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.     

Spatial correlation for horizontal and vertical components of acceleration from northern Iran seismic events

The evaluation of seismic risk of spatially distributed systems requires the spatial correlation model for ground motion intensity measures. This study investigates the spatial correlation of four earthquakes recorded in northern Iran. The intra-event spatial correlation for both horizontal and vertical components of spectral acceleration at eight periods in the range of 0.0–3.0 s is estimated using geostatistical tools. An exponential form is chosen to fit experimental semivariograms, and the correlation ranges of spectral accelerations as a function of period are derived. The results show similar trend of correlation ranges for both components. It should be mentioned that the ranges for the vertical component, in general, are higher than those observed for the horizontal one. For both components, the correlation ranges as a function of period are divided into three segments. The first and the third one are increasing while the second one is decreasing with increasing period.     

A general weighted total Kalman filter algorithm with numerical evaluation

An applicable algorithm for Total Kalman Filter (TKF) approach is proposed. Meanwhile, we extend it to the case in which we can consider arbitrary weight matrixes for the observation vector, the random design matrix and possible correlation between them. Also the updated dispersion matrix of the predicted unknown is given. This approach makes use of condition equations and straightforward variance propagation rules. It is applicable to data fusion within a dynamic errors-in-variables (DEIV) model, which usually appears in the determination of the position and attitude of mobile sensors. Then, we apply for the first time the TKF algorithm and its extended version named WTKF to a DEIV model and compare the results. The results show the efficiency of the proposed WTKF algorithm. In particular in the case of large weights, WTKF shows approximately 25% improvement in contrast to TKF approach.     

Paleoenvironment of geomorphic surfaces of an alluvial fan in the eastern Isfahan,Iran, in the light of micromorphology and clay mineralogy

Alluvial fans are one of the most important landforms in geomorphological and paloenvironmental studies. The objective of this study was the application of clay mineral assemblages and micromorphological properties of the studied paleosols in the geomorphic surfaces of an alluvial fan in the eastern Isfahan as proxies for paleoenvironmental and paleoclimatic changes. Micromorphology, X-ray diffraction, and scanning electron microscopy approaches were used to study the representative pedons. The results indicated that the illuviation process in the calcareous soils of the arid regions of the eastern Isfahan was probably in response to Quaternary moist conditions. There was no significant difference between clay coating properties of the studied relict and buried paleosols. Clay mineralogical study suggested that kaolinite and illite were inherited from the parent materials, while smectite and palygorskite were formed in the soil environment. Paleoargillic horizon was characterized by smectite and calcic (especially the calcrete) horizons were dominated by palygorskite. Palygorskite was accumulated by both neoformation and illuviation processes. High clay content, high intensity of smectite peak, and activity of the illuviation process in paleoargillic horizon demonstrated the seasonality of climate (rainfall) even in the moist periods of Quaternary in Central Iran. Clay mineralogical assemblages suggested a trend of increasing environmental aridity in the study area. This study, therefore, highlighted the role of clay mineralogical investigations in arid lands’ geomorphological and paleoenvironmental researches.     

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.     

Local site effect of a clay site in Shiraz based on seismic hazard of Shiraz Plain

Propagation of seismic waves through soil layers would drastically change the frequency content and amplitude-based features of ground motions at the surface. These alterations are known as seismic site effects. Computation of site effects of high-populated areas such as large cities is of great importance (e.g., it is used in development of seismic microzonation of a region). Shiraz is one of the most populous cities of Iran and is located in a high seismic hazardous region. A representative clay site in this city is selected to assess local site effects. The time series and random vibration theory procedure in the frequency domain are implemented to analyze the aforementioned site. Furthermore, the nonlinear dynamic soil behavior is simulated by the equivalent linear method and the nonlinear method via DEEPSOIL program. Three types of soil column uncertainties such as shear wave velocity, modulus reduction, and damping ratio of soil layers as well as depth of underlying rock half-space (D _bed) are considered herein. The mean amplification and standard deviation of natural logarithm of amplification factors are computed for a variety of analysis types. The results of the current study show that the computed mean and standard deviation of amplification factor in ln units by considering only V _S uncertainty are in good agreement with the corresponding ones by considering V _S and modulus reduction and damping ratio variabilities simultaneously for the studied site. Furthermore, it seems that the effect of bedrock depth in definition of spectral shapes of the Iranian seismic building code should be taken into account.     

Mantle flow patterns at the Neyriz Paleo-spreading center, Iran

The mantle peridotites of Neyriz record two successive episodes of plastic deformations; the first one related to the igneous accretion of the lithosphere and the second one developed during the first stage of the emplacement of the peridotites. These two events have been distinguished on the basis of microstructural criteria. The diapiric pattern, particularly relevant to the mantle process beneath spreading ridges, features vertical flow lines and elliptic flow plane trajectories in a pipe and extends along the ridge axis about 5 km. These structures rotate to horizontal and diverge in every direction in a narrow transition zone, a few hundred meters thick, below the Moho discontinuity. Such a diapiric pattern has been recognized in a few places along the Neyriz paleo-ridge. A large amount of magma passed through these mantle diapirs that were probably the main zones feeding the overlying magma chamber. The most common pattern features very regular structures over several kilometers along the strike of the paleo-ridge: the flow plane dips away from the ridge axis, and the flow line is parallel to the spreading direction. This flow pattern is frozen during the gradual accretion of the lithospheric mantle away from the ridge in a steady-state spreading regime. A shear-sense inversion at just below the Moho is commonly observed, pointing to forced asthenospheric flow. The reconstructed orientation of the Neyriz paleo-spreading center is 105°, compatible with the geometry and orientation of harzburgite foliations and lineations and sheeted dikes.     

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.