气候模式模拟中海量数据的并行抽取与可视化分析诊断

在地球动力学和气候模拟等领域, 数值模拟产生的数据规模达到Tb至Pb量级。 实现这些海量数据的实时可视化和实时诊断分析面临很大的困难, 时空多尺度数据抽取可以解决这一瓶颈。 高精度数据的可视化结果展示需要高分辨率的显示设备, 并行的大屏幕显示技术是解决这一问题的有效手段。 地球科学数值模拟、 并行数据抽取和高分辨率显示都需要搭建高性能计算机集群。 本文在基于Lagrange插值的多维度、 多尺度、 多分辨率并行数据抽取算法的同时, 利用并行计算节点及LCD显示器, 基于Rocks cluster系统搭建起一个176核, 4×10×1024×1280分辨率的高性能计算模拟、 数据抽取和并行显示输出的集成平台, 并将该平台成功应用到气候模式模拟产生的海量数据的并行抽取和并行显示。     

一种去除CT图像中同心椭圆伪影的变分模型

在计算机断层成像（CT）中,伪影会降低重建图像的质量。针对该问题,本文提出一种去除CT图像中同心椭圆伪影的方法。该方法基于方向全变分（DTV）的思想,将椭圆伪影的去除问题建模为能量最小化问题,并通过对椭圆伪影边缘特征的分析,建立适应于椭圆伪影的变分模型。由于提出的模型为具有可分裂结构的非光滑凸优化问题,因此本文使用交替方向乘子法（ADMM）。最后,通过仿真实验验证该模型对于去除椭圆伪影的有效性。      

亚洲MAGSAT卫星磁异常图

使用２０多万个ＭＡＧＳＡＴ卫星磁测数据，编绘亚洲卫星标量（△Ｆ）和矢量（△Ｘ，△ｙ，△Ｚ）磁异常图（２０°ｓ－６０°Ｎ，５０°Ｅ－１６０°Ｅ）．为了提取岩石层卫星磁异常，首先对ＭＡＧＳＡＴ卫星数据进行了筛选，然后进行主磁场、磁层场、电离层场和感应场改正，并消除虚假测点．将卫星数据分成黎明组、黄昏组和联合组，分别绘制相应的卫星磁异常图．根据１°×１°的网点值，用计算机编绘亚洲卫星磁异常图．初步分析了卫星磁异常与大地构造的关系．     

利用数值计算方法（Lagrange）研究伽师及邻近地区重力场变化特征

利用数值计算的方法（Lagrange）,结合近4期流动重力资料,分析伽师及邻近地区的重力场时空动态变化特征,初步研究了该区的重力变化与地震活动及构造的关系。结果表明：重力场变化与构造环境变化有关。在地震孕育过程中,重力场出现有规律的变化并伴随重力变化密集区,地震往往多发于重力场变化正负值交替的零线附近。     

组建亚洲地震委员会（ASC）的情况通报

组建亚洲地震委员会（ＡＳＣ）的情况通报１．组建亚洲地震委员会的目的和意义亚洲地区是全球地震活动频度最高、强度最大的地区，也是地震灾害严重的地区。据不完全统计，本世纪以来至１９８０年，全球共发生大地震（Ｍ≥７．５）共３７次，其中５４％发生在亚洲。从地震...     

亚洲理工学院防灾减灾国际培训班概况

亚洲理工学院防灾减灾国际培训班概况亚洲理工学院防灾中心（ＡＤＰＣ－ＡＩＴ），是国际减灾十年（ＩＤＮＤＲ）项目在亚太地区的牵头单位。为了举办第三世界国家防灾减灾国际培训班，日本政府通过日本国际合作事业团体（ＪＩＣＡ）东京和曼谷办事处作了前期工作。１９９...     

亚洲地震委员会（ASC）正式宣告成立

亚洲地震委员会（ＡＳＣ）正式宣告成立１９９６年８月１～３日，在唐山召开的国际地震学与地球内部物理学协会（ＩＡＳＰＥＩ）亚洲区域大会期间，由ＩＡＳＰＥＩ秘书长Ｅｎｇｄａｈｌ博士主持召开了亚洲地震委员会代表会议。来自亚洲和南太平洋地区共１９个国家的代表出...     

