M7.5 Venezuela Earthquake: Line of sight displacement revealed by NISAR, ALOS-2 and Sentinel-1 Interferometry Rubi Garcia Gonzalez1, Molly Zebker1, Matt Brandin1, Baoning Wu1, Yohai, Magen1, Ekaterina Tymofyeyeva2, and David Sandwell1 1 - Institute of Geophysics and Planetary Physics, University of California, San Diego, USA 2 - NASA Jet Propulsion Laboratory, California Institute of Technology On June 24, 2026, a Mw7.5 earthquake struck Venezuela at 18:05:11 local time (10 km depth), and was preceded by a Mw7.2 foreshock (18:04:34 local time) just 37 seconds earlier at a depth of ~27km. Sparse local seismic and geodetic stations complicated initial mainshock location efforts, making satellite remote sensing a critical tool for identifying impacted zones. The coastal city of La Guaira suffered the heaviest destruction, with 5069 fatalities and 16,740 injured (ABC NEWS). The first interferometric products available for the epicentral area were provided by the European Space Agency (ESA) using C-band radar imagery ($\lambda$ = 5.6 cm). However, dense vegetation cover hindered the measurement of line-of-sight (LOS) surface displacement near key coastal cities (within a 14 km radius of Caracas). Despite these limitations, C-band interferograms we used to successfully resolve up to ∼100 cm of LOS ground displacement near the epicenters (~180 km far from the most impacted coastal cities), yielding the first direct satellite measurement of the earthquake sequence. Here we present high-coherence LOS surface deformation measurements derived from L-band Synthetic Aperture Radar (SAR) data. We use ALOS-2 ScanSAR observations from the Japan Aerospace Exploration Agency (JAXA) spanning both ascending (T040A) and descending (T134D & T135D) tracks. In addition, we integrate NASA-ISRO SAR (NISAR) observations from both ascending (T061 & T162) and descending (T054 & T126) tracks. These results highlight the critical role of L-band SAR in maintaining high interferometric coherence across densely vegetated terrain. Given the scarcity of local in-situ instrumentation (such as continuous Global Navigation Satellite Systems, GNSS, networks) in the affected region, these satellite-derived interferometric datasets serve as a primary quantitative resource to measure near-field surface deformation and constrain regional damage models. The epicenter location was relocated for the United States Geological Survey (USGS) ~115 km eastward from the original location once NISAR products became available. This highlights the importance of the NISAR satellite in less than 10 months of being launched. The InSAR data are processed with open source software GMTSAR and mapped using Generic Mapping Tools (GMT). NISAR SweepSAR ascending (T061 & T162) and descending (T054 & T126) coseismic interferograms The longer L-band wavelength (24 cm) can more effectively penetrate the vegetation cover compared to C-band (5.6 cm) (see at the end of this page). However, L-band data are more susceptible to long-wavelength ionospheric artifacts. Here we present the NISAR results for the Venezuela doublet event using both satellite trajectories. Two tracks were acquired for each direction: the ascending data are from tracks T162A (June 13, 2026 - June 25, 2026) and T061A (June 18, 2026 - June 30, 2026) and the descending data are from tracks T054D (June 17, 2026 - June 29, 2026) and T126D (June 22, 2026 - July 4, 2026). All interferograms have a temporal baseline of 12 days. NISAR products were processed from the L1 RSLC product with a spatial filter wavelength of 160m. We used frequency A and B SLC data to correct for the ionospheric phase contributions in the interferograms. The LOS decomposition was performed by assuming zero north displacement because the satellites fly in a near-polar orbit and are not sensitive to the north component. We solved for the east, E, and vertical (up), U, displacements as: $\begin{array}{c}\end{array}$$\begin{array}{c}\end{array}$$\begin{bmatrix} a_{los} \\ d_{los} \end{bmatrix} = \begin{bmatrix} a_e & a_u \\ d_e & d_u \end{bmatrix} \begin{bmatrix} E \\ U \end{bmatrix}$ where $a_{los}$ and $d_{los}$ are the ascending and descending LOS displacements, and [$a_e, a_u$] and [$d_e, d_u$] are the ascending and descending unit look vectors, respectively. Coseismic interferograms for both trajectories show a positive gradient northeast of Caracas, where the most damaged coastal cities are located, with a LOS  displacement of >-900 mm on the ascending track and >600 mm on the descending track. This region also shows a horizontal displacement to the west of >1200 mm, while the vertical component shows minimal displacement. NISAR East and Vertical Displacement East and vertical displacement maps are shown with thin contour lines at 100 mm and thick lines at 300 mm. Processed InSAR data for Modelers: data format is: #lon #lat #look_E #look_N #look_U #displacement(mm) - NSR-T162A: LOS displacement at 500 m resolution - NSR-T61A: LOS displacement at 500 m resolution - NSR-T54D: LOS displacement at 500 m resolution - NSR-T126D: LOS displacement at 500 m resolution NISAR Phase Gradient The unwrapping process is challenging near high decorrelated areas because the phase value exceeds 2pi, resulting in discontinuities or jumps along the phase as it cycles through the