Potential of legacy 2D seismic data for deep targeting and structural imaging at the Neves–Corvo massive sulphide-bearing deposit,Portugal

Seismic methods are becoming an established choice for deep mineral exploration after being extensively tested and employed for the past two decades. To investigate whether the early European mineral-exploration datasets had potential for seismic imaging that was overlooked, we recovered a low-fold legacy seismic dataset from the Neves–Corvo mine site in the Iberian Pyrite Belt in southern Portugal. This dataset comprises six 4–6 km long profiles acquired in 1996 for deep targeting. Using today's industry-scale processing algorithms, the world-class, ca. 150 Mt, Lombador massive sulphide and other smaller deposits were better imaged. Additionally, we also reveal a number of shallow but steeply dipping reflections that were absent in the original processing results. This study highlights that legacy seismic data are valuable and should be revisited regularly to take advantage of new processing algorithms and the experiences gained from processing such data in hard-rock environments elsewhere. Remembering that an initial processing job in hard rock should always aim to first obtain an overall image of the subsurface and make reflections visible, and then subsequent goals of the workflow could be set to, for example understanding relative amplitude ratios. The imaging of the known mineralization implies that this survey could likely have been among one of the pioneer studies in the world that demonstrated the capability of directly imaging massive sulphide deposits using the seismic method.     

Simultaneous estimation of shape factor and depth of subsurface cavities from residual gravity anomalies using feed-forward back-propagation neural networks

We develop a new method of using feed-forward back-propagation (FFBP) neural networks to simultaneously estimate shape factor and depth of gravity anomalies. The advantages compared to neural network methods are the following: no pre-assumptions are made on source shape, the FFBP neural network estimates both depth and shape factor of source bodies and, once trained, works well for any new data in the training space, without repeating the initial calculations.     

An alternative direct method towards mean dynamic topography computations

An alternative “direct method” to “mean dynamic topography” (MDT) computations using satellite altimetry-derived “mean sea surface” (MSS) and “global geopotential model” (GGM), without direct application of the geoid, is devised. The developed approach, which is based on derivation of an equipotential surface of the gravity field of the Earth that fits to global MSS in least squares sense, is formulated via a constrained optimization problem. The validity of our method is numerically tested by computing a global MDT model based on DNSC08 MSS model and EGM2008 GGM as input data.     

Discrimination of sand dunes and loess deposits using grain-size analysis in northeastern Iran

Identification and characterization of aeolian deposits in arid environments provide information on mechanisms of loess and sand accumulation. The objectives of this study were to (i) identify the distribution of aeolian deposits, (ii) discriminate loess and sand deposits using granulometric data, and (iii) describe the aeolian deposition in Sarakhs area, northeastern Iran. Particle size distributions of 26 surface samples were determined using a laser grain-size analyzer. Fine sand, very fine sand, and very coarse silt were dominant fractions in studied sediments, and the sum of these fractions ranged from 46.9% in loess deposits to 93.8% in sand dunes. The mean grain size (M _z ) of sand dunes ranged from 3.31 to 3.54 ?, which gradually changed to 4.09 to 5.50 ? in loess deposits. Sorting, skewness, and kurtosis ranged from 0.84 to 1.94 ?, 0.18 to 0.49, and 0.76 to 2.38, respectively. Aeolian deposits in the area resulted in the incorporation of Hariroud River system and Kopeh Dagh Mountains for aeolian particle production and accumulation. Alluvial comminution in Hariroud River is suggested the main mechanism of sand and silt production and flood plain environment the main reservoir of these particles. The mountains of Kopeh Dagh act as a barrier and play a key role for sand and loess accumulation.     

