TWO COMPUTER PROGRAMS FOR THE CALCULATION OF STANDARD GRAPHS FOR RESISTIVITY PROSPECTING*

This paper includes publication of two algorithms, to be processed on a digital computer. The algorithms are based upon the two mathematical denotations, presented by Dr. J. C. van Dam in this same issue. The computer programs are written in Algol, in the form of procedures (i.e. subroutines to be activated from a surrounding Algol program). Each of the two procedures is preceded by a numerical analysis. Some remarks are made about the reliability of each of the two procedures, the production and the computing time required.     

AN EXTENSION OF THE BORN INVERSION METHOD TO A DEPTH DEPENDENT REFERENCE PROFILE*

An extension of the multidimensional Born inversion technique for acoustic waves is described. In earlier work, a perturbation in reference sound velocity was determined by assuming that the reference velocity was constant. In this extension, we allow the reference velocity to be a function of the depth variable z. The output of this method is a high-frequency bandlimited reflectivity function of the subsurface. The reflectivity function is an array of bandlimited singular functions scaled by the normal reflection strength. Each singular function is a Dirac delta function of a scalar argument which measures distance normal to a reflecting interface. Thus, the reflectivity function is an indicator map of subsurface reflectors equivalent to the map produced by migration. In addition to the assumption of small perturbation, the method requires that the reflection data reside in the high frequency regime in a well-defined sense. The method is based on the derivation of an integral equation for the perturbation in sound velocity from a known reference velocity. When the reference velocity is constant, the integral equation admits an analytic solution as a multifold integral of the reflection data. Further high frequency asymptotic analysis simplifies this integral considerably and leads to an extremely efficient numerical algorithm for computing the reflectivity function. The development of a computer code to implement this constant-reference-velocity solution is published elsewhere. For a reference velocity c(z) we can no longer invert the integral equation exactly. However, we can write down an asymptotic high-frequency approximation for the kernal of the integral equation and an asymptotic solution for the perturbation. The computer implementation of this result is designed along the same lines as the code for constant background velocity. In tests the total processing time for this algorithm with depth-dependent background velocity is usually considerably less than that required by a standard Kirchhoff migration algorithm. The method is implemented as a migration technique and compared with alternative migration algorithms on the flanks of the salt dome.     

一种适于近场理论地震图宽频域计算的矩阵分解算法

理论地震图方法在地震震源过程的研究中得到了广泛的应用。为了研究震源过程的细节,必须利用近场地震资料的高频信息,但是在应用Haskell矩阵法计算近场理论地震图时,格林函数的高频成分的数值不稳定性是计算宽频域理论地震图的一个基本困难。因此,目前仅有不超过10Hz的算例。本研究采用Haskell矩阵的一种新的分解组合形式,在数值计算时有效地避免了在计算过程中数值结果的溢出,实现了近场理论地震图的宽频域计算。同时,由于利用了矩阵运算中的解析关系,减少了运算的次数,从而提高了计算的速度。根据本算法建立的Fortran程序,在Univoc-1100计算机上进行了数值检验。结果表明,计算得到的格林函数至少在0-40Hz频率域范围内仍保持良好的数值稳定性。本文给出的这种Haskell矩阵分解组合形式的矩阵元素简单,且具有一定的对称性。对于这种分解所包含的物理意义,将有待进一步深入研究。     

有序波前重建法的射线追踪

建立了一种新的计算最小走时和射线路径的方法——有序波前重建法. 文中算法按照波前面的实际扩展顺序外推计算走时，采用以计算点为中心的走时计算策略，直接记录计算点获取最小走时的前一节点坐标，同步计算最小走时和射线路径，得到一种全局算法. 该方法具有原理简单、易于实现、不受介质速度差异大小限制、计算速度快等优点. 数值实验表明有序波前重建法具有较高的计算精度和运行效率.     

点电源二维起伏界面三层介质边界单元法

本文在文献[5]的基础上,以三层介质下边两层界面上隆模型为例,讨论了边界元法在解决层状介质中存在二维横向不均匀情况下点源场视电阻率计算问题方面的应用。在本文中,由于选取了适当的基本解,故在所导出的方程中避免了对水平层状界面的积分计算,只对起伏边界进行积分计算,并在理论推导中设法消除了位场在边界上的法向导数,因而最后形成的代数方程组规模相当小,从而大大减少了所需的边界部分点和未知量的数目,大量地节省了计算机内存,加快了计算速度。这给数值计算带来了很大的方便,并为向二维及三维多层介质中存有异常体的问题推广提供了便利条件。     

