Use of Noise to Augment Training Data: A Neural Network Method of Mineral–Potential Mapping in Regions of Limited Known Deposit Examples

One of the main factors that affects the performance of MLP neural networks trained using the backpropagation algorithm in mineral-potential mapping isthe paucity of deposit relative to barren training patterns. To overcome this problem, random noise is added to the original training patterns in order to create additional synthetic deposit training data. Experiments on the effect of the number of deposits available for training in the Kalgoorlie Terrane orogenic gold province show that both the classification performance of a trained network and the quality of the resultant prospectivity map increasesignificantly with increased numbers of deposit patterns. Experiments are conducted to determine the optimum amount of noise using both uniform and normally distributed random noise. Through the addition of noise to the original deposit training data, the number of deposit training patterns is increased from approximately 50 to 1000. The percentage of correct classifications significantly improves for the independent test set as well as for deposit patterns in the test set. For example, using ±40% uniform random noise, the test-set classification performance increases from 67.9% and 68.0% to 72.8% and 77.1% (for test-set overall and test-set deposit patterns, respectively). Indices for the quality of the resultant prospectivity map, (i.e. D/A, D × (D/A), where D is the percentage of deposits and A is the percentage of the total area for the highest prospectivity map-class, and area under an ROC curve) also increase from 8.2, 105, 0.79 to 17.9, 226, 0.87, respectively. Increasing the size of the training-stop data set results in a further increase in classification performance to 73.5%, 77.4%, 14.7, 296, 0.87 for test-set overall and test-set deposit patterns, D/A, D × (D/A), and area under the ROC curve, respectively.     

Rates of deformation of an extensional growth fault/raft system (offshore Congo, West African margin) from combined accommodation measurements and 3-D restoration

The aim of this paper is to quantify the evolution in time and space of the accommodation (space available for sedimentation) in the case of a growth fault structure resulting from gravity‐induced extension comprising a listric fault/raft system located along the West African margin. To achieve this, use was made of an original approach combining two complementary techniques (accommodation variation measurements and 3‐D restoration) in order to quantify vertical and horizontal displacement related to deformation, using a data set made up of a 3‐D seismic survey and well logs. We applied sequence stratigraphic principles to (i) define a detailed stratigraphic framework for the Albo‐Cenomanian and (ii) measure subsidence rates from accommodation variations. 3‐D restoration was used to (iii) reconstruct the evolution of the 3‐D geometry of the fault system. The rates of horizontal displacement of structural units were measured and linked to successive stages in the growth of the fault system. Subsidence of the structural units exhibits three scales of variation: (1) long‐term variation (10 Ma) of c. 80 m Ma^?1 for a total subsidence of about 1400 m, compatible with the general subsidence of a passive margin, and (2) short‐term variations (1–5 Ma) corresponding to two periods of rapid subsidence (about 150–250 m Ma^?1) alternating with periods of moderate subsidence rate (around 30 m Ma^?1). These variations are linked to the development of the fault system during the Albian (with downbuilding of the raft and development of the initial basin located in between). During the Cenomanian, the development of the graben located between the lower raft and the initial basin did not seem to affect the vertical displacements. (3) High‐frequency variations (at the scale of genetic unit sets) range between ?50 and 250 m for periods of 0.2–2 Ma. Accommodation variations governed these cycles of progradation/retrogradation rather than sediment flux variations. In addition, the nine wells display a highly consistent pattern of variation in accommodation. This suggests that the genetic unit sets were controlled at a larger scale than the studied system (larger than 20 km in wavelength), for example, by eustatic variations. Translation rates are between 3 and 30 times higher than subsidence rates. Therefore, in terms of amplitude, the main parameter controlling the space available for sedimentation is the structural development of the fault system, that is to say, the seaward translation of the raft units, itself resulting from a regional gravity‐driven extension.     

Wind velocity and sand transport on a barchan dune

We present measurements of wind velocity and sand flux performed on the windward side of a large barchan dune in Jericoacoara, northeastern Brazil. From the measured profile, we calculate the air shear stress using an analytical approximation and treat the problem of flow separation by an heuristic model. We find that the results from this approach agree well with our field data. Moreover, using the calculated shear velocity, we predict the sand flux according to well-known equilibrium relations and with a phenomenological continuum saltation model that includes saturation transients and thus allows for nonequilibrium conditions. Based on the field data and theoretical predicted results, we indicate the principal differences between saturated and nonsaturated sand flux models. Finally, we show that the measured dune moves with invariant shape and predict its velocity from our data and calculations.     

Seasonal variation of the soil seed bank of grasses in central Argentina as related to grazing and shrub cover

The objectives of this research were to (a) study seasonal variation of the seed bank of grasses in two areas with different grazing history, (b) analyse the effect of the presence of shrubs on the seed bank and (c) study seed viability and dormancy in buried seeds of Piptochaetium napostaense and Stipa tenuis. The seed density of grasses was in general low and showed seasonal variation. Most of the grasses showed maximum seed density in December when seed dispersal occurs. Shrub cover did not show a marked influence in seed accumulation. The relatively high number of damaged seeds suggests that predation may be an important factor in seed bank dynamics in the Caldenal. Seeds of dominant perennial forage grasses showed dormancy that could explain, at least in part, persistence of these species in spite of periodical disturbance.     

SHRIMP Age and Geochemistry of the Bikou Volcanic Terrane: Implications for Neoproterozoic Tectonics on the Northern Margin of the Yangtze Craton

The Bikou volcanic terrane is predominated by subalkaline tholeiitic lavas. Rock samples display lower initial ratios of Sr and Nd, 0.701248-0.704413 and 0.511080-0.512341 respectively. 207Pb and 208Pb are significantly enriched in the lavas. Most samples have positive εNd, which indicates that the magma was derived from EM-type mantle source, while a few samples with negative εNd indicate that there was contamination in the magma evolution. Magma differentiation is demonstrated by variations of LREE and LILE from depletion to enrichment. Additionally, normalized REE patterns and trace elements showed that lavas from the Bikou volcanic terrane have similar characteristics to those of basalts in arc settings caused by subduction and collision. Analyses showed that the Bikou volcanic terrane is a volcanic arc. New evidence proved that the Hengdan Group, north of the Bikou arc, is a turbidite terrane filling a forearc basin. Consequently, the Bikou volcanic terrane and the Hengdan turbidite terrane const     

Analytical Model for Permeability Evolution in Microcracking Rock

— Analytical expressions to predict the enhancement of permeability due to stress-induced microcracking in initially low porosity rock are presented. A fracture mechanical microcrack model is employed to derive integrated effective hydraulic variables as a function of stress, which are then used to calculate the evolution of permeability using the statistically-based Dienes model. The model enables determination of permeability enhancement as a function of two loading parameters and three material parameters. Results are in reasonable agreement with experimental measurements and indicate that appreciable increases in permeability can be anticipated during brittle failure. The analytical nature of the model makes it easily incorporatable into numerical models that require quantification of the permeability evolution as a function of stress, for which there is currently no law.     

Analysis of Numerical Procedures for Solving Mass Transport Problems at a Regional Scale

The principal mass transport models and the ways of their implementation are considered. Some problems associated with the application of numerical methods for mass transport modeling in problems of dissolved chemicals migration at a regional scale are discussed. A new method proposed in this study allows the determination of the space structure of a groundwater flow and the simulation of steady-state convective mass transport.