A review of the microtremor horizontal-to-vertical spectral ratio (MHVSR) method

Journal of Seismology - The single-station microtremor horizontal-to-vertical spectral ratio (MHVSR) method was initially proposed to retrieve the site amplification function and its resonance...     

Statistical correlation of earthquake and ambient noise spectral ratios

The Horizontal-to-Vertical Spectral Ratio from earthquake (HVSR) and from ambient noise (HVN) recordings realistically indicate the fundamental frequency of soil response but, for the majority of the worldwide examined sites, they do not provide reliable amplification curves as calculated by the earthquake standard Spectral Ratio (SSR). Given the fact that HVSR and especially HVN can be easily obtained, it is challenging to search for a meaningful correlation with SSR amplification functions for the entire frequency band and to use the results for the SSR estimate at a further site where only noise measurements are available. To this aim we used recordings from 75 sites worldwide and we applied a multivariate statistical approach (canonical correlation analysis) to investigate and quantify any correlation among spectral ratios. The canonical correlation between SSR and HVN is then used to estimate the expected SSR at each site by a weighted average of the SSR values measured at the other sites; the weights are properly set to account more for sites with similar behaviour in terms of the canonical correlation results between HVN and SSR. This procedure, repeated for all sites in turn, constitutes the basis of a cross validation. The comparison between the inferred and the original SSR highlights the improvements of site response estimation with respect to the use of ambient noise techniques. The goodness and limitations of the reconstruction procedure are explained by specific geological settings.     

Comparison between seismic vulnerability models and experimental dynamic properties of existing buildings in France

Elastic fundamental frequency is a key-parameter of simplified seismic design and vulnerability assessment methods. Empirical relationships exist in codes to estimate this frequency but they miss experimental data to validate them accounting for national feature of building design and, above all, corresponding uncertainties. Even if resonance frequency extracted from ambient vibrations may be larger than the elastic frequency (at yield) generally used in earthquake engineering, ambient vibration recordings may provide a large set of data for statistical analysis of periods versus building characteristics relationships. We recorded ambient vibrations and estimated the fundamental frequency of about 60 buildings of various types (RC and masonry) in Grenoble City (France). These data complete the set existing yet, made of 26 RC-buildings of Grenoble (Farsi and Bard 2004) and 28 buildings in Nice (France) (Dunand 2005). Statistical analysis of these experimental data was performed for fundamental frequencies of RC shear wall structures and the results are compared with existing relationships. Only building height or number of stories has a statistical relevancy to estimate the resonance frequency but the variability associated to the proposed relationships is large. Moreover, we compared the elastic part of capacity curves of RC and masonry buildings used in the European Risk-UE method for vulnerability assessment with the experimental frequencies. The variability is also large and the curves may not be consistent with French existing buildings.     

Prediction of strong ground motion at EURO-SEISTEST site using the empirical Green's function method

This paper presents some tests on the empirical Green's function method, performed in order to check its effectiveness to predict strong ground motion during future large earthquakes. These investigations basically consist of blind or pseudo-blind tests using part of the data set obtained at the Volvi (Greece) test-site within the framework of the EURO-SEISTEST project. In a first step, a moderate event (M_L=4.1) is simulated by using a small event (M_L=2.5), and taking into account the a priori knowledge of the relevant source parameters (magnitude and stress-drop) for both the base and target events. This check emphasizes the sensitivity of the prediction to the stress-drop values. In a second step, a pseudo-blind prediction is carried out in which the information on the target event is only the magnitude (M_L=5.3) besides the geometrical parameters such as source location, strike and dip. The other important parameters (seismic moment, fault area, stress-drop) are determined on the basis of specific empirical scaling laws derived from several small and moderate events occurring in the area. The synthetic motions are computed for two nearby sites, which are located on the southern edge and in the center of the Mygdonian graben, respectively, and correspond to much different site conditions: weathered rock and thick sediments. They are found in good agreement with the observations, which were unveiled after the simulation. In particular, the amplitude and the phase of the late, local surface waves generated on the southern edge of the graben, are very well reproduced at valley center. Finally, the last step is an attempt to predict strong ground motion for a hypothetical large earthquake of magnitude M_L=6.5. The results are shown to depend very significantly on the scaling laws.     