亚洲中东部“塑性流动－地震”网络系统及板内构造单元

对于地震的网络状分布特征的研究表明，在亚洲的中东部地区存在着两个网络系统，即分布于大部分地区的中东亚网络系统和位于其东南的华东南网络系统。根据多层构造模型，这些地震网络系统实际上是岩石圈下层（含下地壳和岩石圈地幔）塑性流动网络的一种显示。每一“塑性流动－地震”网络系统为不同类型的边界所围限，其中包括一段驱动边界以及若干段约束边界和泄流边界。本地区的两个塑性流动网络系统分别以喜马拉雅弧和台湾弧为驱动边界，对板块内部的构造变形、构造应力场、地震活动性、以及构造单元（亚板块、地体等）的划分起着控制的作用。     

非均匀介质热蠕变流动的数值求解

针对非均匀介质中热蠕变流动问题,给出了有限单元方法与网格－粒子方法联合求解新技术,即有限单元方法求解欧拉网格节点上的未知量,分布于单元内部作为物质成分标记的粒子反映变形过程.有限元法求解动量方程和连续性方程时引入了速度场和压力场等阶插值的压力场稳定的Petrov Galerkin方法,求解能量方程时采用了流线迎风Petrov Galerkin方法,网格－粒子算法中采用双线性插值与有限单元插值函数对应.有限单元计算与网格－粒子计算相对独立,两种方法计算的数据通过有限单元节点传递.同时,实现了三角形单元的算法和程序,解决了复杂结构条件下不规则网格计算的问题.通过经典方腔热对流问题验证了程序,给出了不规则形态块体沉降算例,并分析了数值解的稳定性.     

亚洲中东部“塑性流动—地震”网络系统及板内构造单元

对于地震的网络状分布特征的研究表明，在亚洲的中东部地区存在着两个网络系统，即分布于大部分地区的中东亚网络系统和位于其东南的华东南网络系统。根据多层构造模型，这些地震网络系统实际上是岩石圈下层（含下地壳和岩石圈地幔）塑性流动网络的一种显示，每一“塑性流动－地震”网络系统为不同类型的边界所围限，其中包括一段驱动边界以及若干段约束边界和泄流边界。本地区的两个塑性流动网络系统分别以喜马拉雅弧和台湾弧为驱动     

Mantle flow beneath the Asian continent and its force to the crust

Introduction In order to understand the formation of the geological structure and its evolution of the Asian continent, it is needed to investigate the driving forces from the around continents and oceans and from its underlying mantle as well. The super parallel computer has been used to simulate the geodynamical processes in the lithosphere and mantle for a long time in many international works (Bird, 1995; Tackley, 1995). We also calculated the patterns of the global mantle convection on o…     

An improved optimal elemental method for updating finite element models

The optimal matrix method and optimal elemental method used to update finite element models may not provide accurate results. This situation occurs when the test modal model is incomplete, as is often the case in practice. An improved optimal elemental method is presented that defines a new objective function, and as a byproduct, circumvents the need for mass normalized modal shapes, which are also not readily available in practice. To solve the group of nonlinear equations created by the improved optimal method, the Lagrange multiplier method and Matlab function fmincon are employed. To deal with actual complex structures, the float-encoding genetic algorithm (FGA) is introduced to enhance the capability of the improved method. Two examples, a 7degree of freedom (DOF) mass-spring system and a 53-DOF planar frame, respectively, are updated using the improved method.The example results demonstrate the advantages of the improved method over existing optimal methods, and show that the genetic algorithm is an effective way to update the models used for actual complex structures.     

基于GPU并行的时间域全波形优化共轭梯度法快速GPR双参数反演

探地雷达（GPR）时间域全波形反演计算量巨大,内存要求高,在微机上计算难度大.本文中作者基于GPU并行加速的维度提升反演策略,采用优化的共轭梯度法,避免了Hessian矩阵的计算,在普通微机上实现了时间域全波形二维GPR双参数（介电常数和电导率）快速反演.论文首先推导了二维TM波的时域有限差分法（FDTD）的交错网格离散差分格式及波场更新策略.然后,基于Lagrange乘数法,将约束问题转化为无约束最小问题,构建了共轭梯度法反演目标函数,采用Fletcher-Reeves公式与非精确线搜索Wolfe准则,确保了梯度方向修正因子及迭代步长选取的合理性.而GPU并行计算及维度提升反演策略的应用,数倍地提升了反演速度.最后,开展了3个模型的合成数据的反演实验,分别从观测方式、梯度优化及天线频率等方面,分析了这些因素对雷达全波形反演的影响,说明双参数的反演较单一的介电常数反演,能提供更丰富的信息约束,有效提高模型重建的精度.     

