Uncertainty assessment and optimization of hydrological model with the Shuffled Complex Evolution Metropolis algorithm: an application to artificial neural network rainfall-runoff model

Assessment of parameter and predictive uncertainty of hydrologic models is an essential part in the field of hydrology. However, during the past decades, research related to hydrologic model uncertainty is mostly done with conceptual models. As is accepted that uncertainty in model predictions arises from measurement errors associated with the system input and output, from model structural errors and from problems with parameter estimation. Unfortunately, non-conceptual models, such as black-box models, also suffer from these problems. In this paper, we take the artificial neural network (ANN) rainfall-runoff model as an example, and the Shuffled Complex Evolution Metropolis algorithm (SCEM-UA) is employed to analysis the parameter and predictive uncertainty of this model. Furthermore, based on the results of uncertainty assessment, we finally arrive at a simpler incomplete-connection artificial neural network (ICANN) model as well as with better performance compared to original ANN rainfall-runoff model. These results not only indicate that SCEM-UA can be a useful tool for uncertainty analysis of ANN model, but also prove that uncertainty does exist in ANN rainfall-runoff model. Additionally, in some way, it presents that the ICANN model is with smaller uncertainty than the original ANN model.     

GPS接收机仪器偏差的短期时变特征提取与建模

卫星和接收机仪器偏差(Differential Code Biases, DCB)是利用GPS(Global Positioning System)研究电离层的两类主要误差源.由于所处的空间环境恒定,且可被全球跟踪站连续观测,GPS卫星的DCB具备长期稳定性和较高的估计精度.但针对各类型接收机而言,受测站环境、硬件设施等影响,其DCB可能会呈现明显的短期变化.精确地模型化接收机DCB的短期变化特征,将有助于提高GPS电离层产品的可靠性,以及基于这些产品反演空间和地球科学现象的准确性.采用零/短基线GPS数据,本文改进了提取和分析接收机DCB变化的现有方案.随后,本文推导了一种能直接估计接收机DCB的函数模型.当检验出接收机DCB的短期变化服从随机游走时,通过对比接收机DCB的直接估值与间接提取值之间的符合程度,可"试探出"过程噪声标准差的最优经验值.实验分析选用4台双频接收机(共形成1条零基线和2条短基线,间距最大为15 m)多天的观测数据,主要结论包括:1)改进的接收机DCB提取方案能较好地克服低频伪距噪声和多路径效应的影响; 2)针对零基线,其接收机DCB在各天内的变化量级小于1 TECu,变化趋势则可采用过程噪声标准差为1.0~1.5 mm的随机游走加以描述; 3)对应于某短基线的接收机DCB在某天内的变化可达12 TECu,当采用随机游走描述其趋势时,过程噪声标准差的经验值超过2 mm.     

基于谐波建模和自相关的磁共振信号消噪与提取方法研究

磁共振探测技术（Magnetic Resonance Sounding,MRS）以其无损、定量、直接等优势,被广泛应用于地下水调查、水文环境评价以及灾害水源预警等领域.在实际应用中,强工频谐波和随机噪声等严重影响MRS信号的质量,导致后续水文参数解释不准确.针对这一问题,提出谐波建模和自相关相结合的方法进行消噪以及信号特征参数提取.首先构建工频谐波模型,针对建模算法严重依赖工频基频精度的问题,采用自适应扫描方式搜索方案,大幅提高搜索准确度和速度;其次推导了MRS信号自相关表达式,提出了自相关参数提取的非线性拟合方法.仿真数据结果表明,建模消噪方法有效消除了工频谐波,信噪比平均提升了17.03 dB;自相关处理后,信噪比进一步提升了16.10 dB,初始振幅和弛豫时间参数提取结果的准确度比处理前分别提高了3.8倍和2.8倍.通过不同信噪比和弛豫时间的重复实验,得到当噪声水平小于200 nV和弛豫时间大于200 ms时,自相关参数提取具有较高的稳定性.最后,通过野外实测数据处理实验,进一步验证了联合消噪和参数提取方法的有效性.

