Previous research highlights the importance of thoroughly characterizing burrow attributes [e.g., burrow percentage (BP), morphology, and diameter] to understand their effects on petrophysical properties. While traditional samples, such as hand specimens and core plugs, and outcrops aid in detailed characterization, their limitations are evident when analyzing large 3-D that can exceed 8 cm in diameter. This study introduces a new approach using ground penetrating radar (GPR) to capture these attributes, a technique not previously utilized in such research. We focus on the Upper Cretaceous Aruma Formation in central Saudi Arabia, characterized by extensive burrow networks with BPs reaching up to 40%. Using a 2.7-GHz antenna over a 1-m2 area, we aimed to assess burrow attributes with GPR, comparing the results to field observations and computed tomography (CT) scans of two samples for a thorough analysis and to evaluate GPR’s effectiveness against field and CT scan methods. This comparison highlighted GPR’s ability to detect complex burrow networks, revealing significant variations in burrow attributes that are not fully captured by other methods. Specifically, GPR data showed that the estimated BP deviated by 5% from CT scans and by 9% from field observations, attributed to its more extensive volume of coverage and deeper penetration. Further analysis confirms GPR’s precision in measuring burrow diameters ranging from 0.45 to 8.5 cm and its effectiveness in providing detailed descriptions of burrow morphology. GPR offers unique insights into large burrow systems, providing data hard to obtain otherwise and enabling a deeper understanding of their characteristics.

The first arrival time of seismic waves is a crucial parameter for seismic data analysis, which is used to determine the depth and location of subsurface structures. However, the estimation of first arrival time is often challenging due to the presence of noise and uncertainties in the data. In this study, we propose a novel approach that utilizes the Kalman filter with generalized likelihood ratio (GLR) to estimate the first arrival time in seismic data. First, a discrete time linear system with unknown amplitudes changes occurring at unknown time instants is used to reformulate the convolutional model. Then, to provide residual signals, we apply a Kalman filter based on the no change hypothesis to the linear system. Finally, to estimate the first arrival time, a generalized likelihood ratio (GLR)-based change detection technique is utilized. Using the simulated data, we verified the performance of the proposed approach. Overall, this study presents a promising approach for improving the accuracy of first arrival picking in the presence of different noise levels.

Signal attenuation estimation is a critical task in signal processing and is essential for analyzing media characteristics and compensating for energy loss. Current centroid frequency shift-based methods for estimating attenuation are mostly based on the assumptions of full-band analysis and frequency-dependent or independent quality factor (Q). In this paper, we propose a novel frequency-independent centroid frequency shift (FiCFS) method for signal attenuation estimation with higher adaptability. It is based on arbitrary frequency bands instead of the full-band spectrum defined by the conventional centroid frequency shift (CFS) method, and accordingly, the derivation is performed by incomplete gamma functions instead of ordinary gamma functions. Through rigorous mathematical derivations, the first moment (centroid frequency) and the second moment (variance) are proved to be frequency insensitive for arbitrary frequency bands, and then the arbitrary frequency band-based CFS method, i.e., the FiCFS method, is derived with the frequency-weighted exponential spectrum assumption. The matching ability of the frequency-weighted exponential spectrum to other signal spectra is verified, demonstrating the method’s adaptability to most attenuated signals. Experimental results using synthetic and field data sets demonstrate that the proposed method is adaptive, noise-immune, and reliable.

A physics-based approach called waveform parsimonious refraction interferometry (WPRI), which interpolates head or refraction waveform, is introduced in this work. WPRI yields a significant advantage in exploration and engineering applications, as it mitigates the excessive time and labor cost in 2-D field acquisitions that require dense receivers and shots while improving coverage in refraction seismic imaging processes. Our proposed method generates the virtual seismic refraction waveform, which involves kinematic and dynamic information with near-perfect accuracy for near-surface seismic waveform inversion and migration. To achieve this, we record data from two shot locations situated at opposite ends of the seismic profile, as well as a handful of near-offset traces along the profile. The virtual head or refraction wavefield is then determined through the convolution and cross correlation of the recorded wavefields with the objective of eliminating common wavepath. Furthermore, we introduce a source wavelet deconvolution step to correct dynamic discrepancies present in the virtual waveform. By implementing this type of technique, we are able to produce virtual seismic data that are highly accurate and can be effectively employed in near-surface seismic imaging applications.

