Ground roll extraction using the Karhunen-Loeve transform in the time-frequency domain

Ground roll suppression is critical for seismic reflection data processing. Many standard methods, i.e., FK filtering, fail when spatially aliased surface wave interference is present in the data. Spatial aliasing is a common problem; receiver spacing is often not dense enough to extract wavenumbers of low-velocity surface waves. It has long been known that the Karhunen-Loeve transform can be used to suppress aliased ground roll. However, the ground roll should be flattened before suppression, which is challenging due to the dispersion of surface wave velocities. I propose to solve this problem via the time-frequency domain. I apply the S-transform, which was previously shown to perform well in the multichannel analysis of surface waves. A simple complex-valued constant phase shift is a suitable model of surface wave propagation in common-frequency S-transform gathers. Therefore, it is easy to flatten the corresponding S-transform narrow-band frequency surface wave packet and extract it from the data by principal component analysis of the corresponding complex-valued data-covariance matrix. As the result, the proposed S-transform Karhunen-Loeve (SKL) method filters the aliased ground roll without damaging the reflection amplitudes. The advantages of SKL filtering have been confirmed by synthetic- and field-data processing.

Imaging Complex Subsurface Structures for Geothermal Exploration at Pirouette Mountain and Eleven-Mile Canyon in Nevada

Accurate imaging of subsurface complex structures with faults is crucial for geothermal exploration because faults are generally the primary conduit of hydrothermal flow. It is very challenging to image geothermal exploration areas because of complex geologic structures with various faults and noisy surface seismic data with strong and coherent ground-roll noise. In addition, fracture zones and most geologic formations behave as anisotropic media for seismic-wave propagation. Properly suppressing ground-roll noise and accounting for subsurface anisotropic properties are essential for high-resolution imaging of subsurface structures and faults for geothermal exploration. We develop a novel wavenumber-adaptive bandpass filter to suppress the ground-roll noise without affecting useful seismic signals. This filter adaptively exploits both characteristics of the lower frequency and the smaller velocity of the ground-roll noise than those of the signals. Consequently, this filter can effectively differentiate the ground-roll noise from the signal. We use our novel filter to attenuate the ground-roll noise in seismic data along five survey lines acquired by the U.S. Navy Geothermal Program Office at Pirouette Mountain and Eleven-Mile Canyon in Nevada, United States. We then apply our novel anisotropic least-squares reverse-time migration algorithm to the resulting data for imaging subsurface structures at the Pirouette Mountain and Eleven-Mile Canyon geothermal exploration areas. The migration method employs an efficient implicit wavefield-separation scheme to reduce image artifacts and improve the image quality. Our results demonstrate that our wavenumber-adaptive bandpass filtering method successfully suppresses the strong and coherent ground-roll noise in the land seismic data, and our anisotropic least-squares reverse-time migration produces high-resolution subsurface images of Pirouette Mountain and Eleven-Mile Canyon, facilitating accurate fault interpretation for geothermal exploration.

Physics-constrained deep learning for ground roll attenuation

We propose a method to combine unsupervised and supervised deep learning approaches for seismic ground roll attenuation. The method consists of three components that have physical meaning and motivation. The first component is a convolutional neural network to separate a seismic record into ground roll and signal, while minimizing the residual between the sum of the generated signal and ground roll from two sub-networks and the input seismic record. The second component creates a maximum separation of signal and ground roll in the FK domain, by training a supervised classifier. The third component is a convolutional neural network mapping signal to ground roll, which overcomes the problem of finding appropriate masks in traditional methods. Each component in our method is closely related to and motivated by the wave characteristics of the ground roll. Test results on field seismic records demonstrate the effectiveness of combining these components in preventing signal leakage and removing ground roll from seismic data.

Simulation of scattered seismic surface waves on mountainous onshore areas: Understanding the “ground roll energy cone”

The accurate simulation of seismic surface waves on complex land areas requires elastic models with realistic near-surface parameters. The SEAM Phase II Foothills model, proposed by the SEG Advanced Modeling (SEAM) Corporation, is one of the most comprehensive efforts undertaken by the geophysics community to understand complex seismic wave propagation in foothills areas. However, while this model includes a rough topography, alluvial sediments, and complex geologic structures, synthetic data from the SEAM consortium do not reproduce the qualitative characteristics of the scattering energy that is generally interpreted as the “ground roll energy cone” on shot records of real data. To simulate the scattering, a near-surface elastic model in mountainous areas ideally must include the following three elements: (1) rough topography and bedrock, (2) low-velocity layer, and (3) small-scale heterogeneities (size approximately Rayleigh wavelength). The SEAM Foothills model only includes element (1) and, to a lesser extent, element (2). We represent a heterogeneous near surface as a random medium with two parameters: the average size of the heterogeneities and fractional fluctuation. A parametric analysis shows the influence of each parameter on the synthetic data and how similar it is compared to real data acquired in a foothills area in Colombia. We perform the analysis in the shot gather panel and dispersion image. Our study shows that it is necessary to include the low-velocity layer and small-scale distributed heterogeneities in the shallow part of the SEAM model to get synthetic data with realistic scattered surface-wave energy.

Seismic structure of the Cheb Basin from high resolution surveying – data quality assessment and first results

<p><span><span>In this study we present data and preliminary results from several shallow high-resolution seismic surveys in the Cheb Basin, CR, a small intracontinental basin in the North-West Bohemian Massif, located at the Western end of the Cenozoic Eger Rift. The area is well known for its intense earthquake activity, with the largest instrumentally recorded magnitude of M</span></span><span><sub><span>L</span></sub></span><span><span>=4.6. Macroseismic reports of local seismicity date back to the early 19</span></span><span><sup><span>th</span></sup></span><span><span> century, with magnitudes possibly above 5. Quaternary volcanoes, CO</span></span><span><sub><span>2</span></sub></span><span><span>-rich moffettes, and the swarm-like occurrence of the earthquakes suggest they are being triggered by crustal fluids. In contrast, most focal mechanisms show a dominant strike-slip component, indicative of tectonics. Investigating the role of fluids in triggering those earthquakes is one of the objectives of an ongoing ICDP program.</span></span></p><p><span>We expect high-resolution images of the basin structure to provide additional constraints regarding the importance of tectonic faulting. To that end, we surveyed several up to 3-km-long reflection and refraction profiles in the basin center across the putative Počátky-Plesná Fault, and at its edge, across the basin-bounding Mariánské Lázně Fault. The up to 350-m-thick basin sediments are mostly of Miocene and Quaternary origin, overlying Paleozoic Variscan units and post-Variscan granites. The main reflectors are around 200-400 ms. The data were collected with a 500-m-long split-spread of single geophones at 2 m spacing, and the raw shots are dominated by ground roll. In this presentation, we will show an overview of the field campaigns and present first results.</span></p>