VRANCEA99—the crustal structure beneath the southeastern Carpathians and the Moesian Platform from a seismic refraction profile in Romania

The VRANCEA99 seismic refraction experiment is part of an international and multidisciplinary project to study the intermediate depth earthquakes of the Eastern Carpathians in Romania. As part of the seismic experiment, a 300-km-long refraction profile was recorded between the cities of Bacau and Bucharest, traversing the Vrancea epicentral region in NNE–SSW direction.

The results deduced using forward and inverse ray trace modelling indicate a multi-layered crust. The sedimentary succession comprises two to four seismic layers of variable thickness and with velocities ranging from 2.0 to 5.8 km/s. The seismic basement coincides with a velocity step up to 5.9 km/s. Velocities in the upper crystalline crust are 5.9–6.2 km/s. An intra-crustal discontinuity at 18–31 km divides the crust into an upper and a lower layer. Velocities within the lower crust are 6.7–7.0 km/s. Strong wide-angle PmP reflections indicate the existence of a first-order Moho at a depth of 30 km near the southern end of the line and 41 km near the centre. Constraints on upper mantle seismic velocities (7.9 km/s) are provided by Pn arrival times from two shot points only. Within the upper mantle a low velocity zone is interpreted. Travel times of a PLP reflection define the bottom of this low velocity layer at a depth of 55 km. The velocity beneath this interface must be at least 8.5 km/s.

Geologic interpretation of the seismic data suggests that the Neogene tectonic convergence of the Eastern Carpathians resulted in thin-skinned shortening of the sedimentary cover and in thick-skinned shortening in the crystalline crust. On the autochthonous cover of the Moesian platform several blocks can be recognised which are characterised by different lithological compositions. This could indicate a pre-structuring of the platform at Mesozoic and/or Palaeozoic times with a probable active involvement of the Intramoesian and the Capidava–Ovidiu faults. Especially the Intramoesian fault is clearly recognisable on the refraction line. No clear indications of the important Trotus fault in the north of the profile could be found. In the central part of the seismic line a thinned lower crust and the low velocity zone in the uppermost mantle point to the possibility of crustal delamination and partial melting in the upper mantle.     

山东省干旱演变规律及其影响

利用近５００年旱涝资料分析了干旱对山东省农业，工业及环境的危害。利用功率谱和模糊判别函数研究了旱涝的周期性和阶段性，指出旱涝的演变规律可为气候预测提供依据。     

莱州湾地区500年来旱涝灾害指数系列的建立

作者根据《山东省历代自然灾害志》的记载，应用Foxpro提供的软件，建立了莱州湾地区500年来旱涝灾害的汉字库，并通过一系列转换，将其转为数字库；将旱涝灾情的文字记我转换为区域旱涝指数，并与现代降水资料相对接，建立了1470－1990年共521年的历年逐季旱涝灾害指数序列，为分析该地区旱涝灾害发生规律和预测提供了依据。     

铜陵市区的烟幕分析及其预测

作者应用气象资料分析烟幕的发生规律，并应用灰色理论对烟幕进行预测。     

海安暴雨洪涝灾害预报探讨

暴雨洪涝预报是目前气象和水利，水文学科正在致力攻克的预报难题。作者通过调查海安洪涝发生与暴雨的关系，利用水文，地形资料找出径流量和主河道水位升降的相关，结合暴雨预报，推导暴雨洪涝发生的可能以及大致区域和洪涝程度。     

随机地震动场激励下拱坝多点输入的抗震可靠度分析

本文首次将空间随机地震场输入用于拱坝随机振动分析之中，与拱坝三维动力学有限元法结合，完成了三维拱坝在空间随机地震动场作用下的随机振动分析方法及算法的研究，另外考虑动力和静力的叠加，提出了三维空间随机地震动场的拱坝抗震可靠度分析方法，最后编制了相应的程序，对小湾拱坝进行了抗震可靠度分析，并与随机变量的结果进行了比较。     

地震灾场的模拟与模拟控制问题研究（Ⅰ）地震灾场模拟

本文将离散事件系统仿真的方法引入到地震工程研究之中．在论述地震灾场模拟与离散事件系统仿真的共性与差异的基础上，以城市地震次生火灾的发生与扑救过程为背景，发展了场、时域复合仿真模型，提出了城市地震灾场的综合模拟算法．实例研究表明，本文建议的地震灾场模拟算法可以揭示复合型工程系统地震反应的一系列特征参数。     

改进的钢筋混凝土结构双参数地震破坏模型

本文在分析现有结构地震破坏模型的基础下，提出了一种改进掇参数地震破坏模型。文中采用三线退化型恢复力模型对实际钢筋混凝土结构在实际地震作用下的变形与累积耗能，极限变形与极限滞回耗能进行了分析，通过模型计算的破损结果与实际在结果的对比，研究了模型参数，从而建立了钢筋混凝土结构的双参数地震破坏模型。     

云龙台固体潮观测对地震监测能力的评价

对云龙台的水管倾斜仪观测数据进行了倾斜固体潮的潮汐分析，发现ＥＷ和ＮＳ两个分量的振幅因子异常变化与地震有较好的对应关系。由于潮汐分析方法本身的科学性和严密性，故振幅因子误差小、精度高；又因为对潮汐谐波波群的选择，避免了以太阳日为基础的周期性干扰，Ｍ２、Ｏ１波振幅因子的地震异常识别则简易可行，并且可信度高。云龙台的固体潮观测具有较强的地震监测能力。     

Modeling,ray tracing,and block nonlinear travel-time inversion in 3D

We describe an integrated forward and inverse three-dimensional modeling system that can deal with complex geological structures. The system has been designed to handle large-scale problems by using a distributed approach. It uses seismic ray tracing for forward simulation, time-to-depth mapping, and nonlinear travel-time inversion.A novel decomposition method is our tool to attack large-scale problems in a parallel approach. The system is fully implemented and we demonstrate its performance with synthetic examples.This research was partially supported by the National Science Foundation under SBIR Grant III-9300992.