中文摘要：

近地表结构和构造探测是研究活动断层近地表特征和工程场地地震效应的关键环节．对于城市地区，丰富的噪声限制了常规地球物理勘探方法的应用，最近的研究表明，利用噪声也可用来反演近地表结构．我们在河北保定地区，布设了一条由二百多个观测点组成的密集地震噪声剖面，探索利用噪声探测近地表结构的可能性．用噪声水平和垂直向谱比法（HVSR）得到的HVSR曲线表明，剖面普遍存在HVSR振幅显著（≥2）的峰值频率，根据其特征可以划分为A、B、C三个区．除了基岩台无明显峰值频率外，西部A区峰值频率从8Hz向东递降至0．3Hz．中部B区峰值频率在1．3Hz附近微小变化．东部C区则表现为比较宽的波峰，可能是由两个波峰组成，第一个波峰在0．8～1．0Hz之间，大部分台的HVSR曲线都可以拾取出这个波峰，第二个在0．5～0．8之间，只有少数台可以可靠地拾取．以台阵反演的速度剖面为约束，选择出适用于保定地区的峰值频率一土层厚度关系式，将剖面的峰值频率转换成土层界面的深度分布，并与两条浅层地震反射剖面进行了对比分析．结果表明，C区的两个土层界面深度分布和地震反射剖面上的二个晚更新世（Q3）界面深度（分别在约l00m和150m附近）有较好的吻合；B区得到的土层界面深度与反射剖面的界面埋深有比较明显的差别，进一步研究发现，B区观测点的地形严重影响了HVSR曲线的形态，使峰值频率可能偏离了土层的共振频率；A区土层界面从0m向东倾斜至大约500m深，但无反射剖面数据佐证．从HVSR和土层深度剖面可以推测在A区中部和A、B区交界处存在着两条垂直位错明显的正断层，但对垂直错距小的断层则无法识别．本文的结果表明，噪声HVSR方法不仅可以给出土层卓越频率等场地响应特征，还可以作为勘探工具探测近地表土?

英文摘要：

Near-surface structures exploration is an important way to deeply understand fault＇s nearsurface activities and seismic site effects in seismic hazard and risk analysis programs. In urban area, abundant noise often limits the application of traditional exploration methods. But new studies have shown that it is possible to use ambient noise to invert for near-surface structures. So we set up a profile with more than two hundred seismic noise observation sites to seek the feasibility of subsurface exploration with ambient noise. We used Horizontal to Vertical Spectral Ratio method （HVSR） to process noise data and the HVSR results show there exist peak frequencies with notable amplitude （≥2 ） along the profile which can be further delimited to three Zones A-C from west to east according to their peak frequency characteristics. The peak frequencies gradually drop easterly from 8 Hz to 0. 3 Hz excluding stations on rock without peak in Zone A, and have little fluctuation around 1.3 Hz in Zone B and have broad peak in Zone C. The broad peaks in Zone C might be composed of two peaks, one can be picked up in 0.8-1.0 Hz for many sites, and the other can be picked up reliably in 0.5-0.8 Hz only for a few sites. Using the peak frequency and soil thickness relationship appropriate for the velocity structures inverted from array data, the soil depth profile was converted directly from peak frequencies and was compared with two short seismic reflection profiles. The two derived soil interfaces in Zone C conform well to two reflection interfaces which belong to upper Pleistocene （Q3） layer （at about 100 m and 150 m respectively）. But the derived soil interface in Zone B severely deviates from reflection results, and further analysis has shown that the HVSR curve in Zone B may be seriously affected by topography, causing the peak frequency shifted from soil resonant frequency. The derived soil interface in Zone C slopes down easterly from 0 m to about 500 m depth but without further verifications. From HVSR a