地震波传播的哈密顿表述及辛几何算法

地震波传播过程本质上是能量在传播过程中逐步损耗直至殆尽的过程，而在实际应用中，常在无能量损耗假设下，用弹性波动方程或标量波动方程描述它.在哈密顿(Hamilton)体系表述下，地震波传播过程即为一个无限维的哈密顿系统随时间的演化过程.若不计能量损耗，波场演化过程实质上为一个单参数连续的辛变换，因而对应的数值算法应为辛几何算法.本文首先从地震波标量方程出发，给出哈密顿体系下地震波传播的表述，即任意两个时刻的波场是通过辛变换联系起来的.随后，把波场在时间和相空间离散化后，给出了用于波场计算的一些辛格式，如显式辛格式、隐式辛格式和蛙跳辛格式.并进一步讨论了有限差分格式和辛格式的异同.然后，应用显式辛格式和同阶的有限差分方法给出了同一理论速度模型下的波场和Marmousi速度模型下的单炮记录.数值结果表明，辛算法是一类可行的波场模拟的数值算法.在时间步长较小时，有限差分方法是辛算法的一个很好近似.文中的理论和方法，为地震波传播理论及实际应用研究提供了新的途径.     

利用不完全Cholesky共轭梯度法求解点源三维地电场

点源三维地电场的求解是一大型数值计算问题．本文用有限差分方法求解,最后形成一个线性方程组Ax＝b,这里A是大型稀疏的带状对称矩阵．解大型稀疏方程组的完全Cholesky分解直接算法,一般要求巨大的机器内存来存储系数矩阵A,而且计算速度极慢．因此引入不完全Cholesky共轭梯度（ICCG）算法及按行索引的稀疏存储模式,充分利用A的稀疏性,使得计算速度大大提高,而内存要求则大大减少,因此ICCG算法是地电三维正演的强有力工具．     

AN INTERPRETATION SYSTEM FOR GEO-ELECTRICAL SOUNDING GRAPHS *

For the two and three layer cases geo-electrical sounding graphs can be rapidly and accurately evaluated by comparing them with an adequate set of standard model graphs. The variety of model graphs required is reasonably limited and the use of a computer is unnecessary for this type of interpretation. For more than three layers a compilation of model graphs is not possible, because the variety of curves required in practice increases immensely. To evaluate a measured graph under these conditions, a model graph is calculated by computer for an approximately calculated resistivity profile which is determined, for example, by means of the auxiliary point methods. This model graph is then compared with the measured curve, and from the deviation between the curves a new resistivity profile is derived, the model graph of which is calculated for another comparison procedure, etc. This type of interpretation, although exact, is very inconvenient and time-consuming, because there is no simple method by which an improved resistivity profile can be derived from the deviations between a model graph and a measured graph. The aim of this paper is, on the one hand, to give a simple interpretation method, suitable for use during field work, for multi-layer geo-electrical sounding graphs, and, on the other hand, to indicate an automatic evaluation procedure based on these principles, suitable for use by digital computer. This interpretation system is based on the resolution of the kernel function of Stefanescu's integral into partial fractions. The system consists of a calculation method for an arbitrary multi-layer case and a highly accurate approximation method for determining those partial fractions which are important for interpretation. The partial fractions are found by fitting three-layer graphs to a measured curve. Using the roots and coefficients of these partial fractions and simple equations derived from the kernel function of Stefanescu's integral, the thicknesses and resistivities of layers may be directly calculated for successively increasing depths. The system also provides a simple method for the approximative construction of model graphs.     

THE SPECTRAL FUNCTION OF A LAYERED SYSTEM AND THE DETERMINATION OF THE WAVEFORMS AT DEPTH*

Consider the mathematical model of a horizontally layered system subject to an initial downgoing source pulse in the upper layer and to the condition that no upgoing waveforms enter the layered system from below the deepest interface. The downgoing waveform (as measured from its first arrival) in each layer is necessarily minimum-phase. The net downgoing energy in any layer, defined as the difference of the energy spectrum of the downgoing wave minus the energy spectrum of the upgoing wave, is itself in the form of an energy spectrum, that is, it is non-negative for all frequencies. The z-transform of the autocorrelation function corresponding to the net downgoing energy spectrum is called the net downgoing spectral function for the layer in question. The net downgoing spectral functions of any two layers A and B are related as follows: the product of the net downgoing spectral function of layer A times the overall transmission coefficient from A to B equals the product of the net downgoing spectral function of layer B times the overall transmission coefficient from B to A. The net downgoing spectral function for the upper layer is called simply the spectral function of the system. In the case of a marine seismogram, the autocorrelation function corresponding to the spectral function can be used to recursively generate prediction error operators of successively increasing lengths, and at the same time the reflection coefficients at successively increasing depths. This recursive method is mathematically equivalent to that used in solving the normal equations in the case of Toeplitz forms. The upgoing wave-form in any given layer multiplied by the direct transmission coefficient from that layer to the surface is equal to the convolution of the corresponding prediction error operator with the surface seismogram. The downgoing waveform in this given layer multiplied by the direct transmission coefficient from that layer to the surface is equal to the convolution of the corresponding hindsight error operator (i.e., the time reverse of the prediction error operator) with the surface seismogram.     