range. this causes a great source of noise when detecting small scale surface deformation. To overcome this problem, one could work using the wrapped phase and looking at its gradient, accounting for the rate of change of the phase along the range and azimuth, such $\nabla \phi (x) = \frac{R \nabla I + I \nabla R}{R^2 +I^2}$ where$\nabla \phi (x) = \left( \frac{\partial}{\partial r}, \frac{\partial}{\partial a} \right)$ . Therefore the phase gradient, can highlight fractures and faults that slip a few milometers during an earthquake and that can be masked where large displacement signals are observed. Here we present the surface expression of the Bocono Fault due to the M7.5 mainshock and M7.2 foreshock, with a length of ~38 km from the coast, where a difference of ~ 0.7 km is observed from the Active Fault Database from Venezuela (Audemard et al. 2000). Phase gradient along the range and azimuth of T054D. yphase_T54D_kmz fault surface trace on phase gradient along the range fault surface trace on phase gradient along the azimuth Phase gradient along the range and azimuth of T162A. yphase_T162A_kmz fault surface trace on phase gradient along the azimuth Coherence comparison of pre and post coseismic interferograms Using NISAR data from T126D, we form pre-seismic (June 6, 2026 - June 22, 2026) and post-seismic (June 22, 2026 - July 4, 2026) interferograms to measure the change in coherence across the two scenes. This highlight areas where major landslides occurred as a negative change in coherence (darker color), which are comparable to the regions that USGS reports landslides , especially the landslides that occurred north of Caracas. This is possible because the relative motion of random scatterers can be tracked by the decorrelation of the phase within the InSAR resolution cell. Therefore, drops in InSAR coherence is effective for mapping geophysical activity that completely repaves surface features, such as landslides, as these areas will be completely decorrelated in InSAR data. ALOS-2 ScanSAR ascending (T040A) and descending (T134D & T135D) coseismic interferograms We also downloaded ALOS-2 L-band data for three tracks. For the ascending track T040A, the reference and repeat images were acquired with a temporal baseline of 532 days (January 14, 2025 - June 30, 2026), and the descending tracks T134D (June 10, 2026 - July 8, 2026) and T135D (June 15, 2026 - June 29, 2026) have temporal baselines of 28 days and 14 days, respectively. The interferometric phase was processed with a spatial filter wavelength of 400 m and a decimation factor of 4. We did not correct for ionospheric phase contributions so there may be residual signal in the ALOS-2 displacements. There are three main observations for the ascending T040A and descending T134D/T135D interferograms. First, the coseismic wrapped interferograms show a LOS surface displacement of more than +/- 400 mm for both trajectories. Second, we observe an amplitude of 400 mm in the ascending track and -700 mm in the descending track along the San Sebastian Fault, where the mainshock epicenter was relocated on this fault 33 days after the event. Third, we observe a maximum LOS displacement of 800 mm north of Caracas. The wrapped interferogram also shows additional fringes, indicating increased displacement near the coastal cities north of Caracas. ALOS-2 East and Vertical Displacement East and vertical displacement maps are shown with thin contour lines at 100 mm and thick lines at 300 mm. Processed InSAR data for Modelers: data format is: #lon #lat #look_E #look_N #look_U #displacement(mm) - ALOS-2 T040A: LOS displacement at 500 m resolution - ALOS-2 T134D: LOS displacement at 500 m resolution - ALOS-2 T135D: LOS displacement at 500 m resolution Sentinel-1 ascending (T106A & T033A) and descending (T025D) coseismic interferograms C-band radar data offered the first LOS ground displacement measurements in the epicentral area as shown here. We use data from the ascending track T106A (June 18, 2026 - June 24, 2026), the ascending track T033A (June 18, 2026 - June 25, 2026), and the descending track T025D (June 13, 2026 - June 25, 2026), and all temporal baselines are 6 days. The two ascending tracks and the descending track were processed with a spatial filter wavelength of 200 meters and a decimation factor of 2. In the T106A and T033A data, a LOS displacement of 100 cm is observed as well as large areas of decorrelation where phase information is unreliable. Over 100 cm of LOS displacement was observed at Caracas and nearby coastal cities. We note that there are phase unwrapping errors due to the high decorrelation and may lead to misinterpretation. The C-band data do not maintain sufficient coherence to accurately resolve east and vertical displacement here. Acknowledgements: We thank NASA and JAXA for the rapid acquisition and distribution of their data. The development of the GMTSAR software, especially the processing chain for  ALOS-2 and NISAR was supported by NASA and the National Science Foundation through the NASA Earth Surface and Interior program (80NSSC23K0744), the NSF Office of Advanced Cyberinfrastructure program (OAC 2209808). Rubi Garcia Gonzalez and Molly Zebker were supported by the National Geodetic Survey through grant (NA23NOS4000334). Last updated: August 3, 2026