G53A-01
Variations in Creep Rate along the Central San Andreas Fault from InSAR and GPS Observations
The San Andreas Fault is locked along most of its length, but the 170 km-long section between San Juan Bautista and Parkfield undergoes creep. Measurements from creepmeters, alignment arrays and GPS over the last 25 years have shown that surface creep rates reach about 30 mm/year in the central portion, tapering off towards the locked segments at either end. Though useful, these measurements have been spatially isolated and intermittent in time. We present InSAR observations of creep across the fault, which have superior spatial coverage than previous data, and analyse them to investigate spatial variations in creep rate along the fault segment. From multiple ERS-1 and ERS-2 descending interferograms covering 1992 to 2001, we produce a stack which gives the spatial distribution of creep rate up to about 50 km either side of the fault. We find a maximum creep rate of about 32 mm/year. Deformation is step-like most of the way along the segment, but more distributed at the northern end, where the Calaveras Fault comes very close to the San Andreas Fault. We perform a linear inversion for shallow and deep sliding velocity on these faults using both the InSAR stack and GPS velocities from continuous (PBO) and campaign networks. Creep in the top few kilometers is variable along strike, with patches of faster creep interspersed with more slowly-moving patches. Creep at intermediate depths is greatest in the centre of the segment, reaching a few mm/yr less than the relative plate rate. The deep (> 12 km) sliding velocity is constrained to be less than or equal to the estimated long term relative plate velocity, and we estimate it to be a few mm/yr less than this. We compare the depth-averaged creep rate profile along the fault segment with that estimated by Nadeau and McEvilly (2004) from characteristic repeating microearthquakes. Between them, the three datasets utilised in this study suggest that creep is a spatially heterogeneous process.
G53A-02 INVITED
InSAR Measurements of Postseismic Deformation due to the Mojave Desert Earthquakes: Implications for the Driving Mechanisms
The 1992 Landers and 1999 Hector Mine earthquakes in the Eastern California Shear Zone have initiated robust postseismic transients that were used to study the rheology of the Earth's crust and upper mantle. I will present measurements of postseismic deformation due to both earthquakes derived from InSAR observations. A continuous imaging of the earthquake area by the European Space Agency ERS-2 satellite (including recent acquisitions in the Zero Gyro Mode) produced 7-year long timeseries of deformation for the Hector Mine, and 15-year-long timeseries of deformation for the Landers earthquake. A comparison of the InSAR data collected in the post-Landers and post-Hector Mine epochs indicates that the surface patters of postseismic deformation are very similar, implying a common mechanism. For both events the data reveal lobes of the radar range changes predominantly to the west of the respective earthquake ruptures. The lobes have a characteristic wavelength of a few tens of kilometers, and amplitude of several centimeters. High gradients in the radar range changes across the faults are suggestive of a shallow source of deformation. The data demonstrate that the post-Landers transient has persisted for more than 10 years after the earthquake. A number of different mechanisms has been proposed to explain the observed time-dependent deformation, including afterslip, viscous-like response of a substrate below the brittle-ductile transition, and re-distribution of pore fluids in the upper crust. For non-linear (e.g., powerlaw) rheologies, the surface deformation field may be indistinguishable from that due to afterslip at the early stages of relaxation, when the deformation is localized in high stress areas on the downdip continuation of the earthquake fault. However, at later stages of relaxation visco-elastic models predict appreciable changes in the displacement pattern. In particular, vertical velocities may change sign after viscous flow in the ductile substrate becomes more diffuse. No such reversal is observed in case of the Landers - Hector Mine sequence. Fluid flow and poro-elastic effects are incapable of explaining the observed horizontal deformation, but may substantially contribute to vertical postseismic motions, further complicating a discrimination between afterslip and visco-elastic relaxation.
G53A-03
Large-scale deformation of Tibet measured with Envisat ScanSAR interferometry
Scanning synthetic aperture radar (ScanSAR) interferometry has the potential to map deformation over large areas. The large actively deforming area and arid climate of Tibet make it an excellent test area for the ScanSAR interferometry technique, which has been proposed as a possible SAR operation mode for the DESDynI mission recommended by the Decadal Survey. The Envisat C-band (5.6 cm wavelength) ScanSAR is called Wide Swath (WS) mode and images a track that is over 400 km wide, with five subswaths. Envisat WS was not originally designed for interferometry, but about one in five pairs have the required burst synchronization. Since January 2007, the European Space Agency have been attempting to greatly improve the burst synchronization and reduce the variation in baselines. Unfortunately, some of the WS data over Tibet has been acquired in HH polarization and some in VV polarization, causing an additional limitation on usable interferometric pairs. Several Envisat tracks across Tibet have appropriate WS acquisitions to form interferometric pairs. Initial results show a reduction of coherence in eastern Tibet where the plateau climate is wetter and allows more vegetation cover. Processing long strips (>1000 km) with the full WS width gives strong constraints on the baseline between the two orbits of image-pair acquisitions and enables better separation of atmospheric effects and orbit errors from moderate- wavelength (~ 100 km) deformation signals. Long-wavelength control from GPS or other ground-based data is still required for the longest wavelengths (>300 km).
G53A-04
Characterizing and Mapping Ice Sheet Surface Topography Using a Medium-Footprint, Multi- Beam, Waveform-Recording Lidar
Lidar surveys of the Greenland ice sheet have been used to study mass-balance changes since the early 1990's. Sensors include NASA's ATM system (e.g., Krabill et al., 2000), and the ICESat (Schutz et al., 2002), a large- footprint, spaceborne system launched in 2003 for monitoring long-term trends in ice mass balance. To complement these data sets and prepare for the next-generation of spaceborne measurements, the Laser Vegetation Imaging Sensor (LVIS) was flown onboard the NASA P-3 aircraft over Greenland in September 2007. LVIS is an airborne, medium- footprint (25m diameter), full waveform-recording, airborne, scanning lidar system that has been used extensively for mapping forest structure, habitat, carbon and natural hazards. The system digitally records the shape of the returning laser echo, or waveform, after its interaction with the various reflecting surfaces of the earth, providing a true 3-dimensional record of the surface structure. Data collected included ground elevation and vertical extent measurements for each laser footprint, as well as the vertical distribution of intercepted surfaces (the return waveform) from which surface slope, roughness and other metrics can be extracted. During the mission, data were collected along ICESat repeat ground-track "corridors" that encompass a variety of terrain types (e.g., inland ice, crevasses, ponds, sastrugi, ice/rock margins, and bare earth), over sea- ice in northern Greenland, and at Jakobshavn Isbrae, a fast-flowing outlet glacier where discharge rates have increased in recent years. Data from this mission will be used to assess the ability of 25m-footprint, waveform lidar to precisely and accurately characterize and monitor the surface of the Greenland ice sheet and its margins. The data will also be used to assess the effects of across-track slope corrections currently being used on the ICESat data. The study will highlight the complimentary measurement science that can be achieved using a multi-beam, 25m footprint, contiguous beam laser altimeter such as the one proposed for inclusion in the DESDynI mission, especially in the high-slope, highly dynamic areas of Greenland.
G53A-05 INVITED
Ice sheet studies with DESDynI
DESDynI L-band InSAR promises timely and major advances in our monitoring capability and scientific understanding of the evolution of ice sheets and glaciers. InSAR has a long history of glaciology applications which started in 1991 with the launch of ERS-1. The European Space Agency ERS-1/2 satellites demonstrated the capability of InSAR to measure ice motion, grounding line migration, downdraw of ice surfaces, ice-shelf rifting and tidal flexure, glacial surges and other essential characteristics of ice dynamics, however with limited temporal and spatial coverage. Radarsat-1 and Envisat ASAR permitted significant advances in ice motion mapping using speckle tracking techniques, and longer-term observations of glacier evolution, but missed the shorter time scale resolution of ERS-1/2 tandem data. DESDynI will offer both, with higher quality and higher frequency of visit. The higher quality stems from the higher coherence of L-band signals on snow and ice demonstrated with SIR-C and confirmed with ALOS PALSAR, which is important in high- acumulation coastal sectors. Higher frequency of re-visit stems from the 8-day repeat and a plan to systematically acquire data with controlled baselines. ALOS PALSAR early results in West Antarctica and Greenland are very encouraging, with more data to come.Yet ALOS PALSAR 46-day repeat limits its capability to observe short-terms events e.g. calving, grounding line migration, tidal modulation in glacier velocity, downdraw consecutives to changes in basal sliding or subglacial drainage, which are important to observe to characterize the impulse response of glaciers to climate perturbations. DESDynI promises important discoveries and new science advances that are not possible with existing and planned InSAR missions. These advances will be a pillar for the development of more realistic numerical ice sheet models capable of realistic predictions of their evolution in a continously warming climate. This work was performed at Caltech's Jet Propulsion Laboratory under a contract with the National Aeronautics and Space Administration.
G53A-06
Subsidence in the Avcilar district of Istanbul, Turkey measured by satellite radar interferometry 1992 – 1999
Analysis of 14 synthetic aperture radar images acquired by the ERS-1 and ERS-2 satellites between 1992 and 1999 by interferometry (INSAR) reveals subsidence around the Avcilar area of Istanbul. Using the General Inversion for Phase Technique, (GIPhT), we analyze a set of 24 interferometric pairs. The interferometric fringe patterns show a crescent shape. We interpret them as purely vertical subsidence at a secular rate. The maximum subsidence rate of at least 5 mm/yr occurs at a point located at latitude 40.98 deg N and longitude 28.71 deg E. A simple 4-parameter elastic Mogi model consisting of three infinitesimal spherical sinks at a depth of the order of 2 km deflating at 50 to 100 thousand cubic meters per year describes subsidence signal to first order. The model also accounts for tropospheric effects by estimating one vertical phase gradient for each image acquisition epoch. The model fits the data significantly better than a null model with 95 percent confidence. The spatial distribution of the subsidence suggests that most of the subsidence occurs as compaction in a shallow layer of unconsolidated, partially saturated soil with a relatively weak lithology. This layer appears to deform easily in several processes, including compaction, landslide failure, and seismic wave amplification. It may also be susceptible to liquefaction during a large earthquake.
G53A-07
Spatial and Temporal Characterization of the Portuguese Bend Landslide, California, Using InSAR
Here we use InSAR and GPS data to spatially and temporally constrain deformation of the Portuguese Bend Landslide, and correlate movement with seasonal rainfall. The Portuguese Bend Landslide on the Palos Verdes peninsula in southern California has been sliding at a rate of several meters per year since the late 1950's. Earlier work has shown that displacement varies greatly from season to season, with high values occurring during the winter and relatively low values in the summer. Our findings suggest that the landslide continues to move at rates comparable to those found in previous studies. With data collected from European ERS satellites over tracks 170 and 442 between June 1992 and November 2000, we use the ROI_PAC software package to process more than 500 interferograms of the region. The interferograms show a wide variation in phase coherence, with the greatest coherence occurring during the dry summer months when deformation is the slowest. We do not expect winter interferograms to be coherent over the sliding area as the displacement rate is too large to be measured with conventional techniques. Stacking the 22 summer interferograms, we determine the line-of-sight displacement rate during the summer months to average 50±36 mm/yr. The large deviation is due to yearly changes in displacement rate. Additionally, we stack the incoherent areas in the winter months by counting how often each pixel is coherent to create a qualitative spatial map of high deformation. While this stack does not yield a displacement rate for the winter months, it constrains the extent of the slide and suggests that it is confined to the previously mapped region. The map of phase coherence also constrains the initiation and cessation of fast ground motion to the months of December through April. This corresponds to the rainy season, during which displacement rate increases via infiltration and pore pressure generation within the slide mass. We also analyze existing GPS data for the region, spanning the same time interval and area as the InSAR data. The InSAR result for the magnitude of the phase change in the summer months is validated by reducing the three component GPS data to the satellite's line-of-sight. Combining InSAR and GPS with rain gauge data, our long term objective is to better understand the mechanical and hydrological properties of the Portuguese Bend Landslide and monitor its activity over the past decade.
G53A-08
Regional Landslide Mapping and Monitoring in Norway Using SBAS InSAR
Being a mountainous country, with long steep fjords and valley sides, Norway is particularly susceptible to large rock avalanches. In the last 100 years, over 170 people have been killed by tsunamis in fjords caused by large rock avalanches. In each case, the rock avalanche was preceded by many years of slow movement, with acceleration prior to slope failure. At present, three similar unstable areas have been identified in Norway, and are being monitored using extensive instrumentation. With several thousand kilometers of inhabited coastline and valleys, the challenge we currently face is the identification of similar hazards in an efficient manner. The Geological Survey of Norway (NGU) is responsible for landslide mapping throughout the country. Since 2005, Norut AS has been helping NGU establish a Norwegian facility for InSAR processing. The goal is to be able to systematically perform interferometric processing of SAR images from multiple satellites to assist in geohazard mapping and monitoring. Once hazardous slopes are identified, continued monitoring using InSAR can be augmented with ground-based systems. Over 700 ERS and ENVISAT scenes, covering 19 overlapping frames, are currently being processed using the Small Baseline Subset (SBAS) algorithm. These scenes cover the area of northern Norway with highest topographic relief, stretching from the Lofoten islands in the southwest to Alta in the northeast. The first results, based on ERS scenes from 1992-1999 only, are from around the Lyngen peninsula, just east of the city of Tromsø. Processing challenges are atmospheric stratification due to high topographic relief and nonuniform temporal sampling due to long winter season with snow cover possibly lasting from October to May. However, the first results show remarkably good coherence due to the lack of vegetation above 700 m. Nordnesfjellet is one of the three sites currently being monitored. InSAR results clearly show the outline of the moving block, and velocity estimates are in agreement with earlier GPS measurements, though the accuracy of the GPS measurements so far is somewhat questionable due to the short time series (yearly measurements since 2004 only). Numerous other areas within the processed area have downwards velocities of up to one centimeter per year. Extensive field checking this summer has identified active fracture systems with evidence of movement in each of these areas. In at least one case, differential movement across multiple fault scarps is shown by different movement rates of the sub-blocks. InSAR analysis of the almost two decades long time series available in the ERS and ENVISAT archives enables rapid identification of landslides within a large region, allowing us to focus field mapping in areas with known hazards. Without the use of such remote sensing tools, it would take decades to map the same area. Over the course of the next year, we will process the rest of northern Norway, and then begin work on the most hazardous areas of western Norway, where several active landslides have been identified.