Audio-frequency magnetotellurics (AMT) is one of the commonly used methods in geophysical exploration; however, its signal energy is relatively weak and easily submerged by various cultural noises, making denoising a critical step in AMT data processing. Currently, deep learning-based neural networks have achieved superior denoising performance compared to traditional methods in many fields, but in AMT denoising, the neglect of the sparsity characteristic of cultural noise results in degraded denoising performance. To enable the neural network to consider sparsity features during the denoising process and thereby enhance denoising accuracy, we adopt a convolutional neural network (CNN) as the network backbone and design a multilevel wavelet convolutional neural network (MWCNN) from the perspective of sparse representation. This network improves CNN blocks via shortcut connections and enhances feature transmission efficiency by replacing pooling layers and interpolation with wavelet transforms, thereby enabling the network to account for the sparsity of cultural noise, capture underlying noise spectral information and improve denoising performance. Furthermore, we discuss the influence of various network parameters on denoising performance. Finally, we validate the effectiveness of MWCNN in AMT denoising through comparative experiments on both synthetic and field AMT datasets against wavelet transform, bounded influence remote reference processing, data-driven tight frame, CNN and residual networks. Comprehensive evaluations based on signal-to-noise ratio, wavelet time-frequency spectra, denoised results and residuals, apparent resistivity and phase curves, error analysis, one-dimensional inversion results and Nyquist diagrams confirm the superiority of MWCNN for AMT denoising.

Flow Duration Curve (FDC) is an essential graphical tool for illustrating the variability of observed historical streamflow. Achieving an advanced understanding of the physical characteristics governing FDCs is crucial for enhancing predictions of FDCs in ungauged basins. However, this remains challenging due to the complex processes that control streamflow components and their interactions. To address this, a novel framework that integrates regionalization and process-based methods is proposed for predicting FDCs in ungauged basins. This framework implements a hydrological similarity-based regionalization method to estimate hydrological model parameters in ungauged basins, enhancing streamflow prediction reliability. It categorizes streamflow into four distinct components based on delayed flow separation: Short-delay, intermediate-delay, long-delay, and baseline-delay. These components are synthesized to construct the FDC, with their interdependencies modeled using a Vine copula structure. A Bayesian-based estimation technique for the copula function parameters is developed to further improve the precision of the predicted FDC. Applied to nine selected MOPEX ungauged basins, the framework demonstrated superior accuracy, especially for low to middle streamflow phases. Moreover, the framework exhibited superior simulation accuracy during the validation period, highlighting its substantial potential for future-oriented water resource management and planning strategies.

Bearing vibration data are often contaminated with noise, which is detrimental to equipment fault diagnosis and predictive maintenance. Denoising bearing vibration data is crucial. Traditional denoising methods have certain limitations. For instance, when employing wavelet denoising, fixed basis functions may fail to perfectly match all signal structures, potentially compromising denoising accuracy. Similarly, when utilizing data-driven tight frame (DDTF) denoising, the learned basis, due to the lack of noise constraints, may incur a risk of overfitting. To optimize denoising performance in both scenarios, this paper proposes a method that combines wavelet transform and DDTF dictionary learning to extract noise based on a doubly sparse dictionary. The specific approach involves mutually cascading the wavelet transform and DDTF. After applying the wavelet transform to the noisy signal, multi-layer wavelet sparse coefficients are obtained. DDTF processing is then applied to each layer of wavelet sparse coefficients. Subsequent inverse transformation achieves noise suppression. This method integrates the structural constraint capability of wavelet decomposition with the learning capability of DDTF, thereby mitigating their respective limitations to some extent. The denoised data are fed into a residual network model, and training results confirm that the proposed method achieves the best classification performance. Experimental results from both data denoising and deep learning classification demonstrate that the proposed method exhibits superior denoising performance. Although the algorithm structure of this method is more complex compared to other approaches, it is meaningful in scenarios where high-precision denoising is required.

