S U M M A R YTeleseismic surface wave data from the dense seismographic network in Southern California (CISN, California Integrated Seismic Network) are used to construct S-wave velocity maps for the region. Surface wave phase velocity measurements are achieved by way of a waveformmatching spectral-domain technique that was developed for use within the CISN array. A two-station phase velocity measurement technique was developed, which takes advantage of the fact that, for surface wave arrivals from any azimuth, there are multiple pairs of CISN stations, which are almost exactly on the same great-circle paths. To ensure the validity of this assumption, particle motions were analysed for measured pairs within the network, showing that the deviation between expected backazimuth from ray theory and observed backazimuths are small enough to be analysed by a simple distance correction. This method was applied to 114 large earthquakes for time periods between 1999 and 2002 and led to the retrieval of phase velocity maps for Rayleigh waves between 20 and 55 mHz and for Love waves between 25 and 45 mHz.Finite frequency effect for propagating surface waves was approximately taken into consideration using a geometrical technique. An ellipse, determined by wavelength and great-circle distance between the source and receiver can describe an influence zone for the finite frequency effect. Phase velocity map reconstructions, using a method which incorporates this effect, show minimal changes in phase velocity maps for our region, but do extend solvable velocity regions.Inversion for S-wave velocity distribution with depth shows that deep structural contrast across the San Andreas extends into the upper mantle. A slow velocity anomaly is evident under the Southern Sierras and the Salton Sea, the latter region indicating the northward extension of tectonic activity in the Gulf of California. Fast velocity roots that reach 60 km are found under the western part of Transverse Ranges and the Northern Peninsular Ranges. Reconstruction of past location of the Western Transverse Range microplate implies that these two ranges may have been adjacent to one another historically. Both the Western Transverse range fast velocity root and the Peninsular range fast velocity root may be a remnant of old oceanic plate.
It is well known that off-great-circle path propagation causes a technical difficulty for surface wave analysis in higher frequency ranges. We propose a new approach that combines a beamforming technique and two-station phase velocity measurement to resolve this problem. Beamforming allows us to determine the correct azimuth of incoming surface waves which can be taken into account in phase velocity measurement. Beamforming results also support that a plane-wave approximation is mostly acceptable for frequencies up to about 50-60 mHz (millihertz), although evidence of multipathing is occasionally recognized in beamforming results as multiple peaks. Application of this correction scheme for Rayleigh-wave data in Southern California seems to make the largest impact on the results of azimuthal anisotropy. Effects are not large for frequencies up to 30 mHz but fast velocity axes in azimuthal anisotropy maps change significantly for higher frequencies.
Machine learning has been around for decades or, depending on your view, centuries. To consider the tools and underpinnings of machine learning, one would need to go back to the work of Bayes and Laplace, the derivation of least squares, and Markov chains, all of which form the basis and the probability construct used pervasively in machine learning. There has been a flood of progress between 1950 (with Alan Turing's proposal of a learning machine) and early 2000 (with practical applications of deep learning in place and more recent advances such as AlexNet in 2012). Deep learning has demonstrated tremendous success in a variety of application domains in the past few years, and with some new modalities of applications, it continues to open new opportunities. The recent popularity and emergence of machine learning in the oil and gas industry is likely due to the abundance of unused or overlooked data and the economic need to extract additional information from the data currently used. Additionally, there is an unprecedented availability of computing power, easy-to-use coding libraries, and application programming interfaces, as well as recent and significant advances in various flavors of neural networks. In this paper, we will attempt to show how machine learning can assist geoscientists in performing routine tasks in a much shorter time frame. We assert that there is a great opportunity for geoscientists to learn from machines, use these techniques to quality check their work, and gain nuanced insights from their data. Another advantage is that these approaches lead to the optimization of machine learning workflows by providing more accurate training data sets thus driving continuous learning and enhancement of the model.
Full-waveform inversion (FWI) has the potential to be a game changer for the seismic industry since it produces accurate velocity models at a resolution that cannot be matched by conventional traveltime tomography, with the additional advantage that it runs on basically raw data. Today, these two features should be better exploited to maximize the business value that FWI can deliver to an oil and gas operator. Provided that we acquire adequate data sets, FWI can be run starting from a smooth velocity model so that traveltime tomography is completely bypassed. As very little processing is required, FWI can be executed in parallel with the main preprocessing efforts, with the results of shortening considerably the overall project turnaround. For imaging purposes, FWI can be limited to the low-frequency range (∼ 10 Hz) but if run in the high-frequency range (∼ 30 Hz), FWI velocities have such a resolution that they can be used directly for interpretation. In particular, applications in the field of pore-pressure prediction and shallow hazard have great potential. In the current implementation, the main limitations are linked to data quality and computational power, but both these issues can be adequately addressed.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.