Determination of 2011 Van/Turkey earthquake (M = 7.2) effects from measurements of CORS-TR network

Abstract

Turkey is a country that has active fault lines and is prone to earthquake. The positional variations resulted from the Van earthquake (a magnitude of M = 7.2) on 23 October 2011 and were calculated by using data from CORS-TR (Continuously Operating Reference Station network) stations. The GPS data from the CORS-TR stations effected from the earthquake were evaluated by the BERNESE 5.0 software. For calculating the positional displacements in the CORS-TR stations after the earthquake that occurred in Van, the measurements from 16 International GNSS Service (IGS) stations around Turkey, where their coordinates and velocity vectors were considered to be constant at the ITRF2005 datum. The greatest variation was measured from the Muradiye station, which is the closest point to the epicentre of the earthquake. The variation in this station was −42.7 mm in the vertical direction, −60.3 mm in the northern direction and −18.7 mm in the eastern direction.

1. Introduction

Determining the amount of variation at the point coordinates through the Global Satellite Navigation Positioning System (GNSS) network requires continuous observation to study the variations related to earthquake and tectonic displacements. Measuring the amount of displacement created by earthquakes is very important from the point of view in preventing great disasters in Turkey, where earthquake and the tectonic activities are intense.

The positional displacements pertaining to the Van earthquake that occurred in Turkey have been studied in this research (figure 1). In order to measure the tectonic displacements, and the positional displacements happened after the earthquake, the position data of the CORS-TR stations placed around the earthquake region were obtained for the day before, on the day and the day after the earthquake from the International GNSS Service (IGS) stations and were evaluated by using BERNESE 5.0 academic software.

Figure 1. Van earthquake region.

2. CORS-TR System

CORS-TR System is a network that comprises 146 stations in total, scattered in such a way that to cover Turkey and the North Cyprus Turkish Republic. The main control centre of the system has been set up in Ankara. The CORS-TR System became active in 2009 and maintains its activities in an uninterrupted manner. The system collects data from all of the stations at one-second intervals for 24 hours. The distribution of the CORS-TR stations is shown in figure 2. The layout of the stations was established by using the stations providing best data to observe the tectonic-plate displacements (Mekik et al. 2011; Uzel et al. 2011).

Figure 2. Distribution and fault lines of the CORS-TR stations.

3. Tectonic structure of the Van region

Turkey is an actively deforming region and subject to devastating earthquakes for long time. The earthquakes in Turkey are governed by three major geological structures: the Aegean-Cyprean Arc, the dextral North Anatolian Fault Zone and sinistral East Anatolian Fault Zone (Bozkurt 2001). These structures were originated from westward escape of Anatolian plate squeezed between two colliding plates, the Arabian and Eurasian plates. The abovementioned two fault zones meet and form a continental triple junction to the east of Karlıova in eastern Turkey (Bozkurt 2001). The area located at the east of the Karlıova triple junction is represented by N-S compressional tectonic regime in which there are several faults parallel to the North and East Anatolian fault zones (figure 4). These faults are seismically active and cause many earthquakes in the regions. Of these, the Ercis, Caldıran, Hasantimur Lake, Suphan and Malazgirt fault lines are the active faults in the region where the Van earthquake on 23 October 2011 occurred.

Figure 3. Distribution of the IGS stations used as reference.

Figure 4. CORS-TR stations affected by the Van earthquake.

4. Point-position calculations

To study the positional displacements in the CORS-TR stations for the earthquake that occurred in Van, we obtained data from 16 IGS stations around Turkey, where their coordinates and velocity vectors were considered as invariable at the ITRF2005 datum (figure 3).

The GPS data from the CORS-TR stations experiencing the earthquake were evaluated with BERNESE 5.0 software (Dach et al. 2011). In order to define the coordinates of CORS-TR stations where the GPS measurements have done in the ITRF system, measurements from 16 IGS (International GPS Service) stations having appropriate position in Europe and Asia were included in calculation.

The Bernese 5.0 software used in evaluation of GPS observations carried out in CORS-TR stations where earthquake occurred was realized in two steps.

In the first step: The CORS-TR stations coordinates have been calculated approximately in centimetres accuracy by applying PPP (Precise Point Positioning) method. During this stage, code and phase measurements obtained from CORS-TR stations and receiver-satellite clock errors (IGSwwwd.CLK) measured by IGS in 5-minute intervals were used. At the adjustment stage, the coordinates of each point have been calculated separately without having any network structure. Until the first network solution during the parameter estimation stage, the calculated coordinates have been taken as a priori for the evaluation of double difference measurements as being the second stage.

