Sentinel-1 InSAR Surface Deformation Data over the Permian Basin

Abstract

This Sentinel-1 InSAR dataset contains surface deformation that occurred between Nov. 2014 and Jan. 2019 associated with the Permian Basin oil and gas production. For further details of the processing method and uncertainty analysis, please see the associated paper of Staniewicz et al., 2020. Note: click the tree viewing option to see proper organizational layout of files, not the table layout. When using this data for research, please cite: Staniewicz, S., Chen, J., Lee, H., Olson, J., Savvaidis, A., Reedy, R., et al. (2020). InSAR reveals complex surface deformation patterns over an 80,000 square kilometer oil-producing region in the Permian Basin. Geophysical Research Letters, 47, e2020GL090151. Details of generation and data attributes Two paths of Sentinel 1 data were used in the analysis: the ascending path 78 and the descending path 85. For each path, the cumulative radar line-of-sight (LOS) deformation between (1) Nov. 2014 and Jan. 2017; (2) Nov. 2014 and Jan. 2018; and (3) Nov. 2014 and Jan. 2019 are included. Here the pixel spacing for all InSAR grids is 120 meters. All deformation data units are in centimeters. Each of the maps' cumulative results have an uncertainty of ~ 1 cm or less. All the maps using ascending (or descending) Sentinel data are coregistered the same latitude/longitude grid as the digital elevation model (DEM) covering the ascending (or descending) path. Note: We used the SRTM DEM data to generate the interferograms. These DEM data can be found in geotiffs/ascending_path78/dem.tif and geotiffs/descending_path85/dem.tif. The units of the DEMs are in meters. For each path, we provided the names and locations of the GPS stations with continuous coverage between Nov. 2014 and Jan. 2019 as CSV files. The GPS east, north, and vertical daily time series are available through the Nevada Geodetic Laboratory (http://geodesy.unr.edu/ ). For this example, the NMHB station's NA plate-fixed solutions are available at http://geodesy.unr.edu/NGLStationPages/stations/NMHB.sta The file geotiffs/ascending_path78/gps_locations.csv contains the name, latitude, and longitude of the stations within the ascending path, as well as the row and column of that location within the ascending latitude/longitude grid. The GPS stations TXKM was used as the spatial reference point to calibrate all LOS InSAR maps, the rest of GPS stations were used as independent validations for the InSAR results. In addition to providing the data in GeoTIFF format, we have also loaded the data into MATLAB provded .mat files (located in the matlab_version/ folder). We have divided the .mat files into data coregistered on the ascending grid, the descending grid, and the vertical/east deformation solutions in the region where the ascending and descending paths overlap. The definition of the radar LOS direction We note that InSAR measures surface deformation along the radar LOS direction. In the region where the ascending and descending paths overlap, we decomposed the the two LOS deformation solutions (Nov. 2014 to Jan. 2019) using the ascending and descending LOS maps (unitless, as in geotiffs/ascending_path78/los_enu.tif and geotiffs/descending_path85/los_enu.tif) into their horizontal and vertical components. These vertical/horizontal solutions are contained in geotiffs/vertical_horizontal_decomposition/. Further details of the LOS decomposition can be found in the associated paper and supplement of Staniewicz et al., 2020. Converting GPS ENU data to the radar LOS Here we show an example of how you would convert GPS east, north, up (ENU) time series data into measurements comparable to the ascending LOS InSAR measurements using the LOS unit vector coefficients. We use station NMHB as an example, whose metadata is contained in the geotiffs/ascending_path78/gps_locations.csv file. The LOS vector coefficients are in the geotiffs/ascending_path78/los_enu.tif image (or path78_data.mat), which is a 3 band GeoTIFF containing the look vector coefficients. To extract the 3 LOS coefficients from the matrix los_enu we could do the following in MATLAB: load path78_data.mat r = gps_locations.row(1); c = gps_locations.col(1); enu_coeffs = los_enu(r, c, :); alpha_east = enu_coeffs(1); alpha_north = enu_coeffs(2); alpha_up = enu_coeffs(3); Calling the east, north, up time series ts_east, ts_north, ts_up respectively, we can convert this to ts_LOS as follows: ts_LOS = alpha_east * ts_east + alpha_north * ts_north + alpha_up * ts_up