Geodetic observations of postseismic creep in the decade after the 1999 Izmit earthquake, Turkey: Implications for a shallow slip deficit

Abstract

The relationship between aseismic slip and tectonic loading is important for understanding both the pattern of strain accumulation along a fault and its ability to generate large earthquakes. We investigate the spatial distribution and temporal evolution of aseismic creep on the western North Anatolian Fault (NAF) using time series analysis of Envisat interferometric synthetic aperture radar (InSAR) data, covering the full extent of the 1999 Izmit and Düzce earthquake ruptures and spanning 2002–2010. Discontinuities in the line‐of‐sight velocity across the fault imply that fault creep reaches the Earth's surface at an average fault‐parallel rate of ∼5 mm/yr along an ∼80 km section of the NAF. By combining InSAR and published GPS velocities, we are able to extract the vertical and east‐west components of motion and show that the Adapazari basin is subsiding at a rate of ∼6 mm/yr. Vertical motions have biased previous estimates of creep in this region. The displacement time series close to the fault is consistent with an afterslip model based on rate‐and‐state friction, which predicts a rapid deceleration in fault creep rate after the Izmit earthquake to a near‐steady state ∼5 mm/yr after 5 years. Projecting our model 200 years into the future we find that the cumulative displacement of 1–1.3 m is insufficient to account for the shallow coseismic slip deficit observed in previous studies. Distributed off‐fault deformation in the shallow crust or transient episodes of faster slip are likely required to release some of the long‐term strain during the earthquake cycle.

Interseismic strain accumulation across the central North Anatolian Fault from iteratively unwrapped InSAR measurements

“…Therefore, it is impossible to calculate the full 3‐D velocity field without a prior assumption. The common assumption made in previous studies is that there is no vertical motion across the region of interest [e.g., Walters et al , ; Hussain et al , ]. In our case we note that both the ascending and descending tracks are equally insensitive to motion in the north‐south direction.…”

“…For every pixel where information from both ascending and descending geometries are available, we use equation to invert for the east‐west and vertical components of motion following the method described by Wright et al [] and Hussain et al [] while taking into account the local incidence angles: where D LOS is the LOS velocity, θ is the local radar incidence angle, α is the azimuth of the satellite heading vector, and [ D E , D N , D U ] T is a vector with the east, north, and vertical components of motion, respectively.…”

“…However, profile B‐B′ crosses the creeping section of the fault. For this profile we model the fault‐parallel velocity as a combination of two signals: a long wavelength signal that represents interseismic loading at rate S and locking depth d 1 , and a short wavelength signal that represents the fault creep at a rate C from the surface down to depth d 2 [e.g., Wright et al , ; Elliott et al , ; Hussain et al , ]. where is the Heaviside function.…”

“…We find best fit values for each model parameter ( S , d 1 , C , and d 2 ) and an offset a , using a Bayesian approach, implementing the Goodman and Weare [] affine‐invariant ensemble Markov chain Monte Carlo (MCMC) sampler while accounting for the covariance. For details see Hussain et al [].…”

“…In order to understand the role that the NAF plays in regional tectonics and seismic hazard, there have been numerous estimates of the fault slip rate for the NAF using present‐day deformation measured with Global Navigation Satellite Systems (GNSS) [e.g., Straub et al , ; Reilinger et al , ; Ergintav et al , ] or offset geological features [e.g., Hubert‐Ferrari et al , ; Pucci et al , ; Kozaci et al , ]. There have also been several interferometric synthetic aperture radar (InSAR)‐derived estimates of the fault slip rate, which have focused on the western or eastern regions of the NAF where the InSAR coherence is better [e.g., Wright et al , ; Cakir et al , ; Walters et al , ; Kaneko et al , ; Cakir et al , ; Cetin et al , ; Walters et al , ; Cavalié and Jónsson , ; Hussain et al , ].…”