     

Numerical modeling of bedload and suspended load contributions to morphological evolution of the Seine Estuary (France)

This numerical modeling study (i) assesses the influence of the sediment erosion process on the sediment dynamics and subsequent morphological changes of a mixed-sediment environment, the macrotidal Seine estuary, when non-cohesive particles are dominant within bed mixtures (non-cohesive regime), and (ii) investigates respective contributions of bedload and suspended load in these dynamics. A three dimensional (3D) process-based morphodynamic model was set up and run under realistic forcings (including tide, waves, wind, and river discharge) during a 1-year period. Applying erosion homogeneously to bed sediment in the non-cohesive regime, i.e., average erosion parameters in the erosion law (especially the erodibility parameter, E₀), leads to higher resuspension of fine sediment due to the presence of coarser fractions within mixtures, compared to the case of an independent treatment of erosion for each sediment class. This results in more pronounced horizontal sediment flux (two-fold increase for sand, +30% for mud) and erosion/deposition patterns (up to a two-fold increase in erosion over shoals, generally associated with some coarsening of bed sediment). Compared to observed bathymetric changes, more relevant erosion/deposition patterns are derived from the model when independent resuspension fluxes are considered in the non-cohesive regime. These results suggest that this kind of approach may be more relevant when local grain-size distributions become heterogeneous and multimodal for non-cohesive particles. Bedload transport appears to be a non-dominant but significant contributor to the sediment dynamics of the Seine Estuary mouth. The residual bedload flux represents, on average, between 17 and 38% of the suspended sand flux, its contribution generally increasing when bed sediment becomes coarser (can become dominant at specific locations). The average orientation of residual fluxes and erosion/deposition patterns caused by bedload generally follow those resulting from suspended sediment dynamics. Sediment mass budgets cumulated over the simulated year reveal a relative contribution of bedload to total mass budgets around 25% over large erosion areas of shoals, which can even become higher in sedimentation zones. However, bedload-induced dynamics can locally differ from the dynamics related to suspended load, resulting in specific residual transport, erosion/deposition patterns, and changes in seabed nature.     

基于改进差分进化算法的大地电磁测深和重力宽范围物性约束联合反演

宽范围物性约束技术容易实现、具有一定容错性,目前已在大地电磁测深（MT）和地震、MT和重力联合反演中实现,但该技术是结合模拟退火算法实现的.差分进化算法（DE）是一种全局优化算法,但该算法在地球物理联合反演领域应用较少.基于此,本文以双种群设置方案为框架改进了DE算法,并提出了基于改进DE算法的宽范围物性约束技术.MT和重力联合反演的模型试验表明：与传统的DE算法相比,改进的DE算法收敛速度更快,寻优能力更强;基于改进DE算法的宽范围物性约束技术可以促进不同岩石物性参数在一定"范围"内实现耦合,既可以利用岩石物性关联的导向作用,又可以发挥优化算法的寻优能力,进而降低地球物理联合反演对先验信息的要求;此外,该技术的实现也验证了宽范围物性约束思想在联合反演领域中的适用性,具有进一步推广至其他优化算法中的潜质.     

基于改进差分进化算法的大地电磁测深和重力宽范围物性约束联合反演

宽范围物性约束技术容易实现、具有一定容错性,目前已在大地电磁测深（MT）和地震、MT和重力联合反演中实现,但该技术是结合模拟退火算法实现的.差分进化算法（DE）是一种全局优化算法,但该算法在地球物理联合反演领域应用较少.基于此,本文以双种群设置方案为框架改进了DE算法,并提出了基于改进DE算法的宽范围物性约束技术.MT和重力联合反演的模型试验表明：与传统的DE算法相比,改进的DE算法收敛速度更快,寻优能力更强;基于改进DE算法的宽范围物性约束技术可以促进不同岩石物性参数在一定"范围"内实现耦合,既可以利用岩石物性关联的导向作用,又可以发挥优化算法的寻优能力,进而降低地球物理联合反演对先验信息的要求;此外,该技术的实现也验证了宽范围物性约束思想在联合反演领域中的适用性,具有进一步推广至其他优化算法中的潜质.     

