International audienceThe horizontal-to-vertical (H/V) method has the potential to significantly contribute to site effects evaluation, in particular in urban areas. Within the European project, site effects assessment using ambient excitations (SESAME), we investigated the nature of ambient seismic noise in order to assess the reliability of this method. Through 1D seismic noise modeling, we simulated ambient noise for a set of various horizontally stratified structures by computing efficiently the displacement and stress of dynamic Green's functions for a viscoelastic-layered half-space. We performed array analysis using the conventional semblance-based frequence-wavenumber method and the three-component modified spatial autocorrelation method on both vertical and horizontal components and estimated the contribution of different seismic waves (body/surface waves, Rayleigh/Love waves) at the H/V peak frequency. We show that the very common assumption that almost all the ambient noise energy would be carried by fundamental-mode Rayleigh waves is not justified. The relative proportion of different wave types depends on site conditions, and especially on the impedance contrast. For the 1D horizontally layered structures presented here, the H/V peak frequency always provides a good estimate of the fundamental resonance frequency whatever the H/V peak origin (Rayleigh wave ellipticity, Airy phase of Love waves, S-wave resonance). We also infer that the relative proportion of Love waves in ambient noise controls the amplitude of the H/V peak

Dynamic glacier activity is increasingly observed through passive seismic monitoring. We analysed near-regional-scale seismicity on the Arctic archipelago of Svalbard to identify seismic icequake signals and to study their spatialÁ temporal distribution within the 14-year period from 2000 until 2013. This is the first study that uses seismic data recorded on permanent broadband stations to detect and locate icequakes in different regions of Spitsbergen, the main island of the archipelago. A temporary local seismic network and direct observations of glacier calving and surging were used to identify icequake sources. We observed a high number of icequakes with clear spectral peaks between 1 and 8 Hz in different parts of Spitsbergen. Spatial clusters of icequakes could be associated with individual grounded tidewater glaciers and exhibited clear seasonal variability each year with more signals observed during the melt season. Locations at the termini of glaciers, and correlation with visual calving observations in situ at Kronebreen, a glacier in the Kongsfjorden region, show that these icequakes were caused dominantly by calving. Indirect evidence for glacier surging through increased calving seismicity was found in 2003 at Tunabreen, a glacier in central Spitsbergen. Another type of icequake was observed in the area of the Nathorstbreen glacier system. Seismic events occurred upstream of the glacier within a short time period between January and May 2009 during the initial phase of a major glacier surge. This study is the first step towards the generation and implementation of an operational seismic monitoring strategy for glacier dynamics in Svalbard.

International audienc

S U M M A R YModern acquisition of seismic data on receiver networks worldwide produces an increasing amount of continuous wavefield recordings. In addition to manual data inspection, seismogram interpretation requires therefore new processing utilities for event detection, signal classification and data visualization. The use of machine learning techniques automatises decision processes and reveals the statistical properties of data. This approach is becoming more and more important and valuable for large and complex seismic records. Unsupervised learning allows the recognition of wavefield patterns, such as short-term transients and long-term variations, with a minimum of domain knowledge. This study applies an unsupervised pattern recognition approach for the discovery, imaging and interpretation of temporal patterns in seismic array recordings. For this purpose, the data is parameterized by feature vectors, which combine different real-valued wavefield attributes for short time windows. Standard seismic analysis tools are used as feature generation methods, such as frequency-wavenumber, polarization and spectral analysis. We use Self-Organizing Maps (SOMs) for a data-driven feature selection, visualization and clustering procedure. The application to continuous recordings of seismic signals from an active volcano (Mount Merapi, Java, Indonesia) shows that volcanotectonic and rockfall events can be detected and distinguished by clustering the feature vectors. Similar results are obtained in terms of correctly classifying events compared to a previously implemented supervised classification system. Furthermore, patterns in the background wavefield, that is the 24-hr cycle due to human activity, are intuitively visualized by means of the SOM representation. Finally, we apply our technique to an ambient seismic vibration record, which has been acquired for local site characterization. Disturbing wavefield patterns are identified which affect the quality of Love wave dispersion curve estimates. Particularly at night, when the overall energy of the wavefield is reduced due to the 24-hr cycle, the common assumption of stationary planar surface waves can be violated.

No abstract