Short-term water quality variable prediction using a hybrid CNN–LSTM deep learning model

Stochastic Environmental Research and Risk Assessment - Water quality monitoring is an important component of water resources management. In order to predict two water quality variables, namely...     

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.     

Close up observation and inversion of low-altitude ENA emissions during a substorm event

During a series of substorm events on November 12 2004, the Neutral Atom Detector Unit(NUADU) flying onboard the TC-2 spacecraft observed, close to perigee, bright low-altitude Energetic Neutral Atom(ENA) emissions from both north and south poles. Through utilizing high temporal and spatial resolution data inversion techniques we present here a global view the associated processes of energetic ion deposition and injection at the uniquely high temporal resolution of a few minutes. It is demonstrated that the ENA emissions observed came mainly from the, low-altitude, high magnetic latitude polar region. Since precipitating ions follow magnetic field lines and have more chance to produce ENA emissions through charge exchange there.Therefore, the employment of a low-orbit satellite to record at high temporal and spatial resolution ENA image data is inferred to be the best way to monitor those dynamic ring current variations that occur during geomagnetic activity.     

The fragmentation of expanding shells – I. Limitations of the thin-shell approximation

We investigate the gravitational fragmentation of expanding shells in the context of the linear thin-shell analysis. We make use of two very different numerical schemes; the flash adaptive mesh refinement code and a version of the Benz smoothed particle hydrodynamics code. We find that the agreement between the two codes is excellent. We use our numerical results to test the thin-shell approximation and we find that the external pressure applied to the shell has a strong effect on the fragmentation process. In cases where shells are not pressure-confined, the shells thicken as they expand and hydrodynamic flows perpendicular to the plane of the shell suppress fragmentation at short wavelengths. If the shells are pressure-confined internally and externally, so that their thickness remains approximately constant during their expansion, the agreement with the analytical solution is better.     

Reaction of soil enzymes and spring barley to copper chloride and copper sulphate

Study was carried out to compare the effect of the degree of contamination of soil with copper on the soil enzyme activity, depending on the type of chemical compound with which a metal reaches the soil and on the method of soil use. An effect of increasing doses of copper on growth and the development of spring barley has been studied. The study was carried out as a pot experiment on two types of soil: loamy sand and light loam. Copper was introduced to soil as copper(II) sulphate(VI) and copper(II) chloride at four doses: 150, 300, 450 and 600 mg Cu²⁺ kg^?1 d.m. of soil. Soil without copper was used as a control sample. The experiment showed that contamination of soil with copper applied as CuCl₂ and CuSO₄·5H₂O upset the homeostasis of soil. Dehydrogenases and urease were the most sensitive to contamination of soil with Cu²⁺, both as copper(II) chloride and as copper(II) sulphate(VI), and acid phosphatase and alkaline phosphatase were the least sensitive. Greater changes in enzymatic activity were observed in loamy sand than in light loam. The differences in the impact of copper(II) sulphate(VI) and copper(II) chloride on the soil microbiome were relatively small, because mean ED₅₀ for copper(II) chloride was 341 mg Cu²⁺ kg^?1 d.m. of soil, and for copper(II) sulphate(VI) it was 364 mg Cu²⁺ kg^?1 d.m. of soil. This inhibitory effect of copper persisted throughout the period of the experiment. The sensitivity of spring barley to copper compounds was lower than the enzymes under study.     

A Simple Analytical Formula for the Leakage Flux Through a Perforated Aquitard

In this methods note, we present a simple analytical formula to quantify the steady‐state leakage flux (Q) over a perforated aquitard. The flux depends on the aquitard thickness (D), the radius of the perforation (R), the hydraulic conductivity of the material inside the perforation (k_fill), the conductivities of the overlying and underlying aquifers (k₁ and k₂, respectively), and the head difference between the two aquifers (ΔH): This equation assumes an aquitard separating two homogeneous and infinite aquifers (R ? aquifer thickness) in which radial flow to and from the perforation occurs, with no other recharge or discharge boundaries near the perforation. The flux through a perforation in a hypothetical case study with D = 10 m, k₁ = 10 m/d, k₂ = 20 m/d, R = 0.072 m, and ΔH = 1 m ranges between less than 1 mL/d if the hole is backfilled with bentonite (k_fill = 10^?4 m/d), to several liters per day if the perforation is backfilled with sand from the overlying aquifer (k_fill = 10 m/d), to several m³/d if the perforation forms an open conduit (k_fill = 10⁵ m/d). The leakage fluxes calculated with this model agree well with those calculated using a numerical model (MODFLOW).     

Stability of source mechanisms inverted from P‐wave amplitude microseismic monitoring data acquired at the surface

We study the stability of source mechanisms inverted from data acquired at surface and near‐surface monitoring arrays. The study is focused on P‐wave data acquired on vertical components, as this is the most common type of acquisition. We apply ray modelling on three models: a fully homogeneous isotropic model, a laterally homogeneous isotropic model and a laterally homogeneous anisotropic model to simulate three commonly used models in inversion. We use geometries of real arrays, one consisting in surface receivers and one consisting in ‘buried’ geophones at the near‐surface. Stability was tested for two of the frequently observed source mechanisms: strike‐slip and dip‐slip and was evaluated by comparing the parameters of correct and inverted mechanisms. We assume these double‐couple source mechanisms and use quantitatively the inversion allowing non‐double‐couple components to measure stability of the inversion. To test the robustness we inverted synthetic amplitudes computed for a laterally homogeneous isotropic model and contaminated with noise using a fully homogeneous model in the inversion. Analogously amplitudes computed in a laterally homogeneous anisotropic model were inverted in all three models. We show that a star‐like surface acquisition array provides very stable inversion up to a very high level of noise in data. Furthermore, we reveal that strike‐slip inversion is more stable than dip‐slip inversion for the receiver geometries considered here. We show that noise and an incorrect velocity model may result in narrow bands of source mechanisms in Hudson's plots.     

Geophysical constraints on deep weathering and water storage potential in the Southern Sierra Critical Zone Observatory

The conversion of bedrock to regolith marks the inception of critical zone processes, but the factors that regulate it remain poorly understood. Although the thickness and degree of weathering of regolith are widely thought to be important regulators of the development of regolith and its water‐storage potential, the functional relationships between regolith properties and the processes that generate it remain poorly documented. This is due in part to the fact that regolith is difficult to characterize by direct observations over the broad scales needed for process‐based understanding of the critical zone. Here we use seismic refraction and resistivity imaging techniques to estimate variations in regolith thickness and porosity across a forested slope and swampy meadow in the Southern Sierra Critical Zone Observatory (SSCZO). Inferred seismic velocities and electrical resistivities image a weathering zone ranging in thickness from 10 to 35 m (average = 23 m) along one intensively studied transect. The inferred weathering zone consists of roughly equal thicknesses of saprolite (P‐velocity < 2 km s^?1) and moderately weathered bedrock (P‐velocity = 2–4 km s^?1). A minimum‐porosity model assuming dry pore space shows porosities as high as 50% near the surface, decreasing to near zero at the base of weathered rock. Physical properties of saprolite samples from hand augering and push cores are consistent with our rock physics model when variations in pore saturation are taken into account. Our results indicate that saprolite is a crucial reservoir of water, potentially storing an average of 3 m³ m^?2 of water along a forested slope in the headwaters of the SSCZO. When coupled with published erosion rates from cosmogenic nuclides, our geophysical estimates of weathering zone thickness imply regolith residence times on the order of 10⁵ years. Thus, soils at the surface today may integrate weathering over glacial–interglacial fluctuations in climate. Copyright © 2013 John Wiley & Sons, Ltd.