Conference Agenda

Multitemporal InSAR observations have proved to be key observations to characterize spatial and temporal variations of interseismic strain along major faults, allowing not only to retreive average interseismic velocity maps but also transient aseismic slip events, giving new lights on seismic hazard assessment. With their high temporal resolution and wide spatial coverage, Sentinel-1 (S1) InSAR data can be analyzed in time series to tackle the multi-scale issues of seismic hazard, as well as the off-fault deformation and non tectonic signals (such as seasonal hydrological loads) quantification.

We focus here primarily on the eastern border of the Tibetan plateau, from the Himalayan syntax in the south to the Ordos in the north, marked by major faults, recently broken (Longmen Shan thrusts at the origin of the Mw 7.9, Wenchuan earthquake in 2008) or known as seismic gaps unbroken for several hundred years. Geodetic data available to date (GPS, InSAR ERS / Envisat time series) show that some of these gaps are the site of aseismic slow slips (such as some segments along the Haiyuan and Xian Shui He faults, and possibly along the Himalayan front in Bhutan), which, depending on their spatio-temporal characteristics, can help to reduce the seismic hazard on these faults or, conversely, facilitate the initiation of future major ruptures. In addition, the eastern and southern borders of the tibetan plateau are marked by high mountain ranges (Longmen Shan in the east and Himalayas in the south, with elevations variations of several kilometers), subject to erosion and contrasting with basins affected by hydrological loads varying seasonally. We review here our most recent studies over this eastern border of the tibetan plateau, analyzing large scale velocity fields obtained from S1 data time series analysis over descending and ascending orbits, emphasing improvments in InSAR processing specific to S1 data.

The ~1000-km-long Chaman fault system consists of a series of subparallel left-lateral strike-slip faults and thrusts that form the transform to transpressive plate boundary between the Indian Plate and the Eurasian Plate. Studies based on geological and plate closure estimates show that the northward plate motion of India with respect to Eurasia is on the order of 35 mm/yr. Previous InSAR studies (Barnhart, 2016; Fattahi & Amelung, 2016) along the Chaman fault system based on Envisat and ALOS data have shown that the northeast-southwest trending Chaman fault itself only accounts for ~30% of this relative motion. How the remaining 70% is distributed or localized on adjacent structures remain to be determined. Such studies also revealed the existence of shallow creep along more than 300 km of the Chaman fault. The large spatial coverage of the recent Sentinel-1 data allows to tackle both the large-scale strain partitioning issue and the fault-scale creep behavior characterization.

In order to estimate strain accumulation rates along and across the Chaman fault system, we map the present-day interseismic velocity fields using long-swath (> 1250 km) Sentinel-1 (S1) TOPS radar images acquired on both ascending (T42, T71, T144) and descending (T151, T78) orbits, along the western boundary of the Indian subcontinent. Using an automatized processing workflow, we have processed time series of ~150 S1 images acquired between 2014 and 2018. Preliminary results show left-lateral shear velocities of ~20 mm/yr across the distributed plate boundary, with a complex partitioning between the main Chaman left-lateral fault, other adjacent left-lateral faults or secondary structures within the thrust belt. While ascending data are mostly sensitive to the left-lateral component of slip and vertical motion along the Chaman fault, descending data highlight horizontal and vertical motion across secondary structures branching on the main Chaman fault. Surface aseismic creep rate along the Chaman fault seems to reach up to ~10 mm/year and may extend along a ~600 km-long segment, between 28.5 ^oN and 32.5^oN, which appears significantly (50%) longer than that reported in previous studies. Surface creep thus accommodates ~30% of the tectonic loading along a significant portion of this plate boundary. Further data analysis and modelling will provide a better quantification of the creep rate amplitude and depth along fault strike, deep tectonic loading, and strain partitioning on secondary structures.

In the past 10 years after the Wenchuan earthquake, important information was obtained from analysis of InSAR data of the event, including fault geometry, slip-distribution, rupture propagation and dynamics etc. Though some GPS data collected over both sides of the earthquake fault, InSAR data with full coverage of the Sichuan basin and Longmenshan provides crucial information about the kinematic and dynamic processes of the earthquake.

For this particular region with drastic elevation changes, two end-member models were proposed to interpret the deformation mechanism of the Tibetan Plateau and generation of the Wenchuan earthquake, namely the thrust-fold belt or the viscous lower crustal flow models. Here we re-analyze the PALSAR InSAR data acquired ~10 years ago when we published the first results in both Sun et al. (2008) and Shen et al. (2009). A number of correction techniques we developed in these years are applied to the data, particularly for the ionospheric noise in meter scale. Hence the coseismic deformation is greatly improved in this analysis. Then we use a nonlinear-linear-mixed technique to invert the data for detailed rupture features of the Wenchuan earthquake. Our inversion indicates that a shallowly west-dipping segment to the south and a near-vertical segment to the northeast are good enough for fitting the InSAR data, and the displacements of a horizontal detachment extending to the west are not needed.

Thanks to ESA’s Sentinel-1 A/B satellites, a high-temporal resolution dataset over the Longmenshan region is now available, with the earliest acquisition in the middle of 2014. By using two advanced time-series analysis techniques on the TOPS mode data in both ascending and descending pass, we analyzed the deformation process on subswath-by-subwath basis covering the Longmenshan and Minshan regions, where the Wenchuan earthquake occurred in 2008 and the Jiuzhaigou earthquake happened last year respectively. Our first result indicates that the Sentinel-1 SAR dataset is promising for resolving the subtle tectonic deformation process of this region, though other sensors, such as ERS/Envisat, may suffer from heavy decorrelation in the same region.