层状方位各向异性介质的视电阻率计算

从电性各向异性的欧姆定律出发，推导了直流电法层状方位各向异性介质中的电位分布、边界条件及视电阻率计算公式．以四极对称装置系统为例，对具有相同各向异性系数的4层模型采用核函数递推法作了理论数值模拟，得到了不同方向的电阻率测深曲线及其等值线形态．结果表明理论公式是正确的，测深曲线既反映了分层介质的电阻率差异，又反映了各层中电阻率的各向异性特征．      

可控源音频大地电磁数据的反演方法

从反演方程、构造目标函数和求解三方面对用于可控源音频大地电磁法(CSAMT)的实用反演方法中的四种进行了描述．水平层状地层CSAMT法资料的直接反演法首次尝试了一维空间的全资料CSAMT反演，效果较好，但该方法尚难应用于2D、3D复杂介质中；奥克姆反演方法既考虑了横向的光滑函数，又考虑了纵向的光滑函数，得到比较光滑的横向、纵向变化的背景电性结果，但有可能把一些小构造光滑掉．快速松驰反演算法和共轭梯度算法由于计算速度快，占内存少而被用于三维反演中，二者相比，快速松驰算法在求解雅可比矩阵时只做一次正演计算，在更新模型时解小型方程组，所以在速度上更胜一筹．在后三种算法中，由于复杂电性结构无解析解，正演计算都采用数值计算．数值计算的可靠性、速度影响着反演算法的有效性，这方面的研究也将是2D、3D复杂电性结构反演的研究方向之一.     

DIRECT INTERPRETATION OF TWO-ELECTRODE RESISTIVITY SOUNDINGS *

This paper describes the procedure for interpreting the apparent resistivity data measured with the two-electrode array directly with the help of kernel function. The calculation of kernel function from the observed resistivity curve is done by the method of decomposition. In the method of decomposition the resistivity curve is approximated by a sum of certain functions, whose choice is only restricted by the requirement that the contribution to the kernel function corresponding to them should be easily computable. A few such functions are classified. These, and the standard curves for corresponding kernel functions obtained by utilising an integral expression for two-electrode array expressing the kernel explicitly in terms of the apparent resistivity functions, are plotted on log-log scale. The determination of layer parameters, that is, the layer resistivities and thicknesses from the kernel function can be carried out by a method proposed by Pekeris (1940).     

ON THE POSSIBILITY OF SEISMIC EXPLORATION USING SURFACE TORQUE SOURCE AND RELATED TOPICS*

In this study we derive expressions for particle displacement or particle velocity anywhere inside a stratified earth and at its surface due to horizontal torque source located in the top layer. Equivalently, invoking Green's function reciprocity theorem, the solution applies also to the case of a surface or subsurface source when the resulting displacement or velocity is measured within the top layer. In order to evaluate the closed-form analytical solution economically and accurately it is advisable to introduce inelastic attenuation. Causal inelastic attenuation also lends the necessary realism to the computed seismic trace. To provide proof that the analytical solution is indeed correct and applicable to the multilayer case, a thick uniform overburden was assumed to consist of many thin layers. The correctness of the computed particle velocity response can be very simply verified by inspection. The computed response can also serve as a check on other less accurate methods of producing synthetic seismograms, such as the techniques of finite differences, finite elements, and various sophisticated ray-tracing techniques. It is not difficult to construct horizontal surface torque source. It appears that such source is well suited for seismic exploration in areas with a high-velocity surface layer. A realistic source function is analyzed in detail and normalized displacement response evaluated at different incidence angles in the near and the far fields. In an effort to distinguish the features of an SH torque seismogram from a pressure seismogram two models with identical layerings and layer parameters have been set up. As expected the torque seismogram is very different from the compressional seismogram. One desirable feature of a torque seismogram is the fast decay of multiples. Exact synthetic seismograms have many uses; some of them, such as the study of complex interference phenomena, phase change at wide angle reflection, channeling effects, dispersion (geometrical and material), absolute gain, and inelastic attenuation, can be carried out accurately and effortlessly. They can also be used to improve basic processing techniques such as deconvolution and velocity analysis. The numerical evaluation of the analytical solution of the wave equation as described in this paper has a long history. Most of the work leading to this paper was carried out by one of us (M. J. K.) in the years 1957 to 1968 at the Geophysical Research Corporation. However, the full testing of the various computer codes was carried out only very recently at the Phillips Petroleum Company.     

