Fuzzy neural network models for liquefaction prediction

Integrated fuzzy neural network models are developed for the assessment of liquefaction potential of a site. The models are trained with large databases of liquefaction case histories. A two-stage training algorithm is used to develop a fuzzy neural network model. In the preliminary training stage, the training case histories are used to determine initial network parameters. In the final training stage, the training case histories are processed one by one to develop membership functions for the network parameters. During the testing phase, input variables are described in linguistic terms such as ‘high’ and ‘low’. The prediction is made in terms of a liquefaction index representing the degree of liquefaction described in fuzzy terms such as ‘highly likely’, ‘likely’, or ‘unlikely’. The results from the model are compared with actual field observations and misclassified cases are identified. The models are found to have good predictive ability and are expected to be very useful for a preliminary evaluation of liquefaction potential of a site for which the input parameters are not well defined.     

Dynamic processes of a simple linear dune—a study in the Taklimakan Sand Sea, China

Samples of dune sands, surveys of the morphology and field measurements of wind velocity and direction of a simple linear dune in Taklimakan Sand Sea show that the airflow and sand flux vary with the change of wind direction on the dune surface. Decrease of the airflow stress on the lee flank does not result in much decrease of the sand flux because of the low threshold shear velocities and the airflow conditions. There are no significant relations between the sand flux on the lee flank and the angle of incidence of the airflow. The low threshold shear velocities and the maintenance of the sand flux at the lee flank are the main mechanisms keeping the linear shape of the dunes. Measurements of the sand flux shows that it reaches a maximum on the crest of the dune. The grain size of the transported sands has some differences compared to that of the dune surface. The sands transported are finer than that on the dune surface, but better sorted under the influence of the medium to low wind activity. The field experiment results exhibit that it is possible for the dunes to be shaped as linear dunes during the processes of accumulation and elongation.     

Fine structure of Pn velocity beneath Sichuan-Yunnan region

We use 23298 Pn arrival-time data from Chinese national and provincial earthquake bulletins to invert fine structure of Pn velocity and anisotropy at the top of the mantle beneath the Sichuan-Yunnan and its adjacent region. The results suggest that the Pn velocity in this region shows significant lateral variation; the Pn velocity varies from 7.7 to 8.3 km/s. The Pn-velocity variation correlates well with the tectonic activity and heat flow of the region. Low Pn velocity is observed in southwest Yunnan, Tengchong volcano area, and the Panxi tectonic area. These areas have very active seismicity and tectonic activity with high surface heat flow. On the other hand, high Pn velocity is observed in some stable regions, such as the central region of the Yangtze Platform; the most pronounced high velocity area is located in the Sichuan Basin, south of Chengdu. Pn anisotropy shows a complex pattern of regional deformation. The Pn fast direction shows a prominent clockwise rotation pattern from east of the Tibetan block to the Sichuan-Yunnan diamond block to southwest Yunnan, which may be related to southeastward escape of the Tibetan Plateau material due to the collision of the Indian Plate to the Eurasia Plate. Thus there appears to be strong correlation between the crustal deformation and the upper mantle structure in the region. The delay times of events and stations show that the crust thickness decreases from the Tibetan Plateau to eastern China, which is consistent with the results from deep seismic sounding.     

GPS RTK测量技术的应用与体会

GPS RTK测量技术以快捷、精准的特点广泛地被测量人员所接受，但是在到量中仍存在一些问题需要研究。本文介绍了几种经常遇到的棘手问题及其解决方法。     

按不同气候背景制作单站MOS预报方程

在基层业务体制的框架基本形成以后，县局在参考上级指导预报的同时，初步开展对单站数值预报模式的研究，注意中、小尺度天气的变化规律，以提高订正预报的精度。     

地震视应力用于震后趋势快速判定的可能性

对多次强震的地震视应力进行了计算，将各次强震的地震视应力的高低与强震发生后的后续地震的强度进行联系，探索它们之间的关系，进而提出将地震视应力用于强震震后趋势预测的可能性。对多次强震的地震视应力与其后续地震强度的关系的分析结果表明，低视应力的强震的后续地震的强度一般较低，即低视应力的强震发生后，震区发生较强的后续地震的可能性较小。     

加权最小范数平差的伪观测解法

首先证明了加权最小范数平差一般解的一个重要性质，由此提出了利用附加的观测值来求解加权最小范数平差问题的方法，并证明了该法与利用一般解法等价性的充要条件。     

关于地理信息系统学科的理论基础与体系框架

地理信息系统已成为一门独立的学科，关于它的理论基础与体系框架应该是一个值得研究的问题。从地理信息系统的本质看，它是传输地理环境信息的现代化工具，“空间认知”与“地理系统”作为地理信息系统的认识论和方法论，是地理信息系统的理论基础；地理信息系统理论、技术与应用构成了它的体系框架。     

开发PC─E500汉化软件

本文介绍了SHARPPCE500计算机汉化软件的编制，提供了一种较好的汉字库获取方法。     

平面图形演化方法在地震预报中的应用

“平面图形演化方法”是利用无量刚的异常信息量来反映震前异常空间分布及其随时间的演化，从而预报发震时间和地点的一种方法，这种方法μ值等值线图法，μ值异常平面图法组成，它是通过绘制不同的时间上述二种平面图，通过分析水位，水化异常于震前在时间上，空间上的演化规律，寻找μ值异常集中区或高值区，异常丛集图像分布区，条带图像分布带及条带交叉图像地区，并考虑异常分布图像与活动构造带和地震带的关系，从而对强震发生