CWP Seminar - Spring, 2008

Date

Speaker

Subject

Summary

1-16

Milana

Effective reflection coefficients at a bending contact of isotropic and VTI media

I will introduce effective reflection coefficients, which generalize the plane-wave reflection coefficients for finite frequencies and arbitrary interface geometry. In addition, I will present the results of 3D Kirchhoff-type modeling with the effective reflection coefficients, and show that they produce smoother amplitudes, head waves and no artificial diffractions.

Alexandre

Synthetic Models for the Virtual Source Method

The virtual source method is a groundbreaking technique that aims to image below complex topography and perturbed overburden without the knowledge of the overburden velocities and structures. In this presentation, the virtual source method will be tested on different synthetic models on which regular seismic methods give no satisfactory results (salt domes, foothills, perturbed overburdens...) to see how well the method works and under which assumptions and conditions.

1-21

Weitao

A simple introduction to double difference earthquake location algorithm -method and application

Earthquake hypocenter location is very important for the research of seismicity analysis of tectonic processes ,earthquake recurrence and earthquake interaction,etc.I happened to read a paper on the double difference algorithm for precise hypercenter relocation,and will give a introduction of its basic method and application.

Yong Z.

Seismological Records of China Seismic Network and earthquake locate method

Earthquake locating is a classic problem in Seismology. After more than 100 years endeavors, there are still some problems to be solved. In this talk I give some introdution of Chinese seismic records, and give a general explanation of traditional locate method. Then discuss how to improve the precision of locating.

Paul

Micro-earthquake location with sparse data

I return to the issue of imaging with sparse data and illustrate it with a new example. As suggested in the preceding seminar, the interferometric imaging condition serves to attenuate random wavefield fluctuations. Those random wavefield fluctuations can be caused by unknown (random) velocity fluctuations or by sparse sampling of the recorded data.

1-23

Fan

Does the heterogeneity help reducing requied source number in seismic interferometry?

In seismic interferometry, people beleive that heterogeneities in the media help to equilibrize the energy to all directions. Therefore, less sources are needed to extract the wave propergating between two receivers for heterogeneous media than that for homogeneous ones. But is it really a valid statement or are there any other assumptions needed in this statement? We will discuss this in today's talk.

Jae

Brief Introduction about Resistivity Modeling

Resistivity modeling is a kind many geophysical modeling, which is, by nature, effective for geotechnical/environmental application. I give brief introduction about resistivity forward and inversion modeling, and show the effectiveness and limit of the tool with several model test results and a field data interpretation result. Finally, I share some points that I felt we have to be aware of when we encounter any inversion problem.

Ilya

Geometrical seismics for shear waves in homogeneous TI media

It is well known that the geometrical-seismics approximation (and, therefore, dynamic ray tracing) breaks down near low-amplitude directions, such as minima of radiation patterns. By analyzing the Green's function for SV- and SH-waves in transversely isotropic (TI) media, I show that the influence of anisotropy can make the geometrical seismics inadequate even for high-amplitude segments of far-field wavefronts. In particular, the shear wavefield near the symmetry axis should be described using either exact solutions or higher-order ray theory.

1-28

Rosie

Introduction to Monitoring the Subsurface with Multiple Scattered Waves and Quasi-Static Deformation

We are exploring a new technology that is based on using low-frequency strain data to monitor changes in fluid saturation conditions in porous media. This is a brief introduction to the research what I've worked on since this semester. Some of the results will be showed this time, however, I'll update the problems and the results in the following seminars.

Derek

Flattening

In this seminar, I will revisit the flattening algorithm that I talked about last semester. I will quickly review the algorithm. Then, introduce three possible ways to weight the inversion and discus the results of each.

Dave

Apparent horizontal displacements

We know that apparent vertical displacements in time-lapse seismic images are caused mostly by a decrease in seismic wave velocities above compacting reservoirs. If we suppose that the same is true for horizontal displacements, then image ray theory should enable us to compute horizontal displacements from measured vertical displacements. Horizontal displacements computed in this way are indeed comparable to measured horizontal displacements

1-30

Yong M.

