Surface deformation and slip distribution for the 2023 Mw 7.8 Pazarcık and Mw 7.5 Elbistan, Türkiye, earthquakes from subpixel correlation of optical images

The data contained in this Data Release depicts surface deformation and surface slip distribution for the 2023 Mw 7.8 Pazarcık and 7.5 Elbistan, Türkiye, earthquakes derived from subpixel correlation of Planet Dove and Sentinel-2 (S2) optical satellite images. The datasets contained in this data release are (1) raster files of surface displacement fields in north-south and east-west reference frames, (2) point measurements of slip distribution along the East Anatolian fault (EAF) and Çardak fault surface ruptures derived from the Planet Dove subpixel correlation fields, and (3) point measurements of surface slip distribution, manual offset measurements, and rupture width measurements for a 30-km-long study region along the EAF surface rupture where the Narlı fault rupture projects towards the EAF. Details of each dataset and names of the data files are provided below in the “Explanation of Data” section.

 
Explanation of Data:
 
PlanetDove subpixel correlation for both earthquakes
Surface displacement fields (3 m resolution), provided in north-south and east-west components, derived from subpixel correlation of Planet Dove optical satellite images. Fifteen scene pairs (see planet_scene_list.csv) were downloaded as orthorectified RGB (red-green-blue) images and processed individually. Pre-earthquake scenes span January 5 to February 3, 2023, and post-earthquake scenes span February 8 to March 27, 2023. Repeat time ranges from 6 to 76 days with an average of 44.2 days. Scenes were selected manually from near-daily collections based on ground visibility (lack of snow and clouds), pixel quality, collection time of day, and satellite path. Correlation was performed with MicMac (Rosu and others, 2015; Rupnik and others, 2017) using the MM2dPosSism command with a 5x5 pixel correlation window, regularization at native resolution, and minimum correlation threshold of 0.4. MicMac output files were converted to meters with north- and east-positive reference frames. North-south results were destriped by rotating each scene until the stripes were horizontal, subtracting the mean value of each row, and rotating the scene back to its original orientation. All fifteen scenes were then merged, forming one east-west and one north-south raster, and clipped to a 5-km buffer around the surface rupture. Planet images © 2023 Planet Labs PBC accessed via the U.S. Government Plus (USG+) license. The raster datasets are provided in .tif format in WGS84 UTM Zone 37N (epsg=32637). They are best visualized with a diverging color scale limited to +/- ≤7 m.
 
Data Files:
planet.zip
-       planet_EW.tif
-       planet_NS.tif
-       planet_scene_list.csv
 
Sentinel-2 subpixel correlation for a portion of the Pazarcık earthquake:
Surface displacement fields (10 m resolution), provided in north-south and east-west components, derived from subpixel correlation of Sentinel-2 (10 m) optical satellite images that span the 30-km-long detailed study region. To mitigate decorrelation issues, images were chosen one year apart (April 25, 2022 and 2023) to reduce snow, sun angle, and vegetation differences. The image pair was clipped to its overlapping extent and processed with MicMac (Rosu and others, 2015; Rupnik and others, 2017) using the MM2dPosSism command with default values of 9x9 pixel correlation window, regularization at 4x native resolution, and 0.5 minimum correlation threshold. MicMac output files were converted to meters with north- and east-positive reference frames. Sentinel-2 (S2) observations are Copernicus Sentinel data processed by the European Space Agency. The raster datasets are provided in .tif format in WGS84 UTM Zone 37N (epsg=32637). They are best visualized with a diverging color scale limited to +/- ≤7 m.
 
Data Files:
S2.zip
-       S2_EW.tif
-       S2_NS.tif
 
Slip distribution from Planet Dove subpixel correlation:
To construct a surface slip distribution, we extract 45-m-wide (15 pixels) and 5-km-long profiles that are normal to local fault strike and spaced 50 m apart, for a total of 6270 displacement measurements from the Pazarcık earthquake (“planet_eaf_slip_distribution.csv”) and 2800 displacement measurements from the Elbistan earthquake (“planet_cardak_slip_distribution.csv”). We use the simple fault from Reitman and others (2023) to seed the profiles and calculate displacement by projecting the data in each profile into fault-parallel and fault-perpendicular orientations. We then take linear regressions through the fault-parallel results that encompass 1-km and 2.5-km-long sections on either side of the fault zone, which was excluded. Displacement was also calculated based on the mean of the data on each side of the fault zone in 1-km and 2.5-km-long sections, excluding the fault zone. The fault zone was chosen manually for each profile guided by data derivatives (mean, moving mean, gradient) to evaluate where the signal begins to be influenced by deformation. The four measurement methods (means and linear regressions for the 1-km and 2.5-km profiles) result in similar, though slightly different slip distributions that overlap within uncertainty defined by the standard error of the fit parameters. The slip distribution presented here is the mean of the four measurements, with uncertainty bounds defined by the range of the four measurements. Data are provided in eaf_slip_distribution.csv and cardak_slip_distribution.csv, in which column titles are latitude (in degrees, WGS84), longitude (in degrees, WGS84), utm_x_m (in meters, zone 37N), utm_y_m (in meters, zone 37N), dist_eaf_km or dist_cardak_km (distance along the EAF or Çardak surface rupture in km), horiz_offset_m (fault-parallel horizontal displacement in meters), offset_min_m (fault-parallel horizontal displacement minimum in meters), offset_max_m (fault-parallel horizontal displacement maximum in meters). WGS84 is EPSG 4326 and the UTM reference frame is EPSG 32637. Data are also provided for the Gölbaşı splay in planet_golbasi_slip_distribution.csv and for the Narlı fault rupture in planet_narli_slip_distribution.csv with the same columns. For the Narlı and Gölbaşı datasets, the column “dist_eaf_km” is the projected distance along the EAF surface rupture.
 