Uncertainty and importance assessment using differential analysis: an illustration of corrosion depth of spent nuclear fuel canister

Uncertainty of NRC proposed performance assessment for the corrosion depth of storage canister of nuclear spent fuel in time history is preformed using the differential analysis. The method is demonstrated to evaluate uncertainty propagation of a system with incomplete and insufficient data in the preliminary phase of assessment. The result shows that mean corrosion depth presents a non-linear regression while variance demonstrates a first-order linear increase through 1,000-year time history. By incorporating correlation into pitting factor, uniform corrosion factor, oxygen concentration and chlorine concentration, it is found that the covariance of pitting factor and chlorine concentration has the most significant contribution in the variance of total loss. Incorporating assumed covariance into the system, the analytical form of uncertainty and importance for the corrosion depth evaluation can be well demonstrated from the input parameters.     

Thermal evolution of the Ryoke metamorphic belt, southwestern Japan: Tectonic and numerical modeling

Abstract The Ryoke metamorphic belt of southwestern Japan is composed of Cretaceous Ryoke granitoids and associated metamorphic rocks of low-pressure facies series. The Ryoke granitoids are divided into sheet-like bodies (e.g. Gamano granodiorite) and stock-like bodies. The Gamano granodiorite intruded concordantly into the high-grade metamorphic rocks without development of a contact metamorphic aureole, and the intrusion ages of the granodiorite are similar to the ages of thermal peak of the low pressure (low-P) metamorphism. It is suggested that the low-P Ryoke metamorphism resulted from the intrusion of the Gamano granodiorite. In this study, a simple 1-D numerical model of conductive heat transfer was used to evaluate the thermal effects of emplacement of the Gamano granodiorite. Calculated temperature-time ( T-t ) paths are characterized by a rapid increase of metamorphic temperature and a relatively short-lived period of high temperature. For example, the T-t path at the 15-km depth is characterized by a rapid average increase in temperature of 1.4 × 10-³°C/year and high temperatures for < ca 0.5 Ma. The calculated peak temperature for each depth is nearly equal to the petrologically estimated value for each correlated metamorphic zone. The results suggest that the magma-intrusion model is one possible thermal model for low-pressure facies series metamorphism.     

Environmental impact assessment of a new nuclear power plant (NPP) based on atmospheric dispersion modeling

The US Environmental Protection Agency’s short-range atmospheric dispersion model (AERMOD 12345) is a good candidate for radiological dose calculations to the general public and the environment. It advanced capability should provide better confidence in the accuracy of offsite dose and risk assessment. The code has been used to compute the scaling factors for air concentration and ground level deposition of fission products based on routine and hypothetical accident releases from the NPP site in Geregu, Nigeria (7°33′N, 6°41′E). All computations were within the 16 km emergency planing zone of a generic reactor design considered by the study. The scaling factors have been used to assess the potential environmental risk of the NPP using an integrated approach to the assessment and management of environmental risks from ionizing radiation (D-ERICA). Obtained results should provide baseline information for decision making in terms of operation license provision for the pioneer NPP.     

Simulating bed evolution following the Barlin Dam (Taiwan,China) failure with implications for sediment dynamics modeling of dam removal

The failure of the Barlin Dam in Taiwan, China offers an important case study for evaluating concepts in modeling the rapid erosion and channel recovery following intentional and unplanned dam removals. We present a modeling effort that applied a 1D and quasi-2D uncoupled hydraulics and sediment model (NETSTARS) to evaluate how discretization and parameterization influence thefitofbed elevationpredic-tions to observations following dam failure. Our analysis evaluated the model sensitivity to sediment transport function, active layer thickness, and number of stream tubes used to define the cross-section. Results indicate that a) the model is more sensitive to active layer thickness and sediment transport function than to the number of stream tubes, b) development of dam removal models are likely to benefit from varying the active layer thickness in time, and c) increased lateral discretization does not appear to improve model fit in the steep and rapidly changing river environment at our site. We conclude with discussion on differences between, identifying the need for, and general use of 1D, quasi-2D, and fully 2D models in dam removal and failure analysis.     

