一种利用贝叶斯最小判别准则估计未来地震危险的新方法

提出了一种基于最大似然和贝叶斯最小判别准则的新预测方法,它的长处在于对样本数无特殊要求,所以,对低地震活动地区,此模型也能得到与时间有关的稳定的未来强震危险估计。最后,以华北地区为例,说明了本方法的可行性。     

Shallow landslides in deeply weathered slates of the Sierra de Collcerola (Catalonian Coastal Range, Spain)

Stability conditions in an area located NW of Barcelona (Spain) are discussed. Here, several mass movements were observed, mainly affecting weathered Paleozoic slates. Many of these failures involved slopes cut along recent infrastructures: debris flows, wedge and plane failures, generally surficial, occurred more frequently. After a detailed geological and geomorphologic survey, geomechanic characterization was carried out, according to RMR and SMR classifications. This rating gave a prediction of slope behaviour, in fairly good agreement with the real observed one.

Stability numerical analysis was carried out for the main cut slopes, based upon the Limit Equilibrium Method. First of all, the deterministic factor of safety was computed using the mean values of parameters. After that, a simulation technique based upon the Monte Carlo Method was applied in order to obtain factor of safety distributions. The probability of failure was estimated as P(F<1).

Finally, results from deterministic and probabilistic approaches were compared. The effectiveness of different possible remedial measures was highlighted by means of a sensitivity analysis, which showed that the more important parameters in the study area are the geometrical ones (height, slope and failure plane angles). The final technical solutions adopted are briefly outlined.     

逆斜率法及其讨论

经对上海地区51根试桩资料分析表明,同实测值比较,计算值误差一般不超过30％,误差与未次加载的ΔSn/ΔPn比值似有一定的相关性,当比值大于2,误差一般不超过20％,同时对没有达到终至加载条件的试桩也可采用此法估算。     

位移时空综合分析法及其在边坡工程中的应用

以往对监测位移数据分析方法主要是位移-时间关系分析。但监测数据包含了大量对工程设计和施工都很有用的信息,其中不乏位移-空间方面的信息;另一方面,从工程设计和施工角度看,除需要位移-时间信息外,还需要位移-空间信息。位移时空综合分析法的提出旨在综合位移-时间和位移-空间等两方面的信息,为工程设计和施工服务。该方法依表现形式不同又分为竖线法和曲面法两种。本文将通过五强溪船闸边坡工程实例,来进一步说明位移时空综合分析法的原理及其应用。     

The generation and redistribution of overland flow on a massive oxic soil in a eucalypt woodland within the semi-arid tropics of North Australia

There is a dearth of knowledge on the runoff processes of eucalypt woodland communities in the semi-arid tropics of Australia. The work was undertaken on a 100 m transect of a 0·8 degree hillslope typical of the ‘smooth plainlands’ of central-north Queensland. This paper introduces a new experimental design for measuring overland flow in such areas by way of a cascade system of unbounded runoff plots which allow the inputs and outputs between troughs to be calculated. Most storms generate overland flow. Time to overland flow ranges between 1 and 18 min where rain intensities are above 10mm hr⁻¹ and when the average detention storage of 3·6 mm is exceeded. The bare soil surfaces within the scattered grass understory control the runoff generation process through the temporal variability of field saturated hydraulic conductivity. The study demonstrated that overland flow is mainly redistributed over the freely-draining oxic soil. Some areas export more overland flow than they gain from upslope (runoff), others gain more overland flow than they export (runon). Over the study period only 2 per cent of total rain is transferred out of this 100 m transect as overland flow due to the short duration of storms, the relatively high soil permeability, and the low slope angle. The remainder adds to the large soil water store or deep drainage. The variability of runoff–runon over these ‘smooth plainlands’ highlights how results from bounded plots would be misleading in such areas.     