伴随方法在海洋数值模式中的应用

本文通过对变分资料同化方法在气象学和海洋学方面特别是在海洋数值模式中应用的简单回顾，评价了伴随方法在方法论上的优越性，首先引入包含海洋数值模式控制方程组的拉格朗日乘子，说明变分资料同化的数学原理；介绍伴随方法应用中目标函数的构造方法，简单说明了伴随方法中伴随方程的导出和伴随方法实际应用的步骤，最后就近几年国外研究进展做了简单概括。     

The mantle convection pattern and force source mechanism of recent tectonic movement in China

The Runcorn stress equations and 2–30° harmonic coefficients of the geopotential have been applied to determine the mantle convection pattern beneath China. The pattern is compared with geophysical and geological observations and it is found that the directional change belts of mantle flows coincide with the major fault belts between tectonic units of China. The stress field generated by mantle flows, except in the Tian Shan region, also coincide with the stress field of recent tectonic movement in China. The Tarim and Junggar basins are formed by tensional stresses due to divergent mantle convection currents under northwest China. The formation of the Qinhai-Xizang (Tibet) plateau is due mainly to the compression of the Tarim block and Indian plate, caused by convergent mantle convection currents. The shear-fault belts in central China (100–105°E) are generated by the running change belt of mantle flows, a well-known N-S seismic zone. In eastern China, tensional faults, grabens, lake and sea depressions are related to the eastward displacement of continental lithosphere exerted by eastward dispersal mantle flows under this region.This paper provides new material for further study of the force source mechanism of recent tectonic movement from the viewpoint of mantle convection currents.     

上地幔各向异性的反演方法

详细推导了弱各向异性介质的地震波速，在此基础上介绍了用Pn震相研究上地幔各向异性的几种具体算法；推导了用SKS震相和ScS震相反学下地幔各向异性的方法、优缺点及几种方法的相互关系。     

Seismic tomography and continental drift

Based on data of seismic tomography, the structure of the mantle flows of the contemporary Earth and the continental drift are calculated. Results of calculation of the contemporary motion of continents and their future drift for 150 Myr are presented. The present-day positions of six continents and the nine largest islands are taken as an initial state. The contemporary temperature distribution in the mantle is calculated according to the data of seismic tomography. The 3-D distribution of seismic wave velocities is converted into the density distribution and then into the temperature distribution. The Stokes equation is numerically solved for flows in a viscous mantle with floating continents for the given initial temperature distribution. In this way, the velocities of convective flows are determined in the entire present-day mantle and the surface distribution for the Earth’s heat flux is obtained. The reliability of the calculated flows in the mantle is estimated by the comparison of the calculated velocities of the contemporary continents and oceanic lithosphere with data of satellite measurements. Further, evolutionary equations of convection with floating continents were numerically solved. The calculated structure of mantle flows, temperature distribution, and position of continents are presented for a time moment 150 Myr in the future. The resulting successive changes in the position of continents in time show how islands (in particular, Japan and Indonesia) will be attached to continents and how continents will converge, exhibiting a tendency toward the formation of a new supercontinent in the southern hemisphere of the Earth.     

Mantle plumes in the models of quasi-turbulent thermal convection

The process of multiple self-nucleation and ascent of mantle plumes is studied in the numerical models of thermal convection. The plumes are observed even in the simplest isoviscous models of thermal convection that leave aside the more complex rheology of the material, thermochemical effects, phase transformations, etc., which, although controlling the features of plumes, are not necessary for their formation. The origin of plumes is mainly due to the instability of the mantle flows at highly intense (low-viscous) thermal convection. At high viscosity, convective flows form regular cells. As viscosity decreases, the ascending and descending flows become narrower and unsteady. At a further decrease in viscosity, the ascending plumes assume a mushroom-like shape and occasionally change their position in the mantle. The lifetime of each flow can attain 100 Ma. Using markers allows visualizing the evolution of the shape of the mantle plumes.     

Pitfalls in Nonlinear Inversion

— We discuss and illustrate graphically with simple 2-D problems, four common pitfalls in geophysical nonlinear inversion. The first one establishes that the Lagrange multiplier, used to incorporate a priori information in the geophysical inverse problem, should be the largest value still compatible with a reasonable data fitting. This procedure should be used only when the interpreter is sure about the importance of the a priori information used to stabilize the inverse problem relative to the geophysical observations. Because this is rarely the case, the user should use the smallest Lagrange multiplier still producing stable solutions. The second pitfall is an attempt to automatically estimate the Lagrange multiplier by decreasing it along the iterative process used to solve the nonlinear optimization problem. Consequently, at the last iteration, the Lagrange multiplier may be so small that the problem may become ill-posed and any computed solution in this case is meaningless. The third pitfall is related to the incorporation of a priori information by a technique known as “Jumping.” This formulation, from the viewpoint of the class of Acceptable Gradient Methods, is incomplete and may lead to a premature halt in the iteration, and, consequently, to solutions far from the true one. Finally, the fourth pitfall is an inadequate convergence criterion which stops the iteration when the data misfit drops just below the noise level, irrespective of the fact that the functional to be minimized may not have attained its minimum. This means that the a priori information has not been completely incorporated, so that this stopping criterion partially neutralizes the effect of the stabilizing functional, and opens the possibility of obtaining unstable, meaningless estimates.