Climate change and complex anthropogenic activities have raised significant concerns regarding Precipitation-Runoff Relationships (PRR). Traditional methods, assuming stationary and linear conditions, often fail to adequately capture these intricate links. To address the limitations, we proposed an integrated framework, employing the Driving indices for Precipitation-Runoff relationships within the nonStationary and nonLinear theory approaches (DPRS and DPRL) to identify the possible driving mechanisms in PRR. The framework is validated across five sub-basins (WRB1-WRB5) within the Wei River Basin, known for its high spatiotemporal variability and intense anthropogenic activities. Spatiotemporal dynamics, nonstationary processes, and nonlinear interactions among various factors are assessed, including climate forcing, groundwater, vegetation dynamics, and anthropogenic influences. DPRS and DPRL assessments revealed that baseflow significantly influences PRR but with high uncertainty. Potential evapotranspiration plays a dominant role in driving negative PRR changes in WRB5 (weakening the correlation between precipitation and runoff), while vegetation dynamics negatively affect PRR with lower uncertainty. Anthropogenic influences represented by Impervious Surface Ratio (ISR), Night-Time Light (NTL), and population density (POP) exhibit varying driving levels, with ISR having the strongest and direct impact, closely linked to urbanization processes and scale within the study cases. The mutual validation of DPRS and DPRL confirms the dominance of baseflow in the Wei River Basin, with urbanization contributing to high ISR, NTL, and POP driving levels in WRB2 and WRB3. Afforestation policies intensify vegetation dynamics’ impact in WRB4 and WRB5. This framework extends its utility to analyze various land evapotranspiration and soil moisture content at different depths in the PRR, supported by a physically-based hydrological model. Basin complexity is further employed to validate the reliability of the assessment outcomes. These insights contribute to a more comprehensive understanding of hydrological processes and facilitate informed decisions for sustainable water resource management within the basin.

Audio magnetotelluric (AMT) is commonly used in mineral resource exploration. However, the weak energy of AMT signals makes them susceptible to being overwhelmed by noise, leading to erroneous geophysical interpretations. In recent years, deep learning has been applied to AMT denoising and has shown better denoising performance compared to traditional methods. However, current deep learning denoising methods overlook the characteristics of AMT signals, resulting in reduced denoising accuracy. To enhance the denoising performance of deep learning by better matching the features of AMT signals, we propose a convolutional block attention module (CBAM)-based method for AMT denoising. This method focuses on the features of AMT signals and improves the process from three aspects: 1) in the establishment of the sample set, we adopt a multicomponent form based on the correlation of noise to enable the neural network to explore the potential connections among the components of AMT during the training process, thus constructing a stronger network mapping relationship; 2) in the construction of the neural network, we have introduced the CBAM structure into the residual blocks of the ResNet to enhance the network's feature learning capability by focusing on the characteristics of noise; and 3) in the design of the denoising procedure, we adopt a process of identification before denoising to protect the noise-free data segments from being compromised during the denoising process. Finally, through synthetic, field data experiments, and comparative tests, we demonstrate that our proposed method achieves higher denoising accuracy than some traditional methods and conventional deep learning methods.

To streamline fast-track processing of large data volumes, we have developed a deep learning approach to deblend seismic data in the shot domain based on a practical strategy for generating high-quality training data along with a list of data conditioning techniques to improve the performance of the data-driven model. We make use of unblended shot gathers acquired at the end of each sail line, to which the access requires no additional time or labor costs beyond the blended acquisition. By manually blending these data, we obtain training data with good control of the ground truth and fully adapted to the given survey. Furthermore, we train a deep neural network using multi-channel inputs that include adjacent blended shot gathers as additional channels. The prediction of the blending noise is added in as a related and auxiliary task with the main task of the network being the prediction of the primary-source events. Blending noise in the ground truth is scaled down during the training and validation process due to its excessively strong amplitudes. As part of the process, the to-be-deblended shot gathers are aligned by the blending noise. Implementation on field blended-by-acquisition data demonstrates that introducing the suggested data conditioning steps can considerably reduce the leakage of primary-source events in the deep part of the blended section. The complete proposed approach performs almost as well as a conventional algorithm in the shallow section and shows a great advantage in efficiency. It performs slightly worse for larger traveltimes, but still removes the blending noise efficiently.

