement was observed (Figure 12d).

d e and d u components were computed after LOS velocity maps, and the values are −10 to 10 cm, respectively (Figure 13). It is seen that uplift (≥10 cm) and westward movement (≥10 cm) can be recognizable in the d e and d u components in most of Samos Island and the little islands located to the SE of Samos Island. In addition, some parts of the coast show a westward movement, and small-scale subsidence could be seen at some locations of the inland area. Significantly, a westward movement (≥10 cm) was observed along the Karaburun Belt and its eastern part.

As shown in Figures 9 and 10, atmospheric effects in the interferograms before inversion were removed as much as possible after corrections. In Figure 11, the F value of the interferograms, which was mostly overcorrected, was above zero after corrections. Afterwards, the inversion was applied, and finally, the east and up components were obtained. For comparing the vertical and horizontal component values obtained from both InSAR and GNSS results, the east and up components of coseismic displacement are drawn in Figure 13. While 11.5 mm westward movement was observed from the GNSS result at the CESM station shown in Table 2, 28 mm was obtained from the InSAR result. At the DEUG station, 12.4 mm eastward movement was observed as a result of GNSS analysis, while this value was obtained as 62 mm as a result of InSAR. In the DIDI station, 2.3 and 21 mm movements to the east were observed in the GNSS and InSAR results, respectively. At the SAMO station, 69.9 and 164 mm movements to the west were observed from the GNSS and InSAR, respectively. In Figure 13b, according to the up values of the InSAR results, the uplift is seen in the Samos Island and the up value of SAMO is 94.1 mm in GNSS up values; therefore, the InSAR and GNSS values are found coherently in general.

In the study by Sakkas (2021), approximately 15 cm subsidence was observed in the northeast part of Samos Island, but in this study, in Figure 13b, 20.8 cm subsidence was observed in the A point which is located in the northeastern of the Island. In both studies, it can be said that the west of the island of Samos moves to the west, while the south-east of the island remains relatively inactive. Evelpidou et al. (2021) observed an uplift of 35 cm in the west of Samos in their field study, while Mavroulis et al. (2021) observed an uplift of around 30 cm and in our study, and the maximum value was around 22.7 cm with the InSAR result. This difference is thought to be due to correction differences in the InSAR study. The importance of applying atmospheric corrections on interferograms has been shown in the study by Dogru (2020), and the study shows the reduction of atmospheric effects using recent atmospheric methods. In addition, the changes in the quality value of the interferograms as a result of the corrections are also given in this study. In the InSAR results, many interferograms were used together, not a single track like previous studies.

In PGD analysis, the displacements were found in the range of 0.01–3.8 cm in the NS and EW directions (Figure 14). These peak displacements do not show a linear distribution. This situation is directly related to the geological structure of the locations wherein strong motion stations are located. For example, considering the global spread of the seismic wave motion, the peak displacements in the NS direction would be expected to decrease towards the north, reaching the maximum value on the coasts of Seferihisar, considering the epicenter of the earthquake and the station distribution. However, maximum PGD values were measured in the NS direction in Bornova (Izmir) and Menemen plains (Izmir), Aydın City Center, and Didim. The same nonlinear behaviour characteristics of the calculated displacements apply to the PGD values of the EW component. The measured maximum PGD values in the EW direction were measured in Izmir Bornova Plain and Aydın City Center.

In Figure 14, the results of the PGD values indicate that although most of the structural damage in the October 30, 2020 Samos earthquake occurred in the Izmir Bornova Plain, the geologic setting of Aydın City Center could cause high displacements, and, therefore, the damage risk of future earthquakes is high in Aydin City Center. Aydin City Center has a very similar geologic setting similar to the Bornova Plain. They are composed of alluvial low-strength soil layers and possible basin shapes. A similar approach was discussed by Akinci et al. (2021) and Cetin et al. (2022) in detail. In the previous study (Pamukçu and Malaliçi, 2018; DAUM, 2020), the postseismic deformation based on the GNSS data for the years 2009, 2010, 2011, and 2018 which represented the postseismic effects of Sığacık Bay Seferihisar (Izmir) earthquake (M = 5.9) (October 17–21, 2005) (KOERI, 2005) were analyzed. These results represented the preseismic effects of the Samos earthquake. The low-amplitude fields seen in the result of PGD application and the high-amplitude deformation fields obtained from the deformation analysis are compatible with each other. The deployment (more dense and preferably same location at Strong Motion) of GNSS stations is crucial for comparing two sets of data (strong motion and GNSS) and conducting post-earthquake analyses. In addition, determining the permanence levels (elastic, elastoplastic, and plastic) of deformations resulting from earthquakes will be highly useful for the preliminary estimation of potential damage that future earthquakes could cause.