Heidi Anderson Kuzma

Slip rate of the western Garlock fault, at Clark Wash, near Lone Tree Canyon, Mojave Desert, California

The precise tectonic role of the left-lateral Garlock fault in southern California has been controversial. Three proposed tectonic models yield significantly different predictions for the slip rate, history, orientation, and total bedrock offset as a function of distance along strike. In an effort to test these models, we present the first slip-rate estimate for the western Garlock fault that is constrained by radiocarbon dating. A channel (referred to here as Clark Wash) incised into a Latest Pleistocene alluvial fan has been left-laterally offset at least 66 ± 6 m and no more than 100 m across the western Garlock fault, indicating a left-lateral slip rate of 7.6 mm/yr (95% confidence interval of 5.3–10.7 mm/yr) using dendrochronologically calibrated radiocarbon dates. The timing of aggradational events on the Clark Wash fan corresponds closely to what has been documented elsewhere in the Mojave Desert, suggesting that much of this activity has been climatically controlled. The range-front fault, located a few hundred meters northwest of the Gar-lock fault, has probably acted primarily as a normal fault, with a Holocene rate of dip-slip of 0.4–0.7 mm/yr. The record of prehistoric earthquakes on the Garlock fault at this site, though quite possibly incomplete, suggests a longer interseismic interval (1200–2700 yr) for the western Garlock fault than for the central Garlock fault. The relatively high slip rate determined here indicates that the western and central segments of the Garlock fault show similar rates of movement that are somewhat faster than rates inferred from geodetic data. The high rate of motion on the western Garlock fault is most consistent with a model in which the western Garlock fault acts as a conjugate shear to the San Andreas fault. Other mechanisms, involving extension north of the Garlock fault and block rotation at the eastern end of the fault may be relevant to the central and eastern sections of the fault, but they cannot explain a high rate of slip on the western Garlock fault.

Particle Swarm Optimization (PSO): a simple and powerful algorithm family for geophysical inversion

PSO is an optimization technique that has been successfully used in many different engineering fields. The PSO algorithm can be physically interpreted as a stochastic damped mass-spring system. This analogy served us to introduce the PSO continuous model and to deduce a whole family of PSO algorithms arising from the discretization of the PSO continuous model. We also analyze their respective first order and second order stability regions from the stochastic point of view. Their performance is also checked using synthetic functions showing a degree of illposedness similar to that found in many geophysical inverse problems. Finally we present the application of these algorithms to the analysis and solution of a VES inverse problem associated to a seawater intrusion in a coastal aquifer in South Spain.

Non‐linear avo inversion using support vector machines

Model/data pairs, computed using known geophysical forward relationships, can be used to train Support Vector Machines to approximate inverse relationships. Given appropriate training data, a Support Vector Machine statistically reproduces the results of non-linear inversion. Inversion of the non-linear Zoeppritz equations is an illposed problem. The results of an inversion depend heavily on the a priori assumptions used to regularize it. If an SVM is trained using data that contains the same assumptions, it will get the same answer. It captures nonlinear relationships as easily as linear ones by employing a kernel function. A naive SVM can usually be trained using fewer calculations of a forward model than are necessary in an equally naive inversion. Since an SVM only needs to be trained once whereas an inversion needs to be repeated on each new data set, procedures which require multiple inversions of the same data become available. The Jackknife method uses multiple inversions to calculate error bars and improve a model estimate.

A support vector machine for AVO interpretation

A Support Vector Machine (SVM) is used to approximate an inverse to the Zoeppritz equations. It is tested on an AVO data set from the Gulf of Mexico which includes a gas hydrate layer. Physically realistic examples of velocity and density contrasts from the region are used to train the SVM. They are also used to construct a test set. Parameters of the SVM, including the kernel and regularization constant, are selected based on their performance on the test set. The SVM does a better job at interpreting the Gulf data does inversion of standard linearized Zoeppritz equations. Its performance against careful inversion of the full Zoeppritz equations has yet to be determined.

Polynomial Chaos for Uncertainty Quantification in Geophysics

Summary Uncertainty Quantification using Polynomial Chaos (PC) expansions is a method for determining the sensitivity of the output from any type of modeling program to uncertainty in its input parameters. The intuition behind the technique is similar to the intuition of Fourier transformation which is familiar to most geophysicists. There are two ways to explore the effect of input parameters on output simulations; the first is to study the PC coefficients themselves which is akin to looking at a Fourier spectrum. The second is to reconstruct the function after enhancing or removing coefficients which is similar to frequency domain filtering. As an example, a PC expansion is used to explore the effect of uncertainty in Vp, Vs velocity and density as inputs into the full Zoeppritz

Support vector machines implemented on a graphics processing unit

Graphics Processing Units (GPUs) are inexpensive, commercial vector processors developed as engines for high-speed computer games. By substituting geophysical data for color values in arrays that are designed to compute images, GPUs can be used as parallel scientific processors. Support Vector Machines (SVMs) are computer learning algorithms that can be used to emulate non-linear geophysical inversion. SVMs depend on making multiple comparisons between examples of input data. These comparisons can be made in parallel on the GPU, significantly speeding up the algorithm. GPUs only support single precision, however, and are limited by their memory size. This poster presents results of a preliminary study showing that speedups of as great as 250x can be achieved using a GPU to make the comparisons necessary to train an SVM with a nearly 3000 example training set of input vectors with 3000 elements each. Because only the comparison part of the SVM algorithm was implemented on the GPU, this work can likely be generalized to other processes useful for Geophysics that depend on making comparisons , such as correlation and convolution.

A new fractal-interpolation algorithm for seismic data based on iterated function systems

A new fractal-interpolation method called PPA (Pointed Point Algorithm) based on the Iterated Function System (IFS) is proposed to interpolate the signals with the expected interpolation error, solving the problem that the ordinary fractal interpolation cannot get the value of any arbitrary point directly, which has not been found in the existing literature. Experiments on the theoretical data and real field seismic data show that the proposed PPA method can not only get the expected points value, but also get a great accuracy in the reconstruction of the seismic profile, leading to a significant improvement over other trace interpolation methods.

Vehicle Traffic as a Source for Near‐Surface Passive Seismic Imaging

In this paper, we present preliminary results from a field experiment in which we explored the use of cars and traffic as a source for passive seismic imaging. We set out a line of geophones at a 45 degree angle between two intersecting roads at the University of California, Richmond Field Station. We collected data sets including background noise, an idling car at the intersection of the roads, one car driving on one road and two cars driving on both roads. A freight train contributed an appreciable signal. We found that the signal propagated best in the 3-20 Hz range. Average power spectra at each of the geophones showed a constructive interference structure which we had seen in simple synthetic models, computed by modeling the road as a line source. We were able to use the data from the idling car to estimate a dispersion curve.

Monitoring Methane Hydrate Production In the Arctic; a Preliminary Feasibility Study

In this paper we present preliminary results based on several synthetic experiments in which we study to determine if the shape, size and saturation of similar gas lenses from arctic methane hydrates, which can be monitored throughout production using repeated seismic survey either at the surface or in a borehole. We use saturation models presented in the first part of this paper and work by Kowalsky et al. (2010) who showed that timelapse Vertical Seismic Profiling (VSP) is a possible tool for monitoring production of marine hydrates.