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.     

Screening model for volatile pollutants in dual porosity soils

This paper develops mass fraction models for transport and fate of agricultural pollutants in structured two-region soils. Mass fraction index models, based on a semi-infinite domain solution, are derived that describe leaching at depth, vapor losses through soil surface, absorption, and degradation in the dynamic- and stagnant-water soil regions. The models predict that leaching is the result of the combined effect of the upward vapor-phase transport relative to downward advection, residence time relative to half-life, dispersion, and lateral diffusive mass transfer. Simulations show that leached fraction of volatile compounds does not always decrease monotonically with increased residence time relative to the pollutant half-life, as a result of complex interactions among the different physical and biochemical processes. The results show that leaching, volatilization, and degradation losses can be affected significantly by lateral diffusive mass transfer into immobile-water regions and advection relative to dispersion (i.e. Peclet number) in the mobile-water regions. It is shown that solute diffusion into the immobile phase and subsequent biochemical decay reduces leaching and vapor losses through soil surface. Potential use of the modified leaching index for the screening of selected pesticides is illustrated for different soil textures and infiltration rates. The analysis may be useful to the management of pesticides and the design of landfills.     

Predicting average annual groundwater levels from climatic variables: an empirical model

On the basis of one-dimensional theoretical water flow model, we demonstrate that the groundwater level variation follows a pattern similar to recharge fluctuation, with a time delay that depends on the characteristics of aquifer, recharge pattern as well as the distance between the recharge and observation locations. On the basis of a water budget model and the groundwater flow model, we propose an empirical model that links climatic variables to groundwater level. The empirical model is tested using a partial data set from historical records of water levels from more than 80 wells in a monitoring network for the carbonate rock aquifer, southern Manitoba, Canada. The testing results show that the predicted groundwater levels are very close to the observed ones in most cases. The overall average correlation coefficient between the predicted and observed water levels is 0.92. This proposed empirical statistical model could be used to predict variations in groundwater level in response to different climate scenarios in a climate change impact assessment.     

Models of passive margin inversion: implications for the Rhenohercynian fold-and-thrust belt, Belgium and Germany

Positive tectonic inversion is related to the transmission of compressional stresses along a décollement into the foreland of an orogenic zone. This stress and strain concentration in regions remote from the main orogenic front is commonly related to the presence of pre-existing rheological heterogeneities such as normal syn-depositional faults. During inversion, these pre-existing normal faults are reactivated as reverse faults. Tectonic inversion in the Rhenohercynian fold-and-thrust belt during the Variscan Orogeny shows that inversion is likely synchronous with the onset of collision in the hinterland. Here, we present the results of a simplified thermo-mechanical model (STM) which allows one to study strain partitioning between two orogenic zones. We show that, if the two orogenic zones have the same mechanical properties, the viscosity of the décollement, which links them, controls the initial strain partitioning. During subsequent finite shortening, erosional processes determine the partitioning of strain rate. The presence of a weak structure in the inverted zone and of a low-viscosity décollement leads to initial strain concentration in the inverted track rather than in the collision zone and a progressive decrease in strain partitioning between the two orogenic zones. The STM results are in good agreement with results of a 2D finite-element model. We conclude that, in the western part of the Rhenohercynian Massif, simultaneous uplift and deformation within the Mid-German Crystalline Rise (the main collision zone) and the Ardenne Anticlinorium (the inverted zone) lead to interpreting this orogenic event as a case of vice tectonic rather than the propagation of a ‘wave of folding’ towards the Variscan front, as suggested by previous authors.     

GIS专题数据综合的研究

随着GIS技术在各种应用领域的推广使用，各领域的专业数据与G1S的空间数据相结合，形成描述特定地理现象的各种专题数据。与GIS空间数据一样，专题数据的应用也由微观的、较为详细地描述小区域地理现象特征，转向宏观的、较为概略地描述较大范围地理现象特征，这是专题数据综合的主要任务。该文在讨论GIS专题数据类型及特征的基础上，着重分析了专题数据综合的基本类型，并提出了专题数据综合的概念模型，为GIS专题数据自动综合的实施提供理论依据和实施方案。     

Geomechanical model for the post-Variscan evolution of the Permocarboniferous–Mesozoic basins in Northeast Germany

