大地电磁、重力、磁法和地震初至波走时的交叉梯度二维联合反演研究

为了准确的探测和描绘地下复杂的地质结构,同时克服地球物理单一方法反演的多解性和单一参数反演模型的不一致性等问题,近年来基于交叉梯度联合反演的综合地球物理解释已经得到了广泛的关注和应用.本文首先研究了两种地球物理方法的交叉梯度联合反演算法,在此基础上,推导并实现了多种地球物理方法(大地电磁,重力,磁法,地震初至波走时)的多交叉梯度约束的二维联合反演算法;其次,我们设计了结构不一致模型和复杂模型,针对多物性联合反演算法的准确性和有效性进行了模拟试算,并对复杂模型的单独反演结果和联合反演结果进行了交叉梯度值和物性交会图的对比;最后,本文将成熟的卫星资料多光谱综合分析技术应用到联合反演中,将多物性参数反演模型结果图通过RGB(红-绿-蓝)模式进行合成,得到融合的RGB合成图.结果表明:通过对结构不一致模型和复杂模型的联合反演结果和单独反演结果的对比分析,可以得出联合反演得到的结果更接近真实模型,并从得到的交叉梯度值进一步证明了联合反演模型相似度高,也从物性交会图中得到联合反演的物性相关性更好的结论,反向证明了算法的正确性.最终从得到的RGB合成图像,我们可以更直观的分析反演结果,更有利于准确划分地下模型结构.     

基于交叉梯度约束的重力、磁法和大地电磁三维联合反演

为了降低单一地球物理方法反演的多解性及受噪声的影响程度,本文围绕重力、磁法和大地电磁法开展了三维联合反演的研究.重、磁采用基于对数障碍法的正则化反演算法,大地电磁使用limited-memory BroydenFletcher-Goldfarb-Shanno(L-BFGS)反演算法,引入交叉梯度函数实现了三种物性结构的相互耦合,最终开发出一套重磁电三维联合反演算法,并实现MPI并行加速计算.通过理论模型算例验证了算法的准确性,结果表明:不论是单棱柱体模型还是组合棱柱体模型,联合反演结果相较单独反演对于异常体的空间形态刻画以及物性数值恢复具有较好的提升;单棱柱体模型算例使得异常体的物性参数(密度、磁化率和电阻率)更加接近于真实的物性参数;组合棱柱体模型的联合反演结果不仅仅消除了围岩物性参数的假异常,而且还增强了异常体边界结构的恢复程度.     

电磁与地震联合反演研究现状与未来发展

不同的地球物理方法有各自的应用、优势和局限性,正是因为局限性的存在,单一地球物理方法在地球物理反演中具有一定的不足,例如反演精确度等,因此将多种地球物理方法进行联合反演要比单一方法反演具有一定的优越性.本文介绍了联合反演的分类方法,说明电磁与地震理论的相似性,以及物性上的联系,重点分析近几年国内外电磁和地震联合反演的研究进展、现状及存在的问题.对于不同类型物性参数在联合反演中的耦合问题,指出岩石物理方法适用范围.最后预测电磁和地震联合反演的发展方向,包括优化算法从线性到非线性,进一步到更高级算法;进行更多方法结合的联合反演;在维度上的拓展,一维到二维、三维,甚至四维.另外,加强联合反演岩石物性等先验信息的结合以及结构法反演前景广阔,需要进行更多实际应用验证.     

重震联合中速度-密度耦合研究综述

基于物性参数耦合的多地球物理数据联合反演方法是21世纪初发展起来的新技术,速度-密度耦合约束下的重震联合反演是其重要分支之一.相比于传统的重震资料综合解释,基于速度-密度耦合的重震联合反演能够减少主观因素的干扰,发挥地震和重力数据的互补作用,产生精度和一致性更高的速度-密度模型.结合国内外现状,本文较为全面地介绍了现有的速度-密度耦合方式,并讨论了速度-密度耦合约束下重震联合反演策略和目标函数的构建及求解等相关问题.不同反演策略和耦合方式的适用性不同,没有绝对的优劣之分.根据研究区的实际情况,在合适的速度-密度耦合约束下开展重震联合反演研究和应用是下一步工作的重点.     