Deep learning has shown a considerable potential to significantly improve processing efficiency but has not yet been widely deployed to production projects of seismic signal separation such as seismic interference attenuation. The main reasons are: First, the industry has high standards for signal fidelity, which are critical for the success of subsequent seismic imaging, and deep neural network methods have not yet matched the required level; second, the network's interpretability issue has affected many geophysicists and sponsors’ trust in the deep learning technique. To develop deep neural network methods towards the end of benefiting real-world production, we first attempt to better understand their performance, especially in how they make use of local and global features of the data. A novel quantitative research of the overall network model behaviour on synthetic data is conducted. We simulate three types of coherent seismic data components in the shot domain, blend them together and then train a network to separate them. In this process, random noise, a component having only learnable local features, is selectively injected into the network's training pairs. Three network models sharing the same architecture are trained individually, and they show distinctive behaviours when applied to the same test data. Step-by-step analysis of each of them reveals that training the network with additional random noise injected into both the input and the output channel of the desired signal can lead to a decent prediction of the coherent noise based on good learning of the global features and, in the meantime, preserve almost all the data information from being lost. We propose this key lesson we learnt as a new method to improve the network's signal fidelity for shot-domain seismic interference attenuation, which is essentially a signal separation task. Its effectiveness is demonstrated on field data from Africa with a comparison to a conventional physics-based seismic interference attenuation method used in production.

Considering the 3D propagation characteristics of seismic waves, theoretically, 3D surface-related multiples elimination (3D SRME) can suppress multiples with high accuracy. However, 3D SRME has strict requirements for acquisition geometry, which makes it difficult to be implemented in practice. In the process of 3D SRME, the multiple contribution gather (MCG) is a collection of wavefields with different propagation paths. The accuracy of the multiple propagation paths in the MCGs can be directly characterized by the inclination of the wavefields, which can achieve the weighted superposition of the wavefields. The direct summation of the sparse MCGs in the crossline direction produces serious spatial aliasing, which can easily cause the contamination of primaries. Based on the kinematic characteristics of multiple propagation, MCGs can be considered as a set of hyperbolas with temporal and spatial characteristics. Then, the direct summation of the sparse MCGs can be transformed into a process of superposition along the hyperbolic integration paths. However, as the stable phase points of the events, the apexes of the hyperbola have different spatial distributions in complex geological structures. Such hyperbolic stacking paths are difficult to be controlled by conventional Radon transform or constrained inversion. In this paper, we modify the apex-shifted hyperbolic Radon transform (ASHRT) to implement the summation of crossline MCGs with variable stable phase points along the hyperbolic integration paths. Improved ASHRT uses local similarity to locate the position of stable phase points, which can improve the stability of the algorithm and the efficiency of the computation. The proposed method is demonstrated on a 3D synthetic data set, as well as on a 3D marine data set, effectively avoiding the spatial aliasing caused by sparse crossline MCGs and improving the accuracy of multiple suppression.