Spatial coherence of seismic wave fields across an alluvial valley (weak motion)

The loss of coherence of seismic waves, even over small distances, can become significant and consequently important for engineering applications. This coherence loss depends on several parameters (frequency, inter-station distance), but their dependencies are not well understood yet. We show the results of a detailed analysis of the spatial variability of the coherence between two signals. This analysis is based on a simple model which is an exponential function of inter-station distance and frequency. Data from a temporary network of 36 weak-motion instruments installed at the EURO-SEISTEST site, north-east of Thessaloniki (Greece), is used in this work. The data offers the unique possibility to study the loss of coherence over a wide range of distances, as the inter-station distances are between 8 m and 5488 m. We test the influence on the coherence of the length of the time-window of the signals, of the type of the waves present in the time-window and of the common energy of the signals. We also show that, at least at this particular site, the loss of coherence with distance is probably marked by a "cross over" distance, distinguishing two different ranges: one for inter-station distances up to 100 m and the other above 100 m. Finally, we find that the coherence determined from noise recordings behaves in a rather similar way to the one determined from coda-waves and more stationary (longer) signals. Therefore, noise can be useful for a rough, but quick estimation of the loss of coherence, at least for inter-station distances larger than 100 m.     

Comparisons among the five ground-motion models developed using RESORCE for the prediction of response spectral accelerations due to earthquakes in Europe and the Middle East

This article presents comparisons among the five ground-motion models described in other articles within this special issue, in terms of data selection criteria, characteristics of the models and predicted peak ground and response spectral accelerations. Comparisons are also made with predictions from the Next Generation Attenuation (NGA) models to which the models presented here have similarities (e.g. a common master database has been used) but also differences (e.g. some models in this issue are nonparametric). As a result of the differing data selection criteria and derivation techniques the predicted median ground motions show considerable differences (up to a factor of two for certain scenarios), particularly for magnitudes and distances close to or beyond the range of the available observations. The predicted influence of style-of-faulting shows much variation among models whereas site amplification factors are more similar, with peak amplification at around 1s. These differences are greater than those among predictions from the NGA models. The models for aleatory variability (sigma), however, are similar and suggest that ground-motion variability from this region is slightly higher than that predicted by the NGA models, based primarily on data from California and Taiwan.     

Empirical evaluation of microtremor H/V spectral ratio

The objective of this work is to perform a purely empirical assessment of the actual capabilities of the horizontal-to-vertical (H/V) spectral ratio technique to provide reliable and relevant information concerning site conditions and/or site amplification. This objective has been tackled through the homogeneous (re)processing of a large volume of earthquakes and ambient noise data recorded by different research teams in more than 200 sites located mainly in Europe, but also in the Caribbean and in Tehran. The original recordings were first gathered in a specific database with information on both the sites and recorded events. Then, for all sites close to an instrumented reference, average site-to-reference spectral ratios (“spectral ratio method” (SSR)) were derived in a homogeneous way (window selection, smoothing, signal-to-noise ratio threshold, averaging), as well as H/V ratios (“HVSRE–RF”) on earthquake recordings. H/V ratios were also obtained from noise recordings at each site (either specific measurements, or extracted from pre- or post-event noise windows). The spectral curves resulting from these three techniques were estimated reliable for a subset of 104 sites, and were thus compared in terms of fundamental frequency, amplitude and amplification bandwidth, exhibiting agreements and disagreements, for which interpretations are looked for in relation with characteristics of site conditions. The first important result consists in the very good agreement between fundamental frequencies obtained with either technique, observed for 81% of the analyzed sites. A significant part of the disagreements correspond to thick, low frequency, continental sites where natural noise level is often very low and H/V noise ratios do not exhibit any clear peak. The second important result is the absence of correlation between H/V peak amplitude and the actual site amplification measured on site-to-reference spectral ratios. There are, however, two statistically significant results about the amplitude of the H/V curve: the peak amplitude may be considered as a lower bound estimate of the actual amplification indicated by SSR (it is smaller for 79% of the 104 investigated sites), and, from another point of view, the difference in amplitude exhibits a questioning correlation with the geometrical characteristics of the sediment/basement interface: large SSR/HV differences might thus help to detect the existence of significant 2D or 3D effects.     

Evaluation of the influence of experimental conditions on H/V results from ambient noise recordings

The H/V-noise technique is now widely used to estimate site effect parameters (fundamental frequency and sometimes the associated soil amplification), and many surveys using this technique have provided convincing results. However, a general agreement on a methodology for data acquisition, data processing and result interpretation has yet to be found. H/V measurements from ambient noise recordings imply both reliability of the results and rapidity of data collection. It is therefore important to understand which experimental conditions (1) influence data quality and reliability, and (2) can help speeding up the recording process. Within the framework of the SESAME European project, a specific task was defined to investigate the reliability of the H/V spectral ratio technique in assessing the site effects. The aim of WP02, one specific Work Package of the SESAME project, is to study the effects of experimental conditions on both stability and reproducibility of H/V results. This study has been conducted in a purely experimental way, by testing the possible influence of various experimental conditions on H/V results both on the frequency peak value and on its amplitude. WP02 results help setting up the experimental conditions under which ambient noise recordings have to be performed in order to provide reproducible, reliable and meaningful H/V results. In this paper we present the results of the WP02 SESAME project concerning the evaluation of the influence of experimental conditions of ambient noise recording on H/V results.