In the second step: Double difference measurements were evaluated. At this step, the IGS points at the date of measurements have also been added into the evaluation. The criteria used in the evaluation of double difference measurements have been given in table 1(a) and 1(b).

5. Van earthquake findings

An earthquake with a magnitude of M = 7.2 took place on 23 October 2011 at 01:41 pm local time in the Van region, Turkey. The epicentre of the earthquake was close to the northern part of central Van province (figure 4). The earthquake was felt in a wide area in East Anatolia including: Diyarbakır, Batman, Sırnak, Mus, Erzurum, Bingol, Bitlis, Siirt, Ağrı and Iğdır. The coordinates of the external epicentre of the earthquake were 38.7578 N–43.3602 E, in accordance with the records of the Kandilli Observatory (Van Earthquake 2011).

The data obtained from the CORS-TR stations (Ağrı, Baskale, Erzurum, Muradiye, Ozalp, Tatvan, Hakkari, Hınıs, Horasan, Iğdır and Malazgirt) in the Van region where the earthquake occurred between 13 and 29 October 2011, together with the positional variations, were examined (figure 4).

After examining the positional data measured before and after earthquake, positional displacements were observed in all stations included in evaluation. The Root Mean Square (RMS) values obtained from the CORS-TR station coordinates are given in table 2.

Table 1(a). Evaluation strategy of GPS measurements with the Bernese 5.0 software (model).

Measurement model
Pre-assessment	On the base-base mode, it has been performed by using phase measurements and triple difference method. Different linear combinations of L1 and L2 carrier phases have synchronously been examined and the cycle slips have been fixed. In cases where cycle slips could not be determined exactly, incorrect measurements have been removed or ambiguity has been added.
Basic measurements	Carrier phase measurements have been used. Code measurements were used for synchronization of receiver clock and GPS time only.
	Cut-off Angle: 10 degrees
	Data interval: 30 seconds
	Weighting: On zenith angle, for double difference measurements independent from ionosphere, 6 mm has been taken. Weighting function dependent from elevation angle has been taken as 1/cos (z) ^ 2.
Modelled measurements	Linear combinations of double difference measurements are considered independent from ionosphere.
GPS antenna phase centre calibration	Elevation angle-dependent phase centre corrections have been applied according to the IGS05.ATX model and during evaluation of measures absolute corrections were applied to the receiver antennas. The PHAS _COD.I05 file was used for receiver antennas. The SATELLIT.I05 file was used for GPS satellite antennas.
Troposphere	A priori model: Hydrostatic component has been modelled with dry-Niell correction function and the Saastamoinen model was applied.
	Meteorological data: Not used
	The unknowns of zenith delay have been calculated by using the wet-Niell projection function for each station in intervals of 2 hours.
	Constraints: 1 m has been defined as pre-condition for relative zenith delay values.
	Correction function: For both dry and wet components, the Niell correction function was used.
Ionosphere	It has not been modelled (first-degree influences have been eliminated by the combination of L1 and L2 carrying phase measurements independent from ionosphere). In addition, for the solution of initial phase of uncertainty (ambiguity), the global ionosphere values obtained from Orbit Determination Centre of Europe (CODE) by using GPS were used.
Earth rotation par.	IGS combined earth rotation parameters (final)
Orbit	IGS precise orbit (final)
Models	Earth geo-potential model: JGM3
	Planet ephemeris: JPL DE200
	Tide touring of solid earth: IERS 1996
	Ocean: GOT00
Atmospheric loading	Not applied.
Crustal movements	The velocities of IGS points have been taken in the ITRF2005 coordinate system.

Table 1(b). Evaluation strategy of GPS measurements with the Bernese 5.0 software (parameters).

Calculated parameters
Adjustment	The least squares method
Point coordinates	The ITRF05 coordinate system has been used by using point coordinates and velocities of the stations given during solution of IGS05.SNX.
Satellite clock errors	Satellite clock errors have been eliminated by the double differences method.
Receiver clock errors	Receiver clock errors have been calculated by using the pseudo-range measures during pre-assessment stage.
Criteria for removing from evaluation	The measurements exceeding 0.004 m from normalized residuals during pre-assessment phase have been removed from the evaluation.
Base selection	OBSMAX principle has been used.
Ambiguity	The QIF has been used.

Table 2. The RMS values of CORS-TR station coordinates.