Passive seismic has recently attracted a great deal of attention because non-artificial source is used in subsurface imaging. The utilization of passive source is low cost compared with artificial-source exploration. In general, constructing virtual shot gathers by using cross-correlation is a preliminary step in passive seismic data processing, which provides the basis for applying conventional seismic processing methods. However, the subsurface structure is not uniformly illuminated by passive sources, which leads to that the ray path of passive seismic does not fit the hyperbolic hypothesis. Thereby, travel time is incorrect in the virtual shot gathers. Besides, the cross-correlation results are contaminated by incoherent noise since the passive sources are always natural. Such noise is kinematically similar to seismic events and challenging to be attenuated, which will inevitably reduce the accuracy in the subsequent process. Although primary estimation for transient-source seismic data has already been proposed, it is not feasible to noise-source seismic data due to the incoherent noise. To overcome the above problems, we proposed to combine focal transform and local similarity into a highly integrated operator and then added it into the closed-loop surface-related multiple elimination based on the 3D L1-norm sparse inversion framework. Results proved that the method was capable of reliably estimating noise-free primaries and correcting travel time at far offsets for a foresaid virtual shot gathers in a simultaneous closed-loop inversion manner.

For economic and efficiency reasons, blended acquisition of seismic data is becoming increasingly commonplace. Seismic deblending methods are computationally demanding and normally consist of multiple processing steps. Furthermore, the process of selecting parameters is not always trivial. Machine-learning-based processing has the potential to significantly reduce processing time and to change the way seismic deblending is carried out. We have developed a data-driven deep-learning-based method for fast and efficient seismic deblending. The blended data are sorted from the common-source to the common-channel domain to transform the character of the blending noise from coherent events to incoherent contributions. A convolutional neural network is designed according to the special characteristics of seismic data and performs deblending with results comparable to those obtained with conventional industry deblending algorithms. To ensure authenticity, the blending was performed numerically and only field seismic data were used, including more than 20,000 training examples. After training and validating the network, seismic deblending can be performed in near real time. Experiments also indicate that the initial signal-to-noise ratio is the major factor controlling the quality of the final deblended result. The network is also demonstrated to be robust and adaptive by using the trained model to first deblend a new data set from a different geologic area with a slightly different delay time setting and second to deblend shots with blending noise in the top part of the record.

Marine seismic interference noise occurs when energy from nearby marine seismic source vessels is recorded during a seismic survey. Such noise tends to be well preserved over large distances and causes coherent artefacts in the recorded data. Over the years, the industry has developed various denoising techniques for seismic interference removal, but although well performing, they are still time-consuming in use. Machine-learning-based processing represents an alternative approach, which may significantly improve the computational efficiency. In the case of conventional images, autoencoders are frequently employed for denoising purposes. However, due to the special characteristics of seismic data as well as the noise, autoencoders failed in the case of marine seismic interference noise. We, therefore, propose the use of a customized U-Net design with element-wise summation as part of the skip-connection blocks to handle the vanishing gradient problem and to ensure information fusion between high- and low-level features. To secure a realistic study, only seismic field data were employed, including 25,000 training examples. The customized U-Net was found to perform well, leaving only minor residuals, except for the case when seismic interference noise comes from the side. We further demonstrate that such noise can be treated by slightly increasing the depth of our network. Although our customized U-Net does not outperform a standard commercial algorithm in quality, it can (after proper training) read and process one single shot gather in approximately 0.02 s. This is significantly faster than any existing industry denoising algorithm. In addition, the proposed network processes shot gathers in a sequential order, which is an advantage compared with industry algorithms that typically require a multi-shot input to break the coherency of the noise.

Processing marine seismic data is computationally demanding and consists of multiple time-consuming steps. Neural network based processing can, in theory, significantly reduce processing time and has the potential to change the way seismic processing is done. In this paper we are using deep convolutional neural networks (CNNs) to remove seismic interference noise and to deblend seismic data. To train such networks, a significant amount of computational memory is needed since a single shot gather consists of more than 10data samples. Preliminary results are promising both for denoising and deblending. However, we also observed that the results are affected by the signal-to-noise ratio (SnR). Moving to common channel domain is a way of breaking the coherency of the noise while also reducing the input volume size. This makes it easier for the network to distinguish between signal and noise. It also increases the efficiency of the GPU memory usage by enabling better utilization of multi core processing. Deblending in common channel domain with the use of a CNN yields relatively good results and is an improvement compared to shot domain.