Three-dimensional simulations of a starburst wind

In order to better understand the formation of a starburst-driven wind, we have performed a series of three-dimensional hydrodynamical simulations in an inhomogeneous interstellar medium. We present the results of these simulations, which provide new insights into the formation of the optical filaments and the origin of the soft X-ray emission.     

COMMENTS ON THE NIPALS ALGORITHM

The Non-linear lterative Partial Least Squares(NIPALS)algorithm is used in principal componentanalysis to decompose a data matrix into score vectors and eigenvectors(loading vectors)plus a residualmatrix.N1PALS starts with some guessed starting vector.The principal components obtained by NIPALSdepends on the starting vector;the first principal component could not always be computed.Wold hassuggested a starting vector for NIPALS,but we have found that even if this starting vector is used,thefirst principal component cannot be obtained in all cases.The reason why such a situation occurs isexplained by the power method.A simple modification of the original NIPALS procedure to avoid gettingsmaller eigenvalues is presented.     

Seismic amplifications from offshore to shore

The objective of this study is to determine numerical estimations of seismic amplifications of waves traveling from offshore to shore considering the effect of sea floor configurations. According to the Boundary Element Method, boundary elements were used to irradiate waves and density force can be determined for each element. From this hypothesis, Huygens’ Principle is implemented since diffracted waves are constructed at the boundary from which they are radiated and this is equivalent to Somigliana's theorem. Application of boundary conditions leads to determine a system of integral equations of Fredholm type of second kind, which is solved by the Gaussian method. Various numerical models were analyzed, a first one was used to validate the proposed formulation and some other models were used to show various ideal sea floor configurations to estimate seismic amplifications. Once the formulation was validated, basic slope configurations were studied for estimating spectra of seismic amplifications for various sea floor materials. In general terms, compressional waves (P-waves) can produce seismic amplifications of the incident wave in the order of 2–5. On the other hand, distortional waves (S-waves) can produce amplifications up to 5.5 times the incident wave. A relevant finding is that the highest seismic amplifications due to an offshore earthquake are always located near the shore-line and not offshore despite the seafloor configuration.     

基于主成分分析和决策级融合的高光谱图像分类方法研究

为了评价主成分分析方法和决策级数据融合方法对高光谱图像地物分类结果的影响,作者利用各分类器和主成分分量的优势,提出了两组地物分类结果的决策级融合策略,利用8种常用监督分类方法对高光谱原始图像和PCA变换后不同分量组合图分类得到的结果进行决策级融合,并应用覆盖黄河入海口新老河道交界处的CHRIS/PROBA高光谱图像开展实验研究。结果表明:直接采用每类地物分类精度最高的、空穴和缝隙采用总体分类精度最高的融合策略,在综合考虑生产者精度和用户精度的情形下,仅使用最大似然法、支持向量机和人工神经网络3种分类方法,按照分类精度从高到低的顺序进行的融合分类效果最好,总体分类精度为87.82%。与8种监督分类方法中效果最好的最大似然法相比,精度提高了2.7个百分点,同时明显减少了错分现象,尤其是对于分布面积较小的翅碱蓬和柽柳,滩涂被误分为翅碱蓬、芦苇被误分为柽柳的现象大大降低。     

华东地区月平均气温统计降尺度方法比较

用中国地面气象观测站的逐日气温观测资料和NCEP/NCAR再分析资料,分别使用基于多元线性回归(MLR)和3种主成分分析(PCA)的统计降尺度方法,对1959-2008年的华东地区的月平均气温分两个时段进行统计降尺度分析并加以检验,比较了不同降尺度方法的结果。结果表明:对于华东地区气温的统计降尺度预报,基于MLR的统计降尺度方法相对于3种PCA方法而言,对单站年际变化模拟方面有一定优势。PCA方法应用于统计降尺度时,预报因子的区域选择是影响统计降尺度结果的重要因素之一。对于温度进行统计降尺度分析时,预报因子中包含温度因子是非常必要的;所试验的4种降尺度方法,对各站点多年平均情况的模拟要好于对区域平均的年际变化的模拟。