High-resolution N-S deformation of active normal faults in SW Turkey derived from Sentinel-1 InSAR time series

Abstract

<p>Active tectonics in south-western Turkey is dominated by rapid N-S extension at a rate of 22 mm/a (e.g. Aktug et al., 2009), which is mostly accommodated by several large E-W trending, graben-forming normal fault zones. Seismic activity of these fault zones appears to vary both spatially and temporally (e.g. Leptokaropoulos et al., 2013). Generally, Synthetic Aperture Radar interferometry (InSAR) is a useful technique to assess the recent deformation of fault zones and locate potentially creeping segments. However, as Sentinel-1 satellites orbit the Earth on approximately N-S directed tracks, line-of-sight (LOS) velocities are relatively insensitive to N-S deformation and therefore it can be a challenge to resolve deformation in this direction. With its rapid N-S extension, the SW-Anatolian graben system is a suitable study area to develop an approach to derive a tectonic N-S deformation signal from Sentinel-1 InSAR.</p> <p>We compute InSAR LOS velocities from Sentinel-1 data for all ascending and descending frames covering the study area. A least-squares inversion is used to decompose the LOS velocities into north, east and up components. To reduce the number of unknowns, we constrain the E-W component with interpolated GNSS velocities, so we effectively only invert for N-S and up components. Mathematically, the inversion requires at least two time series products to be solved, but given the low sensitivity of InSAR to N-S deformation, we use three Sentinel-1 scenes, with at least one from ascending and descending tracks to increase the accuracy. As a result, this approach is limited to regions where either two ascending or two descending tracks are overlapping, which fortunately covers most of the large grabens in Western Turkey. Using our new technique, we compute a smooth velocity field for all three components of motion (N-S, E-W and up-down) on a N-S swath crossing all major E-W-trending normal fault systems in the region, at a pixel resolution of about 100x100 m. With some improvements to come, we are able to calculate swath profiles displaying surface deformation across all fault zones. Our approach resolves both the broad scale velocity field and localised deformation differences across individual fault zones.</p> <p>Compared to GNSS velocities, InSAR has a much higher resolution, allowing us to infer localised information on surface deformation in the vicinity of major fault zones instead of just quantifying a broad, regional trend. This can be used to assess individual fault zones, quantify changes in N-S surface deformation across faults and compare these results with recorded seismicity to reveal detailed insights into the active deformation of the largest fault zones in the region. Once the technique is established, we aim to expand the studied region. This study shows that overlapping tracks of Sentinel-1 data are a valuable resource, enabling detailed analysis of fault zones that are otherwise hard to assess by InSAR data from N-S orbiting satellite systems.</p> <p>References:</p> <p>Aktug et al. (2009). <em>Journal of Geophysical Research</em>, <em>114</em>(B10), B10404. https://doi.org/10.1029/2008JB006000</p> <p>Leptokaropoulos et al. (2013) <em>Bulletin of the Seismological Society of America</em>, <em>103</em>(5), 2739–2751. https://doi.org/10.1785/0120120174</p>

Constraining Historical Earthquake Sequences With Coulomb Stress Models: An Example From Western Türkiye

“…The Coulomb 3.4 Software (Toda et al, 2011) used to calculate Coulomb stress transfer is available from the USGS via https://www.usgs.gov/node/279387. The 3D-Faults software (Diercks et al, 2023), which is used to build the 3D-Fault network and slip distribution, is released at https://doi.org/10.5281/zenodo.7428878, future updates will be made available at https://github.com/MDiercks/3D-faults. significantly improved the quality of the manuscript.…”