位场DFT算法研究

本文推广了经典的抽样定理,并据此导出了函数有限离散傅里叶变换误差方程（简称DFT误差方程,下同）。该方程把有限离散傅里叶变换中固有的离散效应和有限效应表示为确切的数学形式。离散效应被表示为一个含整参变量（参变量取0,1,…,N-1）的复无穷级数;有限效应被表示为一个含整参变量（参变量取0,1,…,N-1）的复无穷级数的DFT。 基于DFT误差方程和位函数特点,作者提出了两种位场数值傅里叶变换新算法--移样法和等效源续尾叠样法。移样法可近百倍地提高位场数值傅里叶反变换的精度,等效源续尾叠样法可数十倍地提高正变换精度。两种算法都不需要增加资料长度和取样密度,因而基本不需要增加计算机时间和内存。文中给出了算例。     

隐伏层的形成与识别

利用折射波场的特点，研究隐伏层，建立不同性质中间层的临界方程和隐伏层形成条件的公式．研究表明，隐伏层形成与剖面中每个地层的厚度、波速、以及折射面的倾角有关．厚度薄而波速低的中间层容易形成隐伏层．还用τ－ｐ变换方法来识别和提取隐伏层的视速度与截距时间，并给出了模型的结果．     

地电阻率季节干扰变化分析

结合我国实际,本文提出四类地电阻率季节变化的干扰模型.经模型理论计算及同有季节变化的四十多台站的实测资料对比得出:(1)地电阻率季节变化是探测深度偏浅时,地表层电性变化,特别是表层地下水水位年动态变化所引起的一种与地震无关的干扰变化.其变化形态与量级取决于干扰模型、地电断面类型以及探测深度.(2)通常地电阻率季节变化与浅层电阻率变化的符号一致,呈现出正常季节变化;但对 K,Q 型地电断面,当干扰层位于地表第一层时,将从某一供电极距开始,出现反常的季节干扰变化.(3)当探测深度大于300m 时,不论正常的还是反常的季节变化,其幅度都将小于2%.因此,只要适当选择台址或加大供电极距使探测深度足够大,便可排除或压低这种干扰.      

测试水井导水性能的简便试验

这种测试水井含水层导水系数的新方法是利用井口水面空气压力阶变试验,计算出聊城井含水层的导水系数为14.4m2/d,而通过抽水试验得出的导水系数为43.6m2/d,二者有相同的数量级,但新方法比抽水试验简单易行和经济快速,用它来解释地下水微动态更合理,容易推广应用     

过渡层中地震P-SV波广义反射系数的两种近似算式

本文导出计算过渡层介质中P-SV波广义反射系数的两种近似算式:一种以层状介质作为过渡层的近似模型;另一种是直接由过渡层模型用数值近似方法计算。文中对比分析了数值试验结果。     

确定地壳形变观测曲线趋势转折点的定量方法与实例

趋势转折是地壳形变变化的一个重要信息,为定量、精确地提取地壳形变观测资料的趋势转折点,基于改进的有序聚类分析法,采用时窗滑动的方法,提取时间序列的趋势转折点,并编制了相应的程序.以新源地震台地倾斜南北向1992～2006年的观测数据为例,以4年窗长、2年步长滑动计算其趋势转折点.研究结果表明：（1）采用改进的有序聚类分析法提取时间序列的趋势转折点具有明确的统计意义,可减少人为判断的主观和任意性.（2）该方法算法简单,易于实现,可用于日常的地壳形变观测资料的分析预报工作.（3）新源地震台地倾斜南北向观测数据计算结果表明,在2001年之前,趋势转折比较频繁,变化也比较明显,而在2001年之后,趋势转折较少,变化更接近线性.     

基于螺旋线上谱因式分解的地震波场隐式辛算法

均匀介质、复杂各向同性介质和各向异性介质中的地震波传播过程,可用统一形式的标量声波方程描述.考虑到在无损耗条件下,地震波方程描述了地震波场这一个无穷维的哈密顿体系随时间的演化过程,该过程为一个单参数连续辛变换,因而可以在其哈密顿形式表述下导出其辛格式.与显式辛算法相比,隐式辛格式对应的隐式辛几何算法具有无条件稳定的特点,可以允许较大的计算步长.但是由于隐式算法不可避免地面临高阶矩阵的求逆,其每一步的计算速度较慢.为实现矩阵快速求逆,文中采用了螺旋边界条件下谱因式分解的方法.在螺旋边界条件下,需要求逆的矩阵化为带状矩阵,而且其各列非零元素的位置和大小具有非常好的相似性,因而可以采用谱因式分解的方法实现快速LU分解.文中采用二阶精度的隐式蛙跳辛格式和谱因式分解方法,计算了常速度、层状介质和Marmousi模型中的波场.计算表明,隐式辛算法不失为波场计算的一种好方法.