Interferometric imaging condition in the frequency domain

Paul has applied interferometric imaging condition for wave-equation migration in the time domain. I will talk about the possibility of using this technique in the frequency domain.

Jyoti

Group attenuation

Interval attenuation obtained from the layer-stripping approach introduced earlier, yields attenuation values that are consistent with the interval phase attenuation and seem to be independent of the inhomogeneity angle. Through numerical modeling and linearized analytic expressions I will show that the spectral ratio method yields phase attenuation coefficients for zero inhogeneity angle.

Eduardo

GeBR: an open-source interface for seismic processing

GeBR is a graphical interface built to make seismic processing easier. GeBR can be used to easily perform the following tasks: management of projects and jobs, creation and execution of processing flows. The GeBR code is written in C language. It uses GTK+ library for graphics and XML tools for information manipulation.

2-4

John M.

3D steerable filters

Last year I was working on the steerable pyramid in two dimensions, with the eventual goal of using it for delineation of channels in 3D seismic data. I am now working with 3D steerable filters. In this talk I describe how the 3D steerable pyramid works, and show how it can be used for analysis of orientation of features within a 3D image.

Gabi

SEG/EAGE salt model

So far, I have been showing the angle decomposition for simple models with one source and relatively simple velocity models. One of the goals of performing this angle decomposition is to be able to study a little about illumination of the reflector. I am going to present a synthetic model using the SEG/EAGE salt dome velocity model and show the angle decomposition for a horizontal reflector underneath the salt.

Farnoush

Study of the extracted Green's function from correlation interferometry for surface source-receiver distribution

Following the discussion from last semester I will complete all the reasons that why we can not get the body-wave Green's function from  correlation interferometry . I also will complete my synthetic examples to demonstrate these reasons. At the end I will  discuss about the stationary zones for different models via correlation gathers.

2-6

Jia

Anisotropic wavefield separators

I talk about what opertators can be used to separate anisotropic wavefields. The basic idea is to obtain the P and S wave mode polarization vectors and transform them back to space domain and use these space domain operators to separate wave modes.

Xiaoxiang

Interval parameter estimation using velocity-independent layer stripping: Case study

We applied the velocity-independent layer stripping to the Rulison Field data which was used by Xiaoxia to make AVO and NMO ellipse analysis. The results of interval nonhyperbolic moveout parameters by the velocity-independent layer stripping and the conventional Dix-type layer stripping will be introduced and compared. :

Yongxia

Amplitude in anisotropic media with Gaussian beam method

Because the arthors did not include an amplitude factor necessary for transforming their migrations into inversions or true amplitude migrations while they evaluate the complex traveltime correctly with Gaussian beam method. Our research try to study the amplitude in anisotropic media with the Gaussian beam method. I will introduce the basic information of Gaussian beam method in anisotropic media this time.

2-11

Math/CS Candidate

Zizhong (Jeffrey) Chen

Scalable Techniques for Fault Tolerant High Performance Computing

Chauvenet Hall, Room 143, 4:00 pm

2-13

Steve

Fracture Mechanisms and Seismic Velocity Changes Around a Heated Tunnel

Seismic data from Yucca Mt. show changes in wavespeed as a function of temperature. I will briefly discuss local stress changes, super-heating, and fluid phase transitions that cause these changes and show corresponding laboratory evidence. Inferences about rock mechanics as a function of time/temperature may be made via these data.

Eduardo

GeBR - Demonstration

GeBr is an open-source interface for seismic processing. In this seminar, I'm demonstrating how GeBR integrate RSF and SU. Also, I will show how to include programs in GeBR.

Tony

Extended imaging condition for migration velocity analysis

By introducing space-lag and time-lag into conventional imaging condition, we can easily obtain the common-image gathers and their angle decomposition. This would provide us the information about the migration velocity we used. I will use a very simple experiment to show some interesting results.

2-14 Thursday

Panos Kelamis, Saudi Aramco

Special Seminar: Land Multiple Elimination – What's New?

3-4 pm, GC263

2-18

President's Day

No Talks

2-20

Ran X.

Micro earthquake imaging

Using FD to calculate wavefield generated by single point source and RTM to locate point source location. When t=0 is unknown, setting the position with maximum amplitude is source position and corresponding time as onset time. Calculating the error of this source location method.