Prediction of the median IDA curve by employing a limited number of ground motion records

A methodology has been proposed which can be used to reduce the number of ground motion records needed for the reliable prediction of the median seismic response of structures by means of incremental dynamic analysis (IDA). This methodology is presently limited to predictions of the median IDA curve only. The reduction in the number of ground motion records needed to predict the median IDA curve is achieved by introducing a precedence list of ground motion records. The determination of such a list is an optimization problem, which is solved in the paper by means of (1) a genetic algorithm and (2) a proposed simple procedure. The seismic response of a simple, computationally non‐demanding structural model has been used as input data for the optimization problem. The presented example is a three‐storey‐reinforced concrete building, subjected to two sets of ground motion records, one a free‐field set and the other a near‐field set. It is shown that the median IDA curves can be predicted with acceptable accuracy by employing only four ground motion records instead of the 24 or 30 records, which are the total number of ground motion records for the free‐field and near‐field sets, respectively. Copyright © 2007 John Wiley & Sons, Ltd.     

Estimation of groundwater storage from seismic data using deep learning

Convolutional neural networks can provide a potential framework to characterize groundwater storage from seismic data. Estimation of key components, such as the amount of groundwater stored in an aquifer and delineate water table level, from active-source seismic data are performed in this study. The data to train, validate and test the neural networks are obtained by solving wave propagation in a coupled poroviscoelastic–elastic media. A discontinuous Galerkin method is applied to model wave propagation, whereas a deep convolutional neural network is used for the parameter estimation problem. In the numerical experiment, the primary unknowns estimated are the amount of stored groundwater and water table level, while the remaining parameters, assumed to be of less of interest, are marginalized in the convolutional neural network-based solution. Results, obtained through synthetic data, illustrate the potential of deep learning methods to extract additional aquifer information from seismic data, which otherwise would be impossible based on a set of reflection seismic sections or velocity tomograms.     

Agent-based tsunami evacuation modeling of unplanned network disruptions for evidence-driven resource allocation and retrofitting strategies

The constant threat from landslides in the northeastern part of Istria, Croatia, calls for the need to apply accurate and reliable methods in landslide hazard assessment in order to prevent landslide damage and to set an early warning system if necessary. Furthermore, landslide susceptibility and hazard assessment enable optimal area management and regional urban planning. The study area is in the northeastern and central part of the Istrian Peninsula, well known as an area of frequent, small and shallow slope instability phenomena. Landslide susceptibility assessment in the area around the city of Buzet was performed using a deterministic landslide susceptibility model in the LS-RAPID software. LS-RAPID was developed to analyze stability at one single location, but the performed analysis has shown that LS-RAPID can be used as a powerful tool in landslide susceptibility and hazard assessment on regional scale. The objective of this paper is to establish the influence of the runout potential on the enlargement of the landslide-susceptible zones, due to expansion of the failure area around the initial failure zone. Performed analysis of rainfall return periods shows the frequency of landslide occurrence and provides the possible correlation with the time component of landslide hazard in the area.     

Flood frequency analysis based on simulated peak discharges

Flood frequency approaches vary from statistical methods, directly applied on the observed annual maximum flood series, to adopting rainfall–runoff simulation models that transform design rainfalls to flood discharges. Reliance on statistical flood frequency analysis depends on several factors such as the selected probability distribution function, estimation of the function parameters, possible outliers, and length of the observed flood series. Through adopting the simulation approach in this paper, watershed-average rainfalls of various occurrence probabilities were transformed into the corresponding peak discharges using a calibrated hydrological model. A Monte Carlo scheme was employed to consider the uncertainties involved in rainfall spatial patterns and antecedent soil moisture condition (AMC). For any given rainfall depth, realizations of rainfall spatial distribution and AMC conditions were entered as inputs to the model. Then, floods of different return periods were simulated by transforming rainfall to runoff. The approach was applied to Tangrah watershed in northeastern Iran. It was deduced that the spatial rainfall distribution and the AMCs exerted a varying influence on the peak discharge of different return periods. Comparing the results of the simulation approach with those of the statistical frequency analysis revealed that, for a given return period, flood quantiles based on the observed series were greater than the corresponding simulated discharges. It is also worthy to note that existence of outliers and the selection of the statistical distribution function has a major effect in increasing the differences between the results of the two approaches.