中国地震台网地电观测以对称四极装置的地电阻率观测为主, 其观测曲线通常存在长期变化、 年变、 日变和阶跃等, 文中以宝昌台为例, 通过反演、 数值模拟等方法分析上述变化成因。 宝昌台地电布极区地下的电阻率基本呈现水平分布, 电测深曲线类型属于KH型。 该台NS、 EW测向的地电阻率自1993年至今一直存在长期下降变化, 且变化速率存在显著的各向异性, 主要是台站所在区域的应力对第3层电阻率持续作用的结果。 该台NS、 EW测向的地电阻率均存在冬春高、 夏秋低的正向年变形态和凌晨及上午高、 下午及晚间低的正向日变形态。 其中, 年变主要是温度、 降雨量季节性变化对第1层电阻率作用的结果, 日变主要是温度日变对第1层电阻率作用的结果。 该台地电阻率阶跃存在冬春频次低、 夏秋频次高的特点, 且多与降雨、 抽水和埋设钢绞线等在时间上吻合, 实验、 数值模拟等结果证实上述因素是引起地电阻率阶跃的主要干扰源。
Fixed-electrode quasi-Schlumberger arrays are mainly used in geo-electric observation of China earthquake networks. The distance between power supply poles is generally about 1km. The detection depth is estimated to be within 0.705km by conventional geophysical and electrical methods in homogeneous medium. The resistivity at seismic station for precursor information monitoring reflects the overall electrical characteristics within the detection range below the polar distribution area, which is also known as apparent resistivity or geo-resistivity. Due to the small distance between power supply poles, small detection depth and great influence from shallow layer, there are usually annual, diurnal and step variations in geo-resistivity curves. Because of the above variations, the characteristics of abnormal variations before earthquakes are usually not obvious, or even annihilated. In this paper, taking Baochang station as an example, the causes of long-term, annual, diurnal and step variations are analyzed by inversion and numerical simulation. Baochang station is located in Baochang Town, Taipusi Banner, Xilin Gol League, Inner Mongolia. Its geographical coordinates are 41.9°N and 115.3°E. The regional geological structure is the eastern segment of Inner Mongolia axis, the fourth-order structural unit. The nearest fault structure is the Chifeng-Kaiyuan Fault, which is the northern boundary fault of North China fault-block region. The resistivity of geo-electric survey area at Baochang station basically presents horizontal distribution characteristics, and the type of electric sounding curve is KH. The inversion results show that the vertical profile of the survey area is divided into four layers:the first layer is frozen soil layer with depth from 0m to 1m, the second layer is sand gravel layer with depth from 1m to 6.5m, the third layer is aquifer with depth from 6.5m to 71.5m, and the fourth layer is quartz porphyry layer with depth greater than 71.5m. When power supply electrode distance AB is 560m and measuring electrode distance MN is 80m, the one dimensional influence coefficients of NS and EW direction in the third layer are all over 0.9, which is one order of magnitude larger than those in the other three layers. This indicates that the variation of resistivity in the range of 7m to 71m can effectively reflect the variation of geo-resistivity. Since 1993, the geo-resistivity at Baochang station has been declining for a long time in NS and EW direction, and the variation rate shows obvious anisotropic characteristics, which is mainly the result of the continuous effect of regional stress on the resistivity of the third layer. There is a normal annual variation pattern of “high in winter and spring, low in summer and autumn” in both directions of geo-resistivity at Baochang station, resulting mainly from the seasonal variation of temperature and rainfall on the resistivity of the first layer. The normal diurnal variation of geo-resistivity at Baochang station is characterized by “high in the morning, low in the afternoon and night”, which is mainly caused by the influence of temperature on surface resistivity. Similar diurnal variation also exists in the hourly value curves of geo-resistivity at the stations of Xiaomiao, Ganzi, Wujiahe, Qingguang and Baodixintai. The geo-resistivity step variation of Baochang station has the characteristics of “low frequency in winter and spring, high frequency in summer and autumn”, and most of them coincide with rainfall, pumping, embedding steel strand, etc. The results of experiments and numerical simulation show that the above factors are the main interference sources of the geo-resistivity step variations.
