An iterative version of the adaptive Gaussian mixture filter

The adaptive Gaussian mixture filter (AGM) was introduced as a robust filter technique for large-scale applications and an alternative to the well-known ensemble Kalman filter (EnKF). It consists of two analysis steps, one linear update and one weighting/resampling step. The bias of AGM is determined by two parameters, one adaptive weight parameter (forcing the weights to be more uniform to avoid filter collapse) and one predetermined bandwidth parameter which decides the size of the linear update. It has been shown that if the adaptive parameter approaches one and the bandwidth parameter decreases, as an increasing function of the sample size, the filter can achieve asymptotic optimality. For large-scale applications with a limited sample size, the filter solution may be far from optimal as the adaptive parameter gets close to zero depending on how well the samples from the prior distribution match the data. The bandwidth parameter must often be selected significantly different from zero in order to make large enough linear updates to match the data, at the expense of bias in the estimates. In the iterative AGM we introduce here, we take advantage of the fact that the history matching problem is usually estimation of parameters and initial conditions. If the prior distribution of initial conditions and parameters is close to the posterior distribution, it is possible to match the historical data with a small bandwidth parameter and an adaptive weight parameter that gets close to one. Hence, the bias of the filter solution is small. In order to obtain this scenario, we iteratively run the AGM throughout the data history with a very small bandwidth to create a new prior distribution from the updated samples after each iteration. After a few iterations, nearly all samples from the previous iteration match the data, and the above scenario is achieved. A simple toy problem shows that it is possible to reconstruct the true posterior distribution using the iterative version of the AGM. Then a 2D synthetic reservoir is revisited to demonstrate the potential of the new method on large-scale problems.     

German research in transport geography: Life in the space between objective analysis and political advice1

Transportation has been a bone of public contention for decades, the discussion ranging from traffic-calming measures in individual streets to the continual growth of global transport movements. In the last 20 years transport topics have also received increased attention within the discipline of geography, be it academic, professional or in schools, but the topics addressed by today's transport geography have almost nothing in common with the roots of the field. This means that transport geography is a handed-down, hyphenated sub-branch of geography in name only. In fact, the name refers to a field of geography that is experiencing not only all the birthing pains and uncertainties of a discipline in the process of defining a new direction for itself, but also the sense of excitement and thrill of the new. This paper sets out to show both the role transport geography plays as part of human geography with its concepts and paradigms, and also the role it plays within the political debate on transport. An appeal is made to geographers to become more involved in this branch of our science.     

The structure of monomeric and dimeric uranyl adsorption complexes on gibbsite: A combined DFT and EXAFS study

We investigated the structure of uranyl sorption complexes on gibbsite (pH 5.6-9.7) by two independent methods, density functional theory (DFT) calculations and extended X-ray absorption fine structure (EXAFS) spectroscopy at the U-L_III edge. To model the gibbsite surface with DFT, we tested two Al (hydr)oxide clusters, a dimer and a hexamer. Based on polarization, structure, and relaxation energies during geometry optimization, the hexamer cluster was found to be the more appropriate model. An additional advantage of the hexamer model is that it represents both edges and basal faces of gibbsite. The DFT calculations of (monomeric) uranyl sorption complexes show an energetic preference for the corner-sharing versus the edge-sharing configuration on gibbsite edges. The energy difference is so small, however, that possibly both surface species may coexist. In contrast to the edge sites, sorption to basal sites was energetically not favorable. EXAFS spectroscopy revealed in all investigated samples the same interatomic distances of the uranyl coordination environment (R_U-Oax ≈ 1.80 Å, R_U-Oeq ≈ 2.40 Å), and towards the gibbsite surface (R_U-O ≈ 2.87 Å, R_U-Al ≈ 3.38 Å). In addition, two U-U distances were observed, 3.92 Å at pH 9.7 and 4.30 Å at pH 5.6, both with coordination numbers of ∼1. The short U-U distance is close to that of the aqueous uranyl hydroxo dimer, UO₂(OH)₂, reported as 3.875 Å in the literature, but significantly longer than that of aqueous trimers (3.81-3.82 Å), suggesting sorption of uranyl dimers at alkaline pH. The longer U-U distance (4.30 Å) at acidic pH, however, is not in line with known aqueous uranyl polymer complexes. Based on the EXAFS findings we further refined dimeric surface complexes with DFT. We propose two structural models: in the acidic region, the observed long U-U distance can be explained with a distortion of the uranyl dimer to form both a corner-sharing and an edge-sharing linkage to neighboring Al octahedra, leading to R_U-U = 4.150 Å. In the alkaline region, a corner-sharing uranyl dimer complex is the most favorable. The U-O path at ∼2.87 Å in the EXAFS spectra arises from the oxygen atom linking two Al cations in corner-sharing arrangement. The adsorption structures obtained by DFT calculations are in good agreement with the structural parameters from EXAFS analysis: U-Al (3.394 Å), U-U (3.949 Å), and U-O (2.823 Å) for the alkaline pH model, and U-Al (3.279 Å), U-U (4.150 Å), and U-O (2.743 Å) for the acidic pH model. This work shows that by combining EXAFS and DFT, consistent structural models for uranyl sorption complexes can be obtained, which are relevant to predict the migration behavior of uranium at nuclear facilities.     