Petrologic evolution of Krakatau (Indonesia): Implications for a future activity

Krakatau Volcano is located along a N35E volcanic lineament running through the Sunda straits (Indonesia). Its last activity has been characterized by successive phases, each beginning with the construction of a cone, and ending with its destruction and the formation of a caldera. The two last (pre- and post-1883) cycles are well known, but the more ancient ones are not so clearly defined.Lavas of Krakatau belong to an andesitic series, in which fractional crystallization plays the most important role. The petrologic evolution is characterized by a cyclicity in good agreement with the structural evolution: the succession is regular: basalts, basic andesites, acid andesites, dacites. A gap between acid and basic andesites occurs in each cycle. The destructive stages correspond to the occurrence of dacitic terms.The Anak cycle was characterized from 1927 to 1979 by basalts and basic andesites; the 1981 eruption involved a more differentiated magma (close to dacitic). Detailed study of the petrologic evolution since 1883 emphasizes the predominant role of fractional crystallization. This process occurred during a very short period, between 1979 and 1981. Separation of labradorite, augite, olivine and magnetite from parental basic andesite may generate the dacitic descendant, in a shallow reservoir (P_H2O estimated about 0.5 kbar). Implications for a future activity are considered.     

Late Quaternary evolution of lower reaches of the Volga River (Raygorod section) based on luminescence dating

The normally-closed Caspian Sea is known for large changes in relative sea-level (of ∼170 m) during the late Quaternary. These transgressive/regressive events influenced the topography, sedimentation and ecosystems of a large area, of up to 1 million km². The Volga River has played an important role in the water balance of the Caspian Quaternary basins but our understanding of the temporal evolution is poorly constrained. Recent studies on the evolution of the Lower Volga have focused mainly on the subaerial sequence of loess-palaeosol series corresponding to a long-duration Caspian low stand (the so-called “Atelian regression” from ∼90 to ∼25 ka). In this study we address, for the first time, the temporal evolution of the Volga River during the late Quaternary, as recorded in the many layers of alluvial sands at the Raygorod reference section. This 50 m high outcrop contains a complicated sequence of different types of interlayered alluvium (channel and floodplain facies), a loess-palaeosol sequence with a weakly developed palaeosol, and marine sediments of the Khvalynian transgression (Chocolate Clay facies). The new chronology, based on 35 samples, is derived using optically stimulated luminescence (OSL) analysis of sand-sized quartz, with support from post-infra-red infra-red stimulated luminescence (post-IR IRSL) from K-rich feldspar grains to date the older parts of the section. The new ages identify five stages of the topography development in the northern parts of the Lower Volga: (1) an MIS 5a flood-plain in deltaic/estuary environments (>90 ka) during a high-stand of the Caspian Sea (Hyrcanian transgression); (2) a transition from deltaic/estuary conditions to a river valley with normal alluvial sedimentation and sporadic stabilization reflected in palaeosol development (80–70 ka); (3) a palaeo-Volga channel migration at elevations of 4–8 m msl during 69–62 ka, evidence of a brief increase in Caspian Sea-level and blocking of the Volga flow; (4) a subaerial stage with high-speed accumulation of loess during MIS 4 to MIS 2, containing one weakly developed palaeosol (MIS 3c) and pedocomplex of three combined palaeosols of the beginning of MIS2 (30–24 ka); (5) a rapid Khvalynian transgression, starting at the Raygorod location at ∼18.3 ka, with relatively weak marine erosion of the top 40–60 cm of loess cover, presumably because of the rapid migration of the coastline in the flat Northern Caspian Lowland.     

Application of singular value decomposition (SVD) and semi-discrete decomposition (SDD) techniques in clustering of geochemical data: an environmental study in central Iran

Common multivariate clustering techniques are ineffective in identifying subtle patterns of correlation, and clustering of variables or samples within complex geochemical datasets. This study compares the combination of singular value decomposition (SVD) and semi discrete decomposition (SDD), with that of hierarchical cluster analysis (HCA), to examine patterns within a multielement soil geochemical dataset from an agricultural area in the vicinity of Pb–Zn mining operations in central Iran. SVD was used to both identify patterns of correlation between variables and samples and to “denoise” the data, and SDD to simultaneously cluster the samples and variables. The results reveal various spatial associations of mining waste-associated metals As, Ba, Pb and Zn, and within the remaining elements whose distribution is largely controlled by the major oxides. SVD–SDD was found to be superior to HCA, in its ability to detect subtle clusters in soil geochemistry indicative of mine-related contamination in the study area.     