There has been a recent rise in the uses and applications of passive seismic data, such as tomographic imaging, volcanic monitoring, and hydrocarbon exploration. Consequently, the sharp increase in passive seismic applications requires real-time event detection capabilities with high accuracy. Proper analysis of such events depends largely on the signal-to-noise ratio improvement through noise suppression techniques. Recent advances in the fields of signal processing and deep learning coupled with the available computational resources provide a great opportunity to address this challenge. In this work, a workflow is proposed where a residual deep neural network is customized and employed to detect passive seismic events. The automated detection is followed by a denoising step to extract the signal of interest from background noise using an IIR Wiener filter. The proposed method does not require any prior knowledge of the signal/noise, and therefore, it can work with various types of signals/noises. Another benefit of the proposed detection method is that the deep neural network is trained only on synthetic seismic data without the need to use real data in the training process. Nevertheless, it exhibits high accuracy in detecting and denoising events from real passive seismic data sets. In particular, field seismic data is recorded in northern Saudi Arabia and used to test the complete detection and denoising method. The detection method proved its capability of detecting events automatically in large data sets and in real time (due to off-line training).

Recorded seismograms are usually distorted by statics owing to complex geological conditions, such as lateral variations in sediment thickness or complex topographies. These distorted and discontinuous signals usually exist in either arrival times or amplitudes of waves, and they are most likely to be smeared as velocity perturbations along their associated raypaths. Therefore, statics may blur images of the target bodies or, even worse, introduce unexpected and false anomalies into subsurface structures. To partly resolve this problem, we develop a weighted statics correction method to estimate unwanted temporal shifts of traces using the closure-phase technique, which is utilized in astronomical imaging. In the proposed method, the source and receiver statics are regarded as independent quantities contributing to the waveform shifts based on their acquisition geometries. Numerical tests on both the synthetic and field cases show noticeable, although gradual, improvements in data quality compared to the conventional plus–minus (PM) method. In general, this method provides a straightforward strategy to reedit the travel times in seismic profiles without inverting for a near-surface velocity model. Moreover, it can be extended to any interferometrical methods in seismic data processing that satisfies the closure-phase conditions.

We present a method that skeletonizes the first arriving seismic refractions by machine learning and inverts them for the subsurface velocity model. In this study, first arrivals can be compressed in a low-rank sense with their skeletal features extracted by a well-trained autoencoder. Empirical experiments suggest that the autoencoder’s$1\times 1$or$2\times 1$latent vectors vary continuously with respect to the input seismic data. It is, therefore, reasonable to introduce a misfit functional measuring the discrepancies between the predicted and the observed data in a low-dimensional latent space. The benefit of this approach is that an elaborated autoencoding neural network not only refines intrinsic information hidden in the refractions but also improves the quality of inversion for a reliable background velocity model. Numerical tests on both synthetic and field data demonstrate the effectiveness of this method, especially in recovering the low-to-intermediate wavenumber parts of the subsurface velocity distribution. Comparisons are made with the other three relevant methods, the wave-equation travel-time (WT) inversion, the envelope inversion, and the full waveform inversion (FWI). As expected, the cycle skipping problem is alleviated due to the reduction of dimensions of data space. This method outperforms the envelope inversion in resolution, and it is no worse than WT. Moreover, there is no need for careful manual travel-time picking with this methodology. In general, this inversion framework provides an extendable strategy to compress any input data for reconstructing high-dimensional physical parameters.