Alexandre

Synthetic models for interferometry

Following the first discussion this semester about synthetic models applied to interferometry, three points will be discussed: the timeshift observed on data constructed by interferometry, the comparison of results obtained by acoustic and elastic models and finally the use of the modeling software Stradivarius.

Milana

Subject 3

Summary 3

2-25

Ilya

Geometrical seismics for S-waves in TI media (II)

I will continue the discussion of problems with applying geometrical seismics to shear-wave propagation in transversely isotropic (TI) media. Since the geometrical-seismics approximation is based on plane-wave solutions, it breaks down in areas of rapid spatial variations of plane-wave polarization, such as the vicinity of the symmetry axis. Accurate description of the polarization and amplitude of SV- and SH-waves in near-symmetry directions requires using either exact modeling methods or higher-order ray theory.

Jae

Integration of Seismic and CSEM data with its uncertainty analysis

I will briefly review the hierarchical Bayesian model which utilizes rock-physics model for the interconnection between seismic and CSEM data. The Metropolis-Hastings algorithm that is designed to analyze a posterior multidimensioanl probability distribution will be introduced thereafter. I will show some results of my rudimentary stochastic joint inversion which brought further questions about uncertainty analysis.

2-27

Rosie

Noise Removal

It is noise removal that I've been working on since last seminar when I gave a brief introduction on monitoring the subsurface with multiple scattered waves and Quasi-static deformation. I would like to show some results or examples I obtained from the Matlab program I wrote.

John M

Channel detection with 3D steerable pyramid

I continue to make progress with 3D steerable pyramid. In this presentation I describe some of the details of the 3D version of the pyramid, and show how the orientation of image features can be found efficiently. I finish with data examples showing enhancement of planar reflectors in a 3D seismic volume. This enhancement of planar features is the first step in our detection of channels.

Paul

Micro-earthquake location with sparse data

This is a continuation of my earlier seminar on imaging with sparse data.

3-3

Yong M.

Pseudo Wigner Distribution -- Window Effect

I will continue the topic of Wigner Distribution Function (WDF), and focus on the Pseudo WDF. In this talk, I will discuss the window effect on the common WDF, and discuss how it helps reduce the cross-term and keep the resolution.

Jyoti

Group attenuation

Continuing my talk on the influence of the inhomogeneity angle on group attenuation, I will present analytical expressions for the phase and attenuation vectors in arbitrarily anisotropic attenuative media. The group attenuation can be easily derived from these expressions.

3-5

Gabi

Relationship between space-lag and reflection angle domain

Today I am going to discuss some aspects of the relationship between the images in the space-lag domain and in the reflection angle domain.

Derek

Isosurfacing

I will present the concept of isosurfacing, specifically, the 3D Marching Cubes algorithm. I will introduce the algorithm in 2D. Then extend to 3D. I will show the output of the algorithm for several medical imaging datasets and conclude by discussing the possible application to geophysical data.

Eduardo

Paraxial ray-tracing

In this seminar I'm presenting a 3D paraxial ray-tracing code. The motivation for developing a 3D paraxial ray-tracing code comes from participation in Norman Bleistein course: Mathematics of Modeling, Migration and Inversion with Gaussian Beams. The purpose is to create a simple and flexible 3D ray-tracing code in ray centered coordinates that can be used in a variety of applications, including Gaussian beams.

3-7

3 pm,

GC265

Albert Tarantola

Special Seminar: Inference from Observations (three remarks)

1. Observations can not provide models of a system (they can only falsify models).

2. The mathematics of falsification are not those of conditional probabilities (Kolmogorov's theory must be completed).

3. Analysis of uncertainties is not easy (I doubt anyone is ready to pay for the price).

3-10

Spring Break

No Talks

3-12

Spring Break

No Talks

3-17

John Anderson, Exxon Mobil Upstream Research

Seismic Simulation With More Physics

Simulating seismic waves from an earth model helps us to understand how key features of our seismic data relate to geology. Historically, most simulations have been acoustic and two-dimensional. Realistic simulations need to include more physics. The physics of elastic wave propagation is necessary to accurately model surface waