Tectonics/Volcanoes 2

More than 82 mm/yr of plate convergence has made Taiwan one of the most active tectonic regions in the world. On the other hand, anthropogenic activities such as groundwater pumping have induced extensive land subsidence in western Taiwan. Recent geodetic work, including continuous GPS and leveling networks, has revealed the first-order patterns of Taiwan crustal deformation. However, due to the relatively low density of geodetic samplings, the details of surface deformation such as fault creep or locking depth in interseismic period, and land subsidence and its relationship with annual discharge/recharge are not well known. In this study, we primarily use Synthetic Aperture Radar (SAR) from the Sentinel-1A satellite since March 2015, and generate time series of surface deformation based on a modified InSAR small baseline method. We are able to separate horizontal movement from vertical by comparing both ascending and descending orbits. We additionally include ERS, Envisat, and ALOS SAR data to extend time series in southwestern Taiwan starting from 1995. Our preliminary results show highest deformation along the Longitudinal Valley up to 50 mm/yr in eastern Taiwan relative to a stable point in the Western Plain, which is higher than the rate in southern Central Range (~30 mm/yr) where the highest peak (3952 m) is located. In southwestern Taiwan, we find uplift rate at more than 10 mm/yr along the southern Western Foothills, implying rapid interseismic deformation along this fold-and-thrust belt. The extended time series in southwestern Taiwan allows us to identify the spatial patterns of seasonal versus long-term deformation. The seasonal displacement is mainly associated with groundwater recharge and withdrawal, whereas the longer-term deformation may indicate tectonic loading during the interseismic period of the earthquake cycle, including elastic strain along or continuous creep on the active faults in Taiwan. These results are potentially useful for evaluating seismic hazards.

Time dependant slip modeling can be a powerful tool to improve our understanding on the interaction of earthquake cycle processes such as interseismic, coseismic, postseismic, and aseismic slip. Interferometric Synthetic Aperture Radar (InSAR) observations allow us to model slip at depth with a higher spatial resolution than when using GNSS alone. Typically the temporal resolution of InSAR has been limited. However, the recent generation of SAR satellites including Sentinel-1, COSMO-SkyMED, and RADARSAT-2 permits the use of InSAR for time-dependant slip modeling, at intervals of a few days when combined. The increasing amount of SAR data makes a simultaneous data inversion of all epochs challenging. Here, we expanded the original Network Inversion Filter (Segall and Matthews, 1997) to include InSAR observations of surface displacements in addition to GNSS. In the NIF framework, geodetic observations are limited to those of a given epoch, where a physical model describes the slip evolution over time. The combination of the Kalman forward filtering and backward smoothing allows all geodetic observations to constrain the complete observation period. Combining GNSS and InSAR allows us to model time-dependant slip at an unprecedented spatial resolution. We validate the approach with a simulation of the 2006 Guerrero slow slip event. In our study, we emphasize the importance of including the InSAR covariance information, and demonstrate that InSAR provides an additional constraint on the spatial extent of the slow slip.

References:

Segall, P., and M. Matthews (1997), Time dependent inversion of geodetic data, J. Geophys. Res., 102 (B10), 22,391 – 22,409, doi:10.1029/97JB01795.

APhoRISM is a collaborative project under the theme FP7-SPACE-2013-1 of the Seventh Framework Programme of the European Commission. It proposes the development and testing of two new methods to combine different types of Earth Observation satellite and ground data: one is related to the monitoring of volcanic crises (MACE method) and the other one concerns seismic events (APE method). This work is focused on the second topic.

The APE method (A Priori information for Earthquake damage mapping), concerns the generation of maps to address the detection and estimate of damage caused by a seism. The novelty of APE relies on a preparedness phase, which exploits the a priori information derived by InSAR time series to measure surface movements, shake maps obtained from seismological data, and vulnerability information. A data fusion algorithm is proposed to integrate the a-priory information layers with remote sensing change detection maps to estimate the damage likelihood. The outputs of APE method are likelihood index damage maps at different scale: medium resolution, i.e. scale of few building blocks, or very high resolution, i.e. scale of individual buildings.

Two seismic relevant events have been selected for validation purposes, at two different spatial scales: Christchurch (New Zealand) 21^st of February 2011 earthquake, and Port au Prince (Haiti) 12^th of January, 2010 earthquake for high and medium resolution, respectively. In both events the area affected by damage is very large and the magnitude of the disaster is high: the 2010 Haiti earthquake was a catastrophic magnitude 7.0 Mw earthquake, with an estimated three million people affected by the quake, and 250,000 residences and 30,000 commercial buildings collapsed or severely damaged; the 2011 Christchurchearthquake was a 6.3 Mw earthquake characterized by the vicinity between two events, which killed 181 people and injured several thousand. The earthquake brought down many buildings that had been damaged in September 2010 earthquake.

In both events a good ground reference dataset is available, and an adequate coverage by EO-data has been verified. The results achieved so far will be analyzed and compared against independently ground truth or reference data derived by local survey or by other EO data, e.g. airborne optical sensors, in order to validate the intermediate results in the preparedness and crisis phases, and the likelihood index damage maps.

Over the last 15 years, Interferometric Synthetic Aperture Radar (InSAR) has emerged as a unique and unsurpassed geodetic tool for the measurement of interseismic crustal velocity on the hundred-kilometre scale. Larger, country-wide or tectonic plate-scale measurements have long been a goal, as they enable study of continental deformation processes across lengthscales ranging over six orders of magnitude, from tens of metres to thousands of kilometres. In addition, these datasets offer a homogenous high-resolution measure of seismic hazard, by relating accumulated geodetic elastic strain in the upper crust to seismic strain released in earthquakes. However, efforts to measure crustal velocities over such wide areas have previously been hampered by the heterogeneous temporal and spatial coverage and long average revisit times of current-generation SAR datasets.

The Sentinel-1 constellation now affords automated measurements of interseismic crustal velocities at the country-wide and tectonic-plate scale. Regular 12 day SAR acquisitions in ascending and descending viewing geometries over Europe provide the two conditions necessary for such work; 1) large data volumes and 2) short temporal baselines. This rapid revisit time is enabled by Sentinel-1’s novel TOPS (Terrain Observation by Progressive Scans) imaging mode.

Turkey is an ideal test location to make interseismic Sentinel-1 measurements due to rapid rates of tectonic deformation and coverage by ESA’s European regional mask for Sentinel-1, which means that regular 12-day SAR acquisitions are available. Much of Turkey falls within the Anatolia tectonic region, which has previously been postulated to be a microplate, caught in the continental convergence zone between the Eurasian, African and Arabian plates. Relative to Eurasia, Anatolia moves westwards towards the Hellenic trench in the Aegean Sea at a rate of around 2 cm/yr, and this motion is largely accommodated by two major active strike-slip faults; the North and East Anatolian Faults.

We combine over 300 interferograms spanning 7 ascending and 7 descending Sentinel-1 tracks over Turkey to produce a country-wide velocity field for the first time. We process interferograms from the first year of Sentinel-1A’s mission using our prototype automated processing chain, apply corrections to mitigate atmospheric contamination, and combine data for each track using a small baseline subset approach. We measure interseismic deformation across the full length of both the North and East Anatolian Faults, and discuss the implications of our measurements for models of continental deformation and seismic hazard in Turkey.

The Mw=8.3 earthquake occurred offshore of Illapel on 16^th September, 2015 at 22:54 UTC. A tsunami several metres high was generated that reached the coast of Chile within minutes and traversed the entire Pacific. We use geodetic displacements from Sentinel-1 radar interferometry and records of tsunami propagation across the Pacific from seafloor pressure gauge data, to constrain the fault slip distribution. We correct the interferograms for long wavelength tropospheric phase delay using the ECMWF weather model. We divide the fault into patches and invert for slip using a Bayesian Markov chain Monte Carlo algorithm to provide the full probability distribution for the slip on each patch. We validate our model by running the simulated tsunami through a flooding model and comparing this to tide gauge and tsunami run-up data. We find that slip reached the trench and also likely re-ruptured part of the fault that ruptured in 1943 M8.1 earthquake. We analyse the slip in the context of other historic earthquakes that have occurred on the subduction zone, and estimate the impact of this earthquake on the hazard for the region.