CORS-TR station	RMS (mm)
	ΔEast (mm)	ΔNorth (mm)	ΔElevation (mm)
AGRD	3.3	12.3	6.1
ERZR	1.1	1.6	4.7
MURA	8.1	27.8	9.9
OZAL	1.3	5.3	7.4
TVAN	3.0	2.0	6.1
HAKK	1.8	8.5	5.3
HINI	0.9	1.5	4.7
HORS	1.8	3.0	5.6
IGIR	1.2	5.5	6.2
MALZ	1.4	6.8	4.4

In statistical theory, precision is measured by the standard deviation (also called standard error or root mean square – RMS). Theoretically (1) where RMS is the standard deviation of a very large sample, v is the true residual and n is the large sample size (Kavanagh & Bird 2000).

The greatest displacement was calculated as ΔNorth = 0.0286 m (figure 5(a)) in the N-S line and in the South direction for Ağrı (AGRD); as ΔElevation = 0.0176 m (figure 5(b)) for Erzurum (ERZR); as ΔNorth = 0.0603 m in the N-S line and in the South direction (figure 5(c)) for Muradiye (MURA); as ΔNorth = 0.0150 m and ΔElevation = 0.0291 m in the N-S line and in the North direction (figure 5(d)) for Özalp (OZAL); as ΔEast = 0.0085 m and ΔElevation = 0.0181 m, especially in the E-W line, West direction (figure 5(e)) for Tatvan (TVAN) where the displacements and the elevations before the earthquake were observed on a separate basis; as ΔEast = 0.021 m in the N-S line and in the North direction, contrary to the expectations (figure 5(f)) for Hakkari (HAKK) where activity in the elevation and displacement before the earthquake was observed on a separate basis; as ΔElevation = 0.0191 m (figure 5(g)) for Hınıs (HINI); as ΔElevation = 0.0208 m (figure 5(h)) for Horasan (HORA); as ΔElevation = 0.0207 m and ΔNorth = 0.0138 m in the N-S line and in the South direction (figure 5(i)) for Iğdır (IGIR); as ΔElevation = 0.0148 m and ΔNorth = 0.163 m in the N-S line in the South direction (figure 5(j)) for Malazgirt (MALZ).

Figure 5. (a) Ağrı CORS-TR station – the magnitude of displacement. (b) Erzurum CORS-TR station – the magnitude of displacement. (c) Muradiye CORS-TR station – the magnitude of displacement. (d) Özalp CORS-TR station – the magnitude of displacement. (e) Tatvan CORS-TR station – the magnitude of displacement. (f) Hakkari CORS-TR station – the magnitude of displacement. (g) Hınıs CORS-TR station – the magnitude of displacement. (h) Horasan CORS-TR station – the magnitude of displacement. (i) Iğdır CORS-TR station – the magnitude of displacement. (j) Malazgirt CORS-TR station – the magnitude of displacement.

6. Results and conclusion

Earthquakes like the one in the Van region in 2011 show that much more measurements are inevitable for the prevention and prediction of future earthquakes in countries such as Turkey, where activities on fault lines are intensive.

All countries have already established their CORS networks. The establishment of the Turkish CORS-TR network started in 2006 and was completed in 2009. It is one of the greatest networks in the world to date. The CORS-TR network can be used both for the real-time applications and also for post-process bureau calculations. One of the main purposes of the network is to focus on tectonics, like determining the plate displacements and forecasting the earthquakes. The CORS-TR network has, therefore, been designed by taking the tectonic structure of Turkey into account.

The magnitude of displacements resulted from the Van earthquake on 23 October 2011 was evaluated by using data from CORS-TR points in Ağrı, Ezurum, Muradiye, Özalp, Tatvan, Hakkari, Hınıs, Horasan, Malazgirt and Iğdır provinces. The data from these stations comprise measurements one week before and the day after earthquake. They were evaluated by taking into account the parameters given in table 1(a) and 1(b) and by using the data at the IGS points given in figure 3 with BERNESE 5.0 software. The maximum displacements for each CORS-TR stations were calculated and shown in figures 6–8. The greatest variation was measured from the Muradiye station, which is the closest point to the epicentre of the earthquake. The variation in this station was –42.7 mm in the vertical direction, –60.3 mm in the northern direction and −18.7 mm in the eastern direction.

Figure 6. Maximum changes in positions determined as a result of the earthquake.

Figure 7. North-East components maximum displacement vectors.

Figure 8. Elevation component maximum displacement vectors.

Acknowledgement

The authors would like to thank the editor and two anonymous reviewers for their valuable comments and suggestions to improve the manuscript. The authors also thank CORS-TR Center for providing data and Prof. Dr. Hükmü ORHAN for his valuable comments and suggestions. Figures 1–4 were plotted using the terrific software of Wessel and Smith (1995).