Data Files:
planet_slip_distribution.zip
-       planet_eaf_slip_distribution.csv
-       planet_cardak_slip_distribution.csv
-       planet_narli_slip_distribution.csv
-       planet_golbasi_slip_distribution.csv
 
Slip distribution from Sentinel-2 subpixel correlation:
Surface slip distribution derived from the Sentinel-2 subpixel correlation fields is calculated via the same method as the Planet subpixel correlation slip distribution, except that profiles are 135-m-wide (15 pixels) and 135 m apart, for a total of 214 unique measurements along the EAF surface rupture in the 30-km-long study region. Data are provided in S2_slip_distribution.csv, in which column titles are latitude (in degrees, WGS84), longitude (in degrees, WGS84), utm_x_m (in meters, zone 37N), utm_y_m (in meters, zone 37N), dist_eaf_km (distance along the EAF surface rupture in km), horiz_offset_m (fault-parallel horizontal displacement in meters), offset_min_m (fault-parallel horizontal displacement minimum in meters), and offset_max_m (fault-parallel horizontal displacement maximum in meters). WGS84 is EPSG 4326 and the UTM reference frame is EPSG 32637.
 
Data Files:
S2_slip_distribution.csv
 
Manual offset measurements from WorldView Images:
We manually measured offset features remotely on post-earthquake WorldView images (≤0.5 m resolution) from February 11 and February 28, 2023, with 39 total offset measurements on the EAF surface rupture in the 30-km-long study area. For each measurement, we project each side of an offset linear feature into the fault zone, with alternative projections providing a 95% uncertainty range (for example, DuRoss and others, 2024). We then calculate displacement as the distance between each projection along the mapped fault trace, which was done automatically after all projections had been made to increase objectivity. We consulted the feature’s pre-earthquake geometry on pre-earthquake images from WorldView if available, or on Microsoft Bing and/or Google Earth satellite images. Data are provided in manual_offsets.csv, in which column titles are latitude (in degrees, WGS84), longitude (in degrees, WGS84), utm_x_m (in meters, zone 37N), utm_y_m (in meters, zone 37N), dist_eaf_km (distance along the EAF surface rupture in km), horiz_offset_m (fault-parallel horizontal displacement in meters), offset_min_m (fault-parallel horizontal displacement minimum in meters), offset_max_m (fault-parallel horizontal displacement maximum in meters), and partial (indicating whether the offset measurement encompasses the entire fault zone or only a part of the fault zone, and thus represents a minimum measurement – yes=partial, no=total, unsure=unsure if partial or total measurement). WGS84 is EPSG 4326 and the UTM reference frame is EPSG 32637.
 
Data Files:
manual_offsets.csv
 
Rupture zone width:
We measured the width of visible surface rupture for the 30-km-long study region every 50 m where data was available, for a total of 565 measurements. Surface rupture width includes only localized, discrete displacement on fault strands, not diffuse deformation accommodated off primary or secondary fault traces that is captured by subpixel correlation.  Surface rupture width was measured perpendicular to local fault strike based on the width of visible surface rupture on primary and secondary faults in WorldView images and/or in the field. Field mapping (DuRoss and others, 2024) was used in areas where WorldView images did not record surface rupture, such as in places covered by forest, clouds, or snow. Data are provided in rupture_width.csv, in which column titles are latitude (in degrees in WGS84), longitude (in degrees in WGS84), utm_x_m (in meters, zone 37N), utm_y_m (in meters, zone 37N), dist_eaf_km (distance along the EAF surface rupture in km), rupture_width_m (rupture zone width in meters), and data_used (dataset(s) used to measure rupture width – wv=WorldView, ge=GoogleEarth, field=field mapping from DuRoss and others, 2024). WGS84 is EPSG 4326 and the UTM reference frame is EPSG 32637.
 
Data Files:
rupture_width.csv
 
DISCLAIMER:
Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

REFERENCES: 
DuRoss, C.B., Reitman, N.G., Hatem, A.E., Mason, H.B., Lavrentiadis, G., Milliner, C., Asimaki, D., Karakaş, M., and Seçen, B., 2024, Field Observations of Surface Rupture and Fault Displacement in the 2023 M7.8 Pazarcık, Türkiye Earthquake, U.S. Geological Survey data release: https://doi.org/10.5066/P969BHT4.
 
Reitman, N.G., Briggs, R.W., Barnhart, W.D., Thompson Jobe, J.A., DuRoss, C.B., Hatem, A.E., Gold, R.D., Akçiz, S., Koehler, R.D., Mejstrik, J.D., Collett, C., 2023, Fault rupture mapping of the 6 February 2023 Kahramanmaraş, Türkiye, earthquake sequence from satellite data (ver. 1.1, February 2024): U.S. Geological Survey data release, https://doi.org/10.5066/P985I7U2.
 
Rosu, A.M., Pierrot-Deseilligny, M., Delorme, A., Binet, R., and Klinger, Y., 2015, Measurement of ground displacement from optical satellite image correlation using the free open-source software MicMac: ISPRS Journal of Photogrammetry and Remote Sensing, v. 100, p. 48–59, https://doi.org/10.1016/j.isprsjprs.2014.03.002
 
Rupnik, E., Daakir, M., and Pierrot-Deseilligny, M., 2017, MicMac – A free, open-source solution for photogrammetry: Open Geospatial Data, Software and Standards, v. 2, no. 14, https://doi.org/10.1186/s40965-017-0027-2