部分区域约束下的交叉梯度多重地球物理数据联合反演

交叉梯度联合反演方法通过对多种地球物理模型实现结构耦合,在岩石物性关系不确定的情况下,既能提高反演结果的可靠性,又能减少反演的多解性,还能减少不同方法解释结果之间的矛盾.当不同的模型观测数据覆盖范围不一致时,交叉梯度联合反演通常需要取出重叠区域数据进行联合反演,并且建模时还要扩展一些模型范围.本文首先提出并实现了部分区域约束下的交叉梯度多重地球物理数据联合反演算法;接着进行了算法的模型试算;最后,我们将该反演算法用于本溪—集安深部地质调查重磁电综合地质地球物理解释中.结果表明:该算法不但能在重叠区域内很好地恢复结构相似的模型,而且在非重叠区域与重叠区域的边界处仍然可以得到平滑变化的模型;在本溪—集安10号剖面所获得的结构上相似的电阻率、密度及磁化率模型较好地反映了该区的深部地质结构,对于确定深部地质体的性质提供了有力的证据.     

砂泥岩储层孔隙度和含水饱和度同步反演

基于岩石物理和地震反演理论,提出了一种同步反演储层孔隙度和含水饱和度的方法.以岩石物理为基础,建立了砂泥岩储层物性和弹性参数之间定量的关系-Simon模型,以贝叶斯理论为手段,结合不同类型的砂泥岩储层,建立了多信息联合约束的物性参数反演目标函数,并通过蒙特卡罗和遗传算法相结合的思路求解该目标函数,最终得到孔隙度和含水饱和度的同步反演结果.将该方法应用于河道砂和砂砾岩两种不同的砂泥岩储层中,孔隙度和含水饱和度数据的联合应用,进一步减少了储层预测的多解性,为石油地质综合研究提供了更加丰富准确的基础数据.     

重震联合反演方法及其应用进展

地球物理反演是探索地下结构的最佳途径之一.地震波可以穿透到地球深部进行直接采样,是探测地球深部的主要方法.重力是结构体密度分布与地表观测点之间距离的体积积分效应,重力异常随着源深度的增加衰减很快,其对浅部结构的灵敏度明显优于地震数据.地震和重力联合反演能够相互补充和约束,提高空间分辨率,使反演结果更加稳定可靠.本文首先介绍了联系地震和重力数据的速度-密度经验关系,随后分别介绍了重震联合反演的3种常用方法—顺序反演、同步反演和交叉梯度反演,简要阐述了各种方法在国内外的应用情况.分析认为顺序反演将两类数据分开独立进行计算,原理简单,易于操作实现.但是该方法依赖于先验模型和速度-密度经验关系,可能存在分辨率较低区域(如模型边界)的误差放大效应.同步反演采用将地震和重力数据放在同一方程组中同时反演的策略,减弱了单一数据先验模型对结果的影响,但两种数据的同时运用势必引入数据权重分配问题.交叉梯度寻求不同物理参数模型在结构上的相似性,对潜在的岩石物性关系做了最少的假设,一定程度上降低了反演的非唯一性,但强制性地匹配模型的结构不一定完全符合地下介质的物性分布.因此使用交叉梯度方法反演时应注意模型的推导需要遵循客观标准,以控制模型的结构相似性和数据拟合度.最后指出重震联合反演中的速度-密度经验关系和数据的权重分配仍是值得探究的问题.     