Ductile deformation of garnet in mylonitic gneisses from the Münchberg Massif (Germany)

Mylonitic gneisses from the Münchberg Massif contain single grains (type I) and polycrystalline aggregates (type II) of garnet displaying a distinct elongation parallel to a macroscopic lineation which is interpreted as the result of ductile deformation. Lattice-preferred orientations of quartz (textures) symmetrical to the macroscopic foliation and lineation and the lack of rotational microfabrics indicate that the bulk deformation was pure shear at least during the latest strain increments. Garnet textures measured by EBSD together with microprobe analyses demonstrate that these two structural types of garnet can be related to two different processes of ductile deformation: (1) For the single grains stretching can be attributed to diffusion creep along grain boundary zones (Coble creep). The related mass transfer is indicated by the fact that primary growth zones are cut off at the long faces of the grains while the related strain shadow domains do not show comparable chemical zoning. Pressure solution and precipitation suitable to produce similar structures can be largely ruled out because retrogressive reactions pointing to the presence of free hydrous fluids are missing. (2) For the polycrystalline garnet aggregates consisting of cores grading into fine-grained mantles, dislocation creep and associated rotation recrystallization can be assumed. Continuous lattice rotation from the core to the outer polycrystalline rim allow a determination of the related dominant slip systems which are {100}<010> and equivalent systems according to the cubic lattice symmetry. The same holds for garnets which appear to be completely recrystallized. For this type of fine-grained aggregates an alternative nucleation model is discussed. Due to penetrative dislocation glide in connection with short range diffusion and the resulting lattice rotation, primary growth zones are strongly disturbed.Since for the considered rock unit of the Münchberg Massif peak metamorphic temperatures between 630 and 670 °C can be assumed, this study clearly demonstrates that the inferred processes of ductile garnet deformation can occur not only in HT regimes as often suggested in the literature even if embedded within a matrix of “low-strength” minerals like quartz, feldspars and micas.     

Transfer and early diagenesis of biogenic silica oxygen isotope signals during settling and sedimentation of diatoms in a temperate freshwater lake (Lake Holzmaar, Germany)