三维几何模型:固定NS、 EW测向的电性参数和极距, 当水平尺寸为4i000m×4i000m时, 地电阻率的有限元模拟值ρ s在最底层厚度h4> 3i500m后不随h4的增加而变化; 当h4=4i000m时, ρ s在水平宽度d> 4i000m后基本不随d的增加而变化。 由此, 确定三维几何模型的尺寸为4i000m×4i000m×4i071.5m。 上述模型是基于2018年4月的观测结果通过反演建立的, 在具体研究中将根据需要对层厚、 层电阻率等参数进行调整, 建立适合研究所需的水平层状模型和三维有限元模型。
宝昌台现有地电观测装置的供电极距AB=560m、 测量极距MN=80m, 则NS测向各层的影响系数为B1=0.007、 B2=0.016、 B3=0.955、 B4=0.022, EW测向各层的影响系数为B1=0.003、 B2=0.012、 B3=0.959、 B4=0.026。 由影响系数结果(图 3)可知:1)第1层和第2层影响系数均为正, 浅层电阻率变化与地电阻率变化具有同向变化特征; 2)第3层影响系数达0.9, 比其他3层的影响系数均大1个数量级, 说明7~71m深度范围内的电阻率变化能够有效地反映在地电阻率的变化中。
表2列出了模型的第1层电阻率、 地电阻率和影响系数B1。 由此可见, 第1层电阻率下降, 则其影响系数将增大, 地电阻率下降。 其中, NS测向变化较快, EW测向变化较缓慢。
本文对宝昌台1993—2016年NS、 EW测向的地电阻率数据进行线性拟合, 得到的变化速率分别为-0.76Ω · m/a和-0.19Ω · m/a, 24a中累积变化的幅度分别为-18.24Ω · m和-4.56Ω · m。 若上述变化由4层电阻率各自贡献, 则由影响系数公式计算获得了各层的变化幅度(表3)。
由于观测环境复杂, 宝昌台的地电阻率观测值易受降雨、 钢绞线埋设、 抽水等干扰因素影响形成阶跃等。 本节统计了2012—2018年间阶跃频次与降雨量的对应情况(图 7, 8), 结果显示, 累计出现116次阶跃, 且冬春频次低、 夏秋频次高, 其中49次阶跃与测区降雨在时间上吻合。
10月12—16日, 布极区温度的逐日间差异性较小且无降雨, 其间地电阻率NS测向值无明显变化, 抽水对NS测向地电阻率观测没有影响。 12—15日, 地电阻率EW测向在每天9—17时的平均变化幅度为0.11Ω · m, 16日9—17时变化幅度达0.21Ω · m, 表明抽水对EW测向影响较大。 在EW测向测量极与供电极之间且靠近测量极处(三维影响系数为负的区域)抽水时, 位于机井下方的水位下降, 且快速形成围绕抽水机井的漏斗状水面, 局部电阻率快速上升, 造成地电阻率快速下降, 没有滞后效应。
(1)高密度电法反演结果显示, 宝昌台地电布极区地下的电阻率基本呈现水平分布, 电测深曲线类型属于KH型。 通过尝试法反演确定, 该台地电布极区的地下共分为4层, 其中第3层为含水层, 层深为6.5~71.5m, 较上覆和下伏地层整体表现为高导层, 当供电极距AB=560m、 测量极距MN=80m时, 第3层NS、 EW向的影响系数均超过0.9, 比其他3个层的影响系数均大1个数量级, 这说明第3层的电阻率变化能够有效地反映于地电阻率变化中, 该层的电阻率能有效携带应力变化、 震前异常变化等信息。
(2)宝昌台NS、 EW测向的地电阻率自1993年至今一直存在长期下降变化, 且变化速率存在显著的各向异性, 这主要是由第3层受台站所在区域的应力持续作用引起的。
(4)宝昌台地电阻率阶跃存在冬春频次低、 夏秋频次高的特点, 且多与测区内降雨、 短期抽水、 埋设钢绞线等吻合, 通过数值模拟、 三维影响系数计算等方法对典型阶跃进行定性和定量角度分析的结果也印证了这一统计结果。
致谢 中国地震局兰州地震研究所杜学彬研究员在成文过程中给予了悉心指导; 中国地震台网中心解滔副研究员为本研究提供了部分计算程序; 内蒙古自治区宝昌地震台贾昕晔高级工程师、 高昌志高级工程师为本研究提供了基础资料; 审稿专家为本文提出了宝贵的修改建议。 在此一并表示感谢!