Seismic behavior and modeling of steel reinforced concrete (SRC) walls

The steel reinforced concrete (SRC) wall consists of structural steel embedded at the boundary elements of a reinforced concrete (RC) wall. The use of SRC walls has gained popularity in the construction of high‐rise buildings because of their superior performance over conventional RC walls. This paper presents a series of quasi‐static tests used to examine the behavior of SRC walls subjected to high axial force and lateral cyclic loading. The SRC wall specimens showed increased flexural strength and deformation capacity relative to their RC wall counterpart. The flexural strength of SRC walls was found to increase with increasing area ratio of embedded structural steel, while the section type of embedded steel did not affect the wall's strength. The SRC walls under high axial force ratio had an ultimate lateral drift ratio of approximately 1.4%. In addition, a multi‐layer shell element model was developed for the SRC walls and was implemented in the OpenSees program. The numerical model was validated through comparison with the test data. The model was able to predict the lateral stiffness, strength and deformation capacities of SRC walls with a reasonable level of accuracy. Finally, a number of issues for the design of SRC walls are discussed, along with a collection and analysis of the test data, including (1) evaluation of flexural strength, (2) calculation of effective flexural stiffness, and (3) inelastic deformation capacity of SRC walls. Copyright © 2014 John Wiley & Sons, Ltd.     

天然气水合物体系动态演化研究(Ⅱ):海底滑坡

天然气水合物被认为是大陆边缘沉积物强度变弱的一个因子，从而能解释大陆边缘海底滑坡的一些观测现象。天然气水合物的形成使沉积物强度增加，而其分解则使沉积物强度变弱。虽然无法直接观测沉积物中天然气水合物的活动过程与相应的海底滑坡，大量的背景资料表明，天然气水合物崩解常常有助于触发海底沉积物块体的运动。此外，大型滑塌可以释放大量的固态天然气水合物，水合物在水柱中上浮。大块天然气水合物可以在分解前到达海洋的上部层，一些甲烷可以直接进入大气中。本文综述与天然气水合物体系演化有关的海底滑坡的研究现状。     

Specific features of water circulation and biogenic substance transformation and transport in Tatar Strait (Sea of Japan): Assessment based on the results of mathematical modeling

This study has no analogues in terms of methodology, as it uses three electronic instruments for studying the oceanographic regime of Tatar Strait. GIS “Sakhalin Shelf” was used to reconstruct annual variations in hydrological and hydrochemical characteristics of the marine environment at standard horizons of oceanographic stations in fixed geographic coordinates. An adapted version of the Oceanic Model created in Bergen University was used to reconstruct the spatial and temporal distribution of seawater temperature and salinity, to calculate the density and the circulation rate of water masses and water exchange parameters between the strait and nearby water areas and between regions within the strait. The hydroecological CNPSi-model was next used to assess the annual dynamics of concentrations of biogenic element compounds; the biomasses, biohydrochemical activity, and bioproductivity of microorganisms (heterotrophic bacteria, three phytoplankton and two zooplankton groups), involved in transformation of biogenic substances, as well as the internal fluxes of biogenic substances, governed by microorganism activity and the external load onto the marine ecosystem. Analysis of calculation results made it possible to reveal new features in water circulation in the strait (including those for winter) and, for the first time, to analyze the transport regime of biogenic substances through the external boundaries of Tatar Strait and through the boundaries of three regions identified within the strait.     

The crustal fluid evolution and the causes of earthquakes (III)

This paper is the third one of a series of three papers on the fluid evolution of the crust-upper mantle and the causes of earthquakes. Based on the last two papers, a model of the crustal resistivity structure and the deep-seated fluid evolution is presented, and also a seismogeny theory is set up, which is called the potential-kinetic energy transformation model. In this model, the crustal deep-seated fluid evolution is considered to take the most important effect on the seismogenic process. Taking the Tangshan M7.8 earthquake of 1976 as an example, the earthquakes occurred in a pull-apart rifting basin are analyzed, and finally the crust outgassing in the seismogenic processes is discussed, referring to the here-presented theory of seismogeny. This paper is supported by the Eight Five-Year Target of the State Seismological Bureau, China.