基于交叉梯度交替结构约束的二维地震走时与全通道直流电阻率联合反演

在利用不同的地球物理勘探方法对地下复杂介质成像时,因观测系统的非完备性及数据本身对某些岩石物性的不敏感性,单独成像的结果存在较大的不确定性和不一致性.对于地震体波走时成像与直流电阻率成像,均面临着成像阴影区问题.对于地震走时成像,地震射线对低速区域覆盖较差形成阴影区,造成低速区域分辨率降低.对于电阻率成像,电场线在高阻区域分布较少,造成高阻区域分辨率较低.为了提高地下介质成像的精度,Gallado和Meju(2003)提出了基于交叉梯度结构约束的联合地球物理成像方法.在要求不同的物性模型拟合各自对应的数据同时,模型之间的结构要求一致,即交叉梯度趋于零.为了更有效地实现基于交叉梯度的结构约束,我们提出了一种新的交替结构约束的联合反演流程,即交替反演不同的数据而且在反演一种数据时要求对应的模型与另一个模型结构一致.新的算法能够更容易地把单独的反演系统耦合在一起,而且也更容易建立结构约束和数据拟合之间的平衡.基于新的联合反演流程,我们测试了基于交叉梯度结构约束的二维跨孔地震走时和直流电阻率联合成像.合成数据测试表明,我们提出的交替结构约束流程能够很好地实现基于交叉梯度结构约束的联合成像.与单独成像结果相比,地震走时和全通道电阻率联合成像更可靠地确定了速度和电阻率异常.     

储层弹性与物性参数地震叠前同步反演的确定性优化方法

储层弹性与物性参数可直接应用于储层岩性预测和流体识别,是储层综合评价和油气藏精细描述的基本要素之一.现有的储层弹性与物性参数地震同步反演方法大都基于Gassmann方程,使用地震叠前数据,通过随机优化方法反演储层弹性与物性参数;或基于Wyllie方程,使用地震叠后数据,通过确定性优化方法反演储层弹性与物性参数.本文提出一种基于Gassmann方程、通过确定性优化方法开展储层弹性和物性参数地震叠前反演的方法,该方法利用Gassmann方程建立储层物性参数与叠前地震观测数据之间的联系,在贝叶斯反演框架下以储层弹性与物性参数的联合后验概率为目标函数,通过将目标函数的梯度用泰勒公式展开得到储层弹性与物性参数联合的方程组,其中储层弹性参数对物性参数的梯度用差分形式表示,最后通过共轭梯度算法迭代求解得到储层弹性与物性参数的最优解.理论试算与实际资料反演结果证明了方法的可行性.     

基于改进差分进化算法的大地电磁测深和重力宽范围物性约束联合反演

宽范围物性约束技术容易实现、具有一定容错性,目前已在大地电磁测深（MT）和地震、MT和重力联合反演中实现,但该技术是结合模拟退火算法实现的.差分进化算法（DE）是一种全局优化算法,但该算法在地球物理联合反演领域应用较少.基于此,本文以双种群设置方案为框架改进了DE算法,并提出了基于改进DE算法的宽范围物性约束技术.MT和重力联合反演的模型试验表明：与传统的DE算法相比,改进的DE算法收敛速度更快,寻优能力更强;基于改进DE算法的宽范围物性约束技术可以促进不同岩石物性参数在一定"范围"内实现耦合,既可以利用岩石物性关联的导向作用,又可以发挥优化算法的寻优能力,进而降低地球物理联合反演对先验信息的要求;此外,该技术的实现也验证了宽范围物性约束思想在联合反演领域中的适用性,具有进一步推广至其他优化算法中的潜质.     

A Monte Carlo study of rainfall forecasting with a stochastic model

A procedure for short-term rainfall forecasting in real-time is developed and a study of the role of sampling on forecast ability is conducted. Ground level rainfall fields are forecasted using a stochastic space-time rainfall model in state-space form. Updating of the rainfall field in real-time is accomplished using a distributed parameter Kalman filter to optimally combine measurement information and forecast model estimates. The influence of sampling density on forecast accuracy is evaluated using a series of a simulated rainfall events generated with the same stochastic rainfall model. Sampling was conducted at five different network spatial densities. The results quantify the influence of sampling network density on real-time rainfall field forecasting. Statistical analyses of the rainfall field residuals illustrate improvement in one hour lead time forecasts at higher measurement densities.     

