Ugur Dogan

Istanbul's earthquake hot spots: Geodetic constraints on strain accumulation along faults in the Marmara seismic gap

During the past century, a series of predominantly westward migrating Mu2009>u20097 earthquakes broke an ~1000u2009km section of the North Anatolian Fault (NAF). The only major remaining “seismic gap” along the fault is under the Sea of Marmara (Main Marmara Fault (MMF)). We use 20u2009years of GPS observations to estimate strain accumulation on fault segments in the Marmara Sea seismic gap. We report the first direct observations of strain accumulation on the Princes Islands segment of the MMF, constraining the slip deficit rate to 10–15u2009mm/yr. In contrast, the central segment of the MMF that was thought to be the most likely location for the anticipated gap-filling earthquakes shows no evidence of strain accumulation, suggesting that fault motion is accommodated by fault creep. We conclude that the Princes Islands segment is most likely to generate the next Mu2009>u20097 earthquake along the Sea of Marmara segment of the NAF.

Onset of aseismic creep on major strike-slip faults

Time series analysis of spaceborne synthetic aperture radar (SAR) data, GPS measurements, and field observations reveal that the central section of the Izmit (Turkey) fault that slipped with a supershear rupture velocity in the A.D. 1999, M w 7.4, Izmit earthquake began creeping aseismically following the earthquake. Rapid initial postseismic afterslip decayed logarithmically with time and appears to have reached a steady rate comparable to the pre-earthquake full fault-crossing rate, suggesting that it may continue for decades and possibly until late in the earthquake cycle. If confirmed by future monitoring, these observations identify postseismic afterslip as a mechanism for initiating creep behavior along strike-slip faults. Long-term afterslip and/or creep has significant implications for earthquake cycle models, recurrence intervals of large earthquakes, and accordingly, seismic hazard estimation along mature strike-slip faults, in particular for Istanbul which is believed to lie adjacent to a seismic gap along the North Anatolian fault in the Sea of Marmara.

Postseismic deformation following the Mw 7.2, 23 October 2011 Van earthquake (Turkey): Evidence for aseismic fault reactivation

Geodetic measurements following the 23 October 2011, Mwu2009=u20097.2 Van (eastern Turkey) earthquake reveal that a fault splay on the footwall block of the coseismic thrust fault was reactivated and slipped aseismically for more than 1.5u2009years following the earthquake. Although long-lasting aseismic slip on coseismic ruptures has been documented following many large earthquakes, long-lasting, triggered slip on neighboring faults that did not rupture during the earthquake has not been reported previously. Elastic dislocation and Coulomb stress modeling indicate that the postseismic deformation can be adequately explained by shallow slip on both the coseismic and splay fault and is likely driven mostly by coseismic stress changes. Thus, the slip deficit on the shallow section of the coseismic fault indicated by interferometric synthetic aperture radar-based models has been partially filled by aseismic slip, suggesting a lower likelihood for a large earthquake on the shallow section of the Van fault than suggested by previous studies.

Kinematics of landslide estimated by repeated GPS measurements in the Avcilar region of Istanbul, Turkey

The aim of this study is to analyze the spatial and temporal behavior of the landslide located in Avcilar region which is situated between Kucukcekmece and Buyukcekmece Lakes in the north-west of Marmara region, Turkey. A network consisting of 10 sites has been surveyed four times from November 2007 to May 2009 using Global Positioning System (GPS). The deformation analysis has been applied to determine the landslide movement parameters of the sites using GPS measurements of the four epochs. The reliable and high precision deformation rates are presented in terms of displacement vectors, velocity vectors and changes of accumulated strain. Landslides of the region are characterized from a regional GPS network. Each site has statistically different temporal behavior and significant relative motions and the region has irregular landslide movements.

Accuracy Analysis of Relative Positions of Permanent GPS Stations in the Marmara Region, Turkey

Abstract The accuracy of GPS (Global Positioning System) derived relative positions of stations depends on several factors. Besides the baseline length and duration of observation session, the methodology and the software used influence the results. In this paper, the observations made in the Marmara Continuous GPS Network (MAGNET) have been analysed to determine how the accuracy of derived relative positions of GPS stations depends on the baseline length and the duration of the observing session. Seven days of GPS observations in the MAGNET collected in 2002 were processed in the ITRF 2000 reference frame using Bernese 4.2 software. The baseline length varies between 6 km and 340 km, the session duration varies between 4h and 24h. The independent baseline components have been analysed.

Comparing GPS positioning errors derived from GAMIT/GLOBK and Bernese GNSS software packages: A case study in CORS-TR in Turkey