We have investigated the transfer of oxygen isotope signals of diatomaceous silica (δ¹⁸O_diatom) from the epilimnion (0-7 m) through the hypolimnion to the lake bottom (∼20 m) in freshwater Lake Holzmaar, Germany. Sediment-traps were deployed in 2001 at depths of 7 and 16 m to harvest fresh diatoms every 28 days. The 7 m trap collected diatoms from the epilimnion being the main zone of primary production, while the 16 m trap collected material already settled through the hypolimnion. Also a bottom sediment sample was taken containing diatom frustules from approximately the last 25 years. The δ¹⁸O_diatom values of the 7 m trap varied from 29.4‰ in spring/autumn to 26.2‰ in summer according to the temperature dependence of oxygen isotope fractionation and represent the initial isotope signal in this study. Remarkably, despite the short settling distance δ¹⁸O_diatom values of the 7 and the 16 m trap were identical only during spring and autumn seasons while from April to September δ¹⁸O_diatom values of the 16 m trap were roughly ∼1.5‰ enriched in ¹⁸O compared to those of the 7 m trap. Isotopic exchange with the isotopically lighter water of the hypolimnion would shift the δ¹⁸O_diatom value to lower values during settling from 7 to 16 m excluding this process as a cause for the deviation. Dissolution of opal during settling with intact organic coatings of the diatom cells and near neutral pH of the water should only cause a minor enrichment of the 16 m values. Nevertheless, opal from the bottom sediment was found to be 2.5‰ enriched in ¹⁸O compared to the weighted average of the opal from the 7 m trap. Thus, resuspension of bottom material must have contributed to the intermediate δ¹⁸O_diatom signal of the 16 m trap during summer. Dissolution experiments allowed further investigation of the cause for the remarkably enriched δ¹⁸O_diatom value of the bottom sediment. Experiments with different fresh diatomaceous materials show an increase of opaline ¹⁸O at high pH values which is remarkably reduced when organic coatings of the cells still exist or at near neutral pH. In contrast, high pH conditions do not affect the δ¹⁸O_diatom values of sub-fossil and even fossil opal. IR analyses show that the ¹⁸O enrichment of the sedimentary silica is associated with a decrease in Si-OH groups and the formation of Si-O-Si linkages. This indicates a silica dehydroxylation process as cause for the isotopic enrichment of the bottom sediment. Silica dissolution and dehydroxylation clearly induce a maturation process of the diatom oxygen isotope signal presumably following an exponential behaviour with a rapid initial phase of signal alteration. The dynamics of this process is of particular importance for the quantitative interpretation of sedimentary δ¹⁸O_diatom values in terms of palaeothermometry.     

Early Jurassic rift structures associated with the Soapaga and Boyacá faults of the Eastern Cordillera, Colombia: Sedimentological inferences and regional implications

The NW-trending Bucaramanga fault links, at its southern termination, with the Soapaga and Boyacá faults, which by their NW trend define an ample horsetail structure. As a result of their Neogene reactivation as reverse faults, they bound fault-related anticlines that expose the sedimentary fill of two Early Jurassic rift basins. These sediments exhibit the wedge-like geometry of rift fills related to west-facing normal faults. Their structural setting was controlled further by segmentation of the bounding faults at approximately 10 km intervals, in which each segment is separated by a transverse basement high. Isopach contours and different facies associations suggest these transverse anticlines may have separated depocenters of their adjacent subbasins, which were shaped by a slightly different subsidence history and thereby decoupled. The basin fill of the relatively narrow basin associated with the Soapaga fault is dominated by fanglomeratic successions organized in two coarsening-upward cycles. In the larger basin linked to the Boyacá fault, the sedimentary fill consists of two coarsening-upward sequences that, when fully developed, vary from floodplain to alluvial fan deposits. These Early Jurassic rift fills temporally constrain the evolution of the Bucaramanga fault, which accommodated right-lateral displacement during the early Mesozoic rift event.     

Assessing the Vibrational Frequencies of the Cathedral of Cologne (Germany) by Means of Ambient Seismic Noise Analysis

Ambient seismic noise measurements were conducted inside the Cathedral of Cologne (Germany) for assessing its frequencies of vibration and for checking whether these occur in the range where soil amplification is expected. If this is the case, damages may increase in case of an earthquake due to an increased structural response of the building. Analysis of the ratio between the horizontal and vertical components of the spectra recorded at stations located inside the building as well as the ratio between the corresponding components of the spectra recorded simultaneously inside the building and at a reference station placed in the basement of the cathedral indicated several modes of vibration. Facilitated by these results an assessment of the seismic vulnerability was attempted for a 2D ground motion scenario using the finite element method.     

