城市微动高阶面波在浅层勘探中的应用：以苏州河地区为例

频率-贝塞尔变换方法(Frequency-Bessel Transform method,简称F-J方法)是一种分析微动信号的新方法,由于该方法采用频率矢量波数变换处理水平层状各向同性弹性模型中时空平稳随机分布的微动信号,所以从理论上可以提取出清晰的瑞利波基阶和高阶模态频散曲线,但是目前还没有相关的野外实验对此进行研究和应用.本文首先采用该方法对上海市苏州河地区采集的城市微动信号进行处理获得了频率-相速度谱,然后提取了多模态瑞利波频散曲线,最后通过粒子群算法对频散曲线进行联合反演,得到了浅地表0~70 m深度范围的S波速度结构,并且利用钻孔数据对反演的速度结构进行了验证.另外,本文还通过对比F-J方法和传统的SPAC(SPatial AutoCorrelation method)方法分别提取的频散曲线,展示了F-J方法在处理城市微动信号方面的优势.本文研究结果表明:(1)F-J方法可以从少量台站(21个台站)短时记录(1小时)的微动信号垂直分量中提取出清晰的基阶和高阶模态瑞利波频散曲线;(2)F-J方法提取的高阶模态频散曲线比传统SPAC方法提取的更加清晰,高频部分(>13 Hz)优势更为明显;(3)联合基阶和高阶模态频散曲线反演的浅地表速度结构更加精确,可以分辨出第四系沉积层中物性相差较小的速度界面和低速异常,在城市浅地表精细结构成像方面具有较好的应用前景.

Application of high-order surface waves in shallow exploration: An example of the Suzhou river,Shanghai

The Frequency-Bessel transform method (abbreviated as F-J method) is a new tool to analyze microtremors. This method utilizes frequency vector wavenumber transformation based on a stationary microtremor source in space and time and a flat multi-layered isotropic elastic model, so it can extract clear fundamental and higher-mode Rayleigh dispersion curves in theory. While there is no field experiment to study and apply the F-J method to engineering exploration so far. In this paper, we use the F-J method to effectively generate frequency-phase velocity spectrograms from urban microtremor recordings in the Suzhou River, Shanghai and then extract multi-mode dispersion curves. Then, the particle swarm algorithm is used to invert fundamental and higher-mode Rayleigh dispersion curves, yielding the S-wave velocity structure in depth range of 0~70 m. And the results are validated by borehole data. In addition, the comparison with traditional SPAC (Spatial Autocorrelation) method shows that the F-J method has advantages in imaging dispersion spectrogram. The results show that (1)The F-J method can effectively extract clear fundamental and higher-mode Rayleigh dispersion curves from the vertical component of urban microtremor recordings. (2)The higher-order dispersion curves extracted by the F-J method are clearer than that extracted by the traditional SPAC method, especially at high frequency (>13 Hz). (3)The inversion results using fundamental and higher-order dispersion curves are more accurate. It can identify velocity interface and low velocity anomalies in Quaternary sediments with small physical property difference. The F-J method has good prospects for imaging fine structures in shallow subsurface of urban areas.