Earthquake Engineering and Engineering Vibration Aims And Scope

正This journal is established by the Institute of Engineering Mechanics(IEM),China Earthquake Administration,to promote scientific exchange between Chinese and foreign scientists and engineers so as to improve the theory and practice of earthquake hazards mitigation,preparedness,and recovery.To accomplish this purpose,the journal aims to attract a balanced number of papers between Chinese and     

Solute concentrations in the lower niger river and the source rock contribution

Water quality analyses for the Niger River for the 1980/81 hydrological year are presented. The samples were collected from the main river at Lokoja, and from two main tributaries, the Kaduna and the Benue Rivers. Different water types were distinguished by the concentrations of major ions. The type Ca > Na > Mg > K - HCO₃ > SO₄ > Cl was represented at all stations during at least part of the year. Chloride was found to dominate the sulphate ion in the Kaduna and Niger, while the Benue maintained a higher concentration of sulphate relative to chloride all year round. Distinct patterns of seasonal variation in the ion concentrations were observed, particularly for the samples collected at Lokoja. Low ion concentrations were prominent during periods of high discharge, while low flow periods coincided with high dissolved ion concentrations. The contribution of rainwater to the total dissolved solids in the river waters was assessed indirectly using rainwater chemistry data from the Gulf of Guinea. The estimated rainwater contribution to the Lower Niger amounts to 5.15 mg 1^?1. Geochemical weathering calculations involving reactions of the four major minerals of granitic rocks - anorthite, biotite, albite, and K-feldspar - with carbon dioxide and water, can account for the average water composition of the Lower Niger. The proportion of the ionic components was also related to the occurrence of the respective element in the minerals.     

Scenario predictions for potential near-field and far-field earthquakes affecting Hong Kong

Historical earthquakes noted in the written records of the South China region, including Hong Kong, are not well delineated along identified prominent fault sources. Despite the lack of any definitive, localised trend in the spatial distribution of seismic activity in the region, there does appear to be some major disparity in the seismic activity rates (especially for large magnitude earthquakes) between the near-field and the far-field regions of Hong Kong. Despite this observation, previous studies of the regional seismic activity and seismic ground motion hazard (the latter using a probabilistic seismic hazard assessment, PSHA) have considered very broad source zone regions, in which uniform levels of seismic activity have been assumed. The present paper further scrutinises this broad source zone (BSZ) approach by adopting a novel expanding circular disc (ECD) method to determine the rates of earthquake recurrence. Such a method is intended to counter-check previously developed models by determining earthquake scenario events in terms of magnitude–distance (M–R) pairs or combinations, having defined values of average return period. Unlike the BSZ approach, the ECD method specifically accounts for the supposed variations in the seismic activity rates between events in the near-field and the far-field of Hong Kong.The form of the developed method is particularly suited to the determination of design-level earthquake ground motions for bedrock sites, since it assumes a directionally-independent attenuation model as described in the companion paper. It is found that, whilst the BSZ approach may indicate the overall average levels of hazard that are representative of the South China region as a whole, it does not capture the large disparity in seismic activity rates between near-field and far-field events. This important feature is expected to have a significant impact upon engineering assessments of the seismic safety of structures in Hong Kong and elsewhere in the South China region. For example, it is found that for events with M≥6, the seismic activity rate (normalised by time and area) in the very far-field is around 3.5 times larger than in the near-field and medium-field of Hong Kong. The resulting design M-R combinations, covering a range of return periods from 70 to 2500 years, are limited, for very long return periods and for distant events, by the maximum credible earthquake (MCE) magnitude. Intensive research to determine this seismic hazard parameter is recommended, in order to refine further the results of the ECD analysis, which presently conservatively assumes the MCE to range between M=6 in the near-field of Hong Kong to M=8 in the very far-field, at distances greater than 280 km from Hong Kong.     