Multi-metal contaminant dynamics in temporarily flooded soil under sulfate limitation

In many river basins, floodplain soils have accumulated a variety of metal contaminants, which might be released during periods of flooding. We investigated the dynamics of copper, cadmium, lead, zinc, and nickel in a contaminated freshwater floodplain soil under a realistic sulfate-limited flooding regime in microcosm experiments. We found that most contaminants were initially mobilized by processes driven by the reductive dissolution of Fe(III) and Mn(IV, III) (hydr)oxides. Subsequently, bacterial sulfate respiration resulted in the transformation of the entire available sulfate (2.3 mmol/kg) into chromous reducible sulfur (CRS). Cu K-edge X-ray absorption fine structure (XAFS) spectroscopy revealed that the soil Cu speciation changed from predominantly Cu(II) bound to soil organic matter (SOM) intermittently to 14% metallic Cu(0) and subsequently to 66% copper sulfide (Cu_xS). These Cu_xS precipitates accounted for most of the formed CRS, suggesting that Cu_xS was the dominant sulfide phase formed in the flooded soil. Sequential metal extractions, in agreement with CRS results, suggested that easily mobilizable Cd was completely and Pb partially sequestered in sulfide precipitates, controlling their dissolved concentrations to below detection limits. In contrast, Zn and Ni (as well as Fe) were hardly sequestered into sulfide phases, so that micromolar levels of dissolved Zn and Ni (and millimolar dissolved Fe(II)) persisted in the reduced soil. The finding that Cu, Cd, and Pb were sequestered (but hardly any Zn, Ni, and Fe) is consistent with the thermodynamically predicted sulfide ladder following the increasing solubility products of the respective metal sulfides. The observation that Cd and Pb were sequestered in sulfides despite the presence of remaining SOM-bound Cu(II) suggested that the kinetics of Cu(II) desorption, diffusion, and/or Cu_xS precipitation interfered with the sulfide ladder. We conclude that the dynamics of multiple metal contaminants are intimately coupled under sulfate limitation by the relative thermodynamic stabilities and formation kinetics of the respective metal sulfides.     

Changes in Zn speciation during soil formation from Zn-rich limestones

In order to better understand the long-term speciation and fractionation of Zn in soils, we investigated three soils naturally enriched in Zn (237–864 mg/kg Zn) from the weathering of Zn-rich limestones (40–207 mg/kg Zn) using extended X-ray absorption fine structure (EXAFS) spectroscopy and sequential extractions. The analysis of bulk EXAFS spectra by linear combination fitting (LCF) indicated that Zn in the oolitic limestones was mainly present as Zn-containing calcite (at site Dornach), Zn-containing goethite (Gurnigel) and Zn-containing goethite and sphalerite (Liestal). Correspondingly, extraction of the powdered rocks with 1 M NH₄-acetate at pH 6.0 mobilized only minor fractions of Zn from the Gurnigel and Liestal limestones (<30%), but most Zn from the Dornach rock (81%). In the Dornach soil, part of the Zn released from the dissolving limestone was subsequently incorporated into pedogenic hydroxy-interlayered vermiculite (Zn-HIV, 30%) and Zn-containing kaolinite (30%) and adsorbed or complexed by soil organic and inorganic components (40%). The Gurnigel and Liestal soils contained substantial amounts of Zn-containing goethite (50%) stemming from the parent rock, smaller amounts (20%) of Zn-containing kaolinite (and possibly Zn-HIV), as well as adsorbed or complexed Zn-species (30%). In the soil from Liestal, sphalerite was only found in trace amounts, indicating its dissolution during soil formation. In sequential extractions, large percentages of Zn (55–85%) were extracted in recalcitrant extraction steps, confirming that Zn-HIV, Zn-containing kaolinite and Zn-containing goethite are highly resistant to weathering. These Zn-bearing phases thus represent long-term hosts for Zn in soils over thousands of years. The capability of these phases to immobilize Zn in heavily contaminated soils may however be limited by their uptake capacity (especially HIV and kaolinite) or their abundance in soil.