Global Navigation Satellite System (GNSS) processing software packages such as GAMIT/GLOBK, Bernese, GIPSY/OASIS etc., have an important role in high-precision GNSS studies. For a single GNSS session, they result in Formal Error (FE) rather than positioning accuracy because it may not be possible to incorporate the correlations in the GNSS data properly into the stochastic model of the estimation procedure. Therefore, a GNSS network used for high-precision studies should be measured repeatedly in different sequential days to have repeatabilities of positions to be used for representing the RMS errors of positioning. This contribution investigates the relationship between the FE and RMS error of Global Positioning System (GPS) positioning for GAMIT/GLOBK (V10.6) and Bernese (V5.2) software packages. For this purpose, seven days of GPS observations in six different sub-networks consisting of totally 36 Continuously Operating Reference Station (CORS) network stations in Turkey are used. The sub-networks were selected such that they were distributed homogenously throughout Turkey. The GPS observations were processed in the ITRF2008 reference frame using both GAMIT/GLOBK and Bernese software packages. Daily repeatabilities of North, East and Up coordinates of each station obtained from two software packages are nearly the same. However, their FEs are very different. GAMIT/GLOBK gives bigger FEs while Bernese results in too smaller FEs compared to their RMS errors of positioning. A linear relationship exists between the average of seven days FEs and the RMS errors for both software packages. This relationship is defined as the ‘Scale Factor (SF) = RMS/FE’. It is shown that a unique SF for the corresponding software package can be given as the mean value of the SFs associated with all points used in the study. According to our analysis, the mean SF in the region is 0.38 and 3.82 for GAMIT/GLOBK and Bernese, respectively.

Evidence for Tear Faulting from New Constraints of the 23 October 2011 Mw 7.1 Van, Turkey, Earthquake Based on InSAR, GPS, Coastal Uplift, and Field ObservationsEvidence for Tear Faulting from New Constraints of the 2011 Van, Turkey, Earthquake

The 23 October 2011 Mw 7.1 Van earthquake in eastern Turkey took place on a previously unknown thrust fault, causing over 600 fatalities and extensive damage mainly in Van city and the northern town of Erciş. Several coseismic models have already been published after the earthquake using the available seismological and geodetic data. These studies generally agree on a coseismic rupture on a north-dipping east–west-trending blind-thrust fault comprising one or two subparallel segments and the existence of a shallow slip deficit. Here, by introducing an additional coseismic pixel-offset dataset from the TerraSAR-X satellite, we conclude that a NNE–SSWtrending left-lateral tear fault that bounds the rupture to the east also slipped both during and after the mainshock, which is in agreement with both field observations and postseismic Global Positioning System (GPS) measurements. This has important implications for the seismic hazard analysis of the region as the extent and geometry of the rupture that we infer differ significantly from those previously proposed. Electronic Supplement: Figures showing comparison of the descending-mode Envisat Interferometric Synthetic Aperture Radar (InSAR) data and fault parameter trade-offs.

Ability of GPS PPP in 2D deformation analysis with respect to GPS network solution

GNSS networks play an important role in monitoring the displacements, movements and deformations of the Earth’s crust and engineering buildings. In this study, we examine how GPS Precise Point Positioning (PPP) is able to determine the horizontal deformations with respect to the GPS network solution. For this purpose, 7 days data of 12 Continuously Operating Reference Stations (CORS) in Turkey (CORS-TR), located in the western part of Turkey, are considered. The Bernese (v5.2)-derived coordinates over 7 days and the ones from four free online PPP services (CSRS, GAPS, APPS, Magic-PPP) are compared using the Bursa-Wolf coordinate transformation model. The errors from these transformations are used to define the RMS values of the PPP solutions in the local coordinate system. These values are relative to the GPS network solution. This fact leads to analysing how the PPP solutions are able to determine the horizontal deformations with respect to the network solution. From many experiments, in which the displacements belonging to the PPP solutions are simulated relative to the network solution, it has been shown that several ppm extensions or contractions may be determined using the free online PPP services. Therefore, we conclude that the online PPP services studied here may be used in 2D deformation studies as an alternative to the GPS network solutions.

Establishment of a gravity calibration baseline with the constrain of absolute gravity measurements after 17 August 1999 Izmit earthquake in Marmara region, Turkey

We described the first results of an on-going study of absolute gravity changes after the 17 August 1999 Izmit earthquake in Marmara region. Repeated absolute gravity measurements were carried out six stations with an A10 absolute gravimeter from 2009 to 2011 in the region. Axa0gravimetric calibration baseline (of the range of about 415 milliGal (mGal), 1xa0mGal=10−5xa0ms−2) was established in the region for the purposes of the calibration of the relative gravimeters. The absolute gravity measurements, repeated twice a year (October, June), can resolve gravity changes with a precision better than 5 microGal (μGal)/yr interval.The absolute gravity stations indicate a statistically significant gravity decrease at one of the absolute gravity stations in the calibration baseline, but the other stations do not. Generally, the absolute gravity difference agrees well with the general trend of the relative gravity changes. In addition, the precise vertical gravity gradients with relative gravity measurements have been determined at the absolute gravity stations for the reduction of measured gravity to the benchmark. The gradients scatter around the nominal constant (3.086xa0μGal/cm) between 2.82–3.73xa0μGal/cm as a result of the local density anomalies. The difference from the nominal gradient causes the error around 7–14xa0μGal by applying the normal vertical gradient instead of the directly measured one.

DISCUSSION ON STANDARD AND ROBUST KALMAN FILTERS, USING POST-EARTHQUAKE DATA SET OF 17 AUGUST 1999 IZMIT EARTHQUAKE

Abstract We discuss on the performances of standard and robust Kalman filters to estimate the post-seismic deformation field produced by the 17 August 1999 lzmit, Turkey, earthquake. The data set covers a subset of the Marmara Continuous GPS (Global Positioning System) Network (MAGNET) and other available survey stations in the region. The results of the models suggest that the standard Kalman filter estimate is not optimal due to serious bias that occurs between epochs. The bias, however, can be detected and removed using a robust Kalman filter.