Seismic microzoning at a local scale of a given site or at a global scale of an urban center requires a fairly detailed knowledge of the lateral variations of the site effects, such as the amplification level, the fundamental soil frequency, and the site classification according to a given design seismic code. Over the last decade, the horizontal-to-vertical spectral ratio of background noise has become a widely used tool. However, it still has some shortcomings related to the windows selection as the microtremor contains many unknown and random sources other than the one related to the site effect. In this contribution, one proposes a new approach based on the frequency content to discriminate between the different sources. Site classification scheme recently proposed for seismic accelerations, which define target numerical and analytical horizontal over vertical spectral ratios for the 4 standard soil classes (Rock, soft rock, stiff soil and soft soil) defined in Eurocode-8, is adapted to background noise. In order to test the reliability of the proposed approach, microtremor measurements were performed on 67 sites with available shear-wave velocities down to a depth of at least 30 m. Free-field microtremor measurements are converted to accelerations instead to velocities, and split into 5 s non-overlapping windows. Automatic computation based statistical approach using a classification index which measures the similarity between the measured and target horizontal over vertical spectral ratios is performed for the 67 sites. The obtained results reveal a similar site classification, using the time-averaged shear-wave velocity to 30 m depth (V_s30) and the new approach, which can reach a success rate of about 84%. The automatic calculation provides the 67 sites with two fundamental parameters required for the preparation of a first-order microzonation map, the fundamental soil frequency, and the classification of the site according to Eurocode-8 guidelines. The proposed approach can be very useful for large-scale seismic micro-zoning of urban centers that may involve hundreds of measurement points, or supplement other geophysical measurements for a site specific seismic analysis.

中文翻译：

---

关于使用微震数据的新站点分类工具的提案

在给定地点的局部规模或城市中心的全球范围内的地震微区划分，需要对地点影响的横向变化有相当详细的了解，例如放大水平，基本土壤频率以及根据给定的设计抗震规范。在过去的十年中，背景噪声的水平与垂直光谱之比已成为一种广泛使用的工具。但是，由于微震中除了与位点效应有关的来源外，还包含许多未知的和随机的来源，因此它仍然与窗口选择有关。在这一贡献中，人们提出了一种基于频率内容的新方法，以区分不同的信号源。最近针对地震加速度提出了场地分类方案，定义了Eurocode-8中定义的4种标准土壤类别（岩石，软岩石，硬土和软土）的目标数值和分析水平与垂直光谱之比的方法，可以适应背景噪声。为了测试所提出方法的可靠性，在67个位置进行了微震测量，可用剪切波速低至至少30 m的深度。自由场微震测量值转换为加速度而不是速度，并分成5 s个非重叠窗口。对67个站点执行了使用分类索引的基于自动计算的统计方法，该分类索引可测量垂直方向上频谱比率与目标水平之间的相似性。获得的结果揭示了相似的网站分类，_s30）和新方法，可以达到约84％的成功率。自动计算为67个站点提供了准备一阶微区划图所需的两个基本参数，基本土壤频率以及根据Eurocode-8指南对站点进行的分类。所提出的方法对于可能涉及数百个测量点的城市中心的大规模地震微分区非常有用，或者可以针对特定地点的地震分析补充其他地球物理测量。