台阵和噪声源分布对微动成像的影响及其在盐矿溶腔探测中的应用

随着节点地震仪的发展, 被动源地震成像方法在不同尺度的地下结构成像中得到了越来越广泛的应用.对于小尺度浅地表成像, 因高频噪声源时空分布不均, 一般采用基于台阵平均的方法, 即所谓的微动成像方法提取高频被动源面波频散数据, 但是不同类型的台阵一定程度上也会受到噪声源分布不均匀的影响.本文以水溶型盐矿溶腔探测为例, 通过实测数据较为系统地分析了不同台阵的响应函数及其对噪声源方向不均匀性的适应能力.研究发现, 当噪声源分布不均时, 三角形台阵可以获得更可靠的频散数据, 但对应的数据采集效率较低, 而其他类型的台阵虽然布设相对简便, 但是提取的相速度与真实速度存在较大的误差.当线性台阵和噪声源方位一致时, 可以提取与三角形台阵一致甚至更好的频散数据.在这个基础上, 我们分别采用三角形台阵和考虑噪声源的线性台阵在湘衡盐矿开展了两条剖面的数据采集, 并利用适应性更好的扩展空间自相关方法(ESPAC)计算面波频散数据.最后, 利用蒙特卡洛方法生成的随机初始模型进行了频散数据反演, 获得研究区600 m以浅的横波速度结构, 并刻画了盐溶腔的发育位置和空间分布形态.

The effect of different arrays and noise source distribution on microtremor imaging and its application in solute salt mine cavity detection

With the rapid development of nodal seismometers, passive source seismic imaging methods have been widely used in the imaging of subsurface structures at different scales. For small-scale near surface imaging, due to the uneven spatial and temporal distributions of high-frequency noise sources, a method so-called microtremor imaging is generally used to extract high-frequency surface wave dispersion data, which is based on array averaging. However, different types of arrays can be affected by uneven distribution of noise sources. In this study, for imaging salt mine cavities, we systematically analyze the response functions of different arrays and their ability to resolve noise sources from different azimuths. It is found that the triangular array can obtain more reliable surface wave dispersion curve, but it is more inefficient to collect the corresponding data. In comparison, for other types of arrays, although it is relatively more efficient to deploy them, the extracted surface wave phase velocities may deviate greatly from true ones. When the linear array approximately aligns with the azimuth of the steady noise source, the surface wave dispersion curve can be reliably extracted, which is consistent with or even better than that from the triangular array. Based on the above analysis, we deploy two profiles that use the triangular array and the noise source aligned linear array for microtremor imaging of cavities in the Xiangheng salt mine, respectively. The extended spatial auto-correlation method (ESPAC) which has the better adaptability is used to calculate the surface wave dispersion data. Finally, the dispersion data is inverted for shear wave velocity model with random initial models generated by the Monte Carlo method. We have obtained detailed shear wave velocity models shallower than 600 m along two profiles and determined the location of the salt cavity and its morphology.