Depositional ages and characteristics of Middle–Upper Jurassic and Lower Cretaceous lacustrine deposits in southeastern Mongolia

Lower Cretaceous lacustrine oil shales are widely distributed in southeastern Mongolia. Due to the high organic carbon content of oil shale, many geochemical studies and petroleum exploration have been conducted. Although most of the oil shales are considered to be Early Cretaceous in age, a recent study reveals that some were deposited in the Middle Jurassic. The present study aims at establishing depositional ages and characteristics of the Jurassic and Cretaceous lacustrine deposits in Mongolia. The Lower Cretaceous Shinekhudag Formation is about 250 m thick and composed of alternating beds of shale and dolomite. The Middle Jurassic Eedemt Formation is about 150 m thick and composed of alternating beds of shale, dolomitic marl, and siltstone. The alternations of shale and dolomite in both formations were formed by lake level changes, reflecting precipitation changes. Shales were deposited in the center of a deep lake during highstand, while dolomites were formed by primary precipitation during lowstand. Based on the radiometric age dating, the Shinekhudag Formation was deposited between 123.8 ±2.0 Ma and 118.5 ±0.9 Ma of the early Aptian. The Eedemt Formation was deposited at around 165–158 Ma of Callovian–Oxfordian. The calculated sedimentation rate of the Shinekhudag Formation is between 4.7 ±2.6 cm/ky and 10.0 ±7.6 cm/ky. Shales in the Shinekhudag Formation show micrometer‐scale lamination, consisting of algal organic matter and detrital clay mineral couplets. Given the average thickness of micro‐laminae and calculated sedimentation rate, the micro‐lamination is most likely of varve origin. Both Middle–Upper Jurassic and Lower Cretaceous lacustrine oil shales were deposited in intracontinental basins in the paleo‐Asian continent. Tectonic processes and basin evolution basically controlled the deposition of these oil shales. In addition, enhanced precipitation under humid climate during the early Aptian and the Callovian–Oxfordian was another key factor inducing the widespread oil shale deposition in Mongolia.     

THEORETICAL APPROACH TO THE SOLUTION OF THE INFILTRATION PROBLEM

ABSTRACT

The one-dimensional transient downward entry of water in unsaturated soils is investigated theoretically. The mathematical equation describing the infiltration process is derived by combining Darcy's dynamic equation of motion with the continuity and thermodynamic state equations adjusted for the unsaturated flow conditions. The resulting equation together with the corresponding initial and boundary conditions constitues a mathematical initial boundary value problem requiring the solution of a nonlinear partial differential equation of the parabolic type. The volumetric water content is taken as the dependent variable and the time and the position along the vertical direction are taken as the independent variables. The governing equation is of such nature that a solution exists for t > 0 and is uniquely determined if two relationships are defined, together with the specified state of the system, at the initial time t = 0 and at the two boundaries. The two required relations are those of pressure versus permeability and pressure versus volumetric water content.

Since the partial differential equation has strong non-linear terms, a discrete solution is obtained by approximating the derivatives with finite-differences at discrete mesh points in the solution domain and integrated for the corresponding initial and boundary conditions. The use of an implicit difference scheme is employed in order to generate a system of simultaneous non-linear equations that has to be solved for each time increment. For n mesh points the two boundary conditions provide two equations and the repetition of the recurrence formula provides n—2 equations, the total being n equations for each time increment. The solution of the system is obtained by matrix inversion and particularly with a back-substitution technique. The FORTRAN statements used for obtaining the solution with an electronic digital computer (IBM 704) are presented together with the input data.

Analysis of the errors involved in the numerical solution is made and the stability and convergence of the solution of the approximate difference equation to that of the differential equation is investigated. The method applied is that of making a Fourier series expansion of a whole line of errors and then following the progress of the general term of the series expansion and also the behavior of each constituent harmonic. The errors (forming a continuous function of points in an abstract Banach space) are represented by vectors with the Fourier coefficients constituting a second Banach space. The amplification factor of the difference equation is shown to be always less than unity which guarantees the stability of the employed implicit recurrence scheme.

Experiments conducted on a vertical column packed uniformly with very fine sand, show a satisfactory agreement between the theoretically and experimentally obtained values. Many experimental results are shown in an attempt to explain the infiltration phenomenon with emphasis on the shape and movement of the wet front, and the effects of the degree of compaction, initial water content and deaired water on the infiltration rate.