The Permocarboniferous basins in Northeast Germany formed on the heterogeneous and eroded parts of the Variscan orogene and its deformed northern foreland. Transtensional tectonic movements and thermal re-equilibration lead to medium-scale crustal fragmentation, fast subsidence rates and regional emplacement of large amounts of mostly acidic volcanics. The later basin formation and differentiation was triggered by reversals of the large-scale stress field and reactivation of prominent zones of weakness like the Elbe Fault System and the Rhenohercynian/Saxothuringian boundary that separate different Variscan basement domains in the area. The geomechanical behaviour of the latter plays an important role for the geodynamic evolution of the medium to large-scale structural units, which we can observe today in three dimensions on structural maps, geophysical recordings and digital models. This study concentrates on an area that comprises the southern Northeast German Basin, the Saale Basin, the Flechtingen High, the Harz Mountains High and the Subhercynian Basin. The presented data include re-evaluations of special geological and structural maps, the most recent interpretation of the DEKORP BASIN 9601 seismic profile and observations of exposed rock sections in Northeast Germany. On the basis of different structural inventories and different basement properties, we distinguish two structural units to the south and one structural unit to the north of the Elbe Fault System. For each unit, we propose a geomechanical model of basin formation and basin inversion, and show that the Rhenohercynian Fold and Thrust Belt domain is deformed in a thin-skinned manner, while the Mid-German Crystalline Rise Domain, which is the western part of the Saxothuringian Zone, rather shows a thick-skinned deformation pattern. The geomechanical model for the unit north to the Elbe Fault System takes account to the fact that the base of the Zechstein beneath the present Northeast German basin shows hardly any evidence for brittle deformation, which indicates a relative stable basement. Our geomechanical model suggests that the Permocarboniferous deposits may have contributed to the structural stiffness by covering small to medium scale structures of the upper parts of the brittle basement. It is further suggested that the pre-Zechstein successions underneath the present Northeast German basin were possibly strengthening during the Cretaceous basin inversion, which resulted in stress transfer to the long-lived master faults, as indicated for example by the shape of the salt domes in the vicinity of the latter faults. Contrary to this, post-Zechstein successions deformed in a different and rather complex way that was strongly biased by intensive salt tectonic movements.     

用几何光学模型对方向变化和直方图建模(英文)

森林和林地的图象二向性反射函数是一个统计函数，较之小尺度的树冠，它更多地用于大尺度的均匀覆盖的地块。用航空象片和高空间分辨率扫描仪数据作的图象热点影响研究在小尺度下显示出很大方差，而且，太阳和观测角度的交互变化进一步增加了这种反射各向异性变化的方差并有规律地继续呈现在分辨率低的图象中，这被称作ＢＲＶＦ或二向性反射方差函数。近年来，作为一种解释结构的手段，高分辨率图象的方向性方差和直方图结构越来越受到重视，这方面的数据也越来越多。这项工作是利用图象方差来解释结构问题（Ｓｔｒａｈｌｅｒ和李小文倡导）的一种延伸，并在过去１５年中由众多人员作了大量工作。在树冠尺度下，森林的直方图和二向反射方差函数可以计算出来，这里利用了近似迭代函数来处理这些数据并和数值积分模拟进行了对比，结果显示可对二向性反射方差函数的测量和引入直方图的各向异性进行准确建模。     

几何光学冠层反射建模中的光谱分量特征的参数化(英文)

该文用几何光学与辐射传输混合模型研究不连续植被冠层的几何光学反射模型的四分量（承照树冠、承照地面、阴影树冠、阴影地面）的参数化。用一个修正的均匀介质层路径散射（反射与传输）参数的解析算法估计路径散射参数（反射与传输），其中也考虑了冠层间隙的影响。光谱分量特征是不连续植被冠层的传输与反射，背景反照率，以直射光通量与天空漫射光通量比例的函数。光谱分量特征的模型与在美国缅因州Ｈｏｌａｎｄ采集的针叶林数据吻合。基于ＬｉＳｔｒａｈｌｅｒ几何光学相互遮蔽模型，用参数化的光谱分量特征对老松林和老云杉林的方向反射进行估计，其结果与在不同太阳与观测方向上的ＰＡＲＡＢＯＬＡ测量值匹配得很好。     

A model for the Am (Km) planetary geomagnetic activity index and application to prediction

We studied the existence of dynamical stochastic relations in the evolution of the am index. A first analysis of the autocorrelation functions showed evidence of several seasonalities. We first used linear (ARMA) models, and it was found that these do not account for the whole internal dynamics of the data series. We then used various non-linear models to provide a better fit to reality. The forecast performances of the non-linear models are not significantly different from those of the linear model. We give a tentative explanation for the failure of the non-linear predictions. Finally, ARCH models were used in order to take into account the fact that the confidence interval for the predicted value depends on past observations.     

A physical statistical approach to erosion

Erosion is a complex process consisting of many components such as surface runoff, impact of raindrops, wind forces, soil and rock mechanics, etc. Trying to integrate all these processes into a physical model seems to be hopeless. In order to understand the variety of natural shapes and patterns produced by erosion we present an integrated statistical approach. Our model is based on simple physical constraints for the separation of amalgamated particles (abrasion) and for the movement of loose particles (denudation) and on the laws of statistics. After some simplifications, we obtain a nonlinear system of partial differential equations which is solved using finite volume techniques. The model is suitable for the formation of different types of rill systems and the episodic behaviour of erosion processes, a kind of self-organized criticality. Besides effects of inhomogeneities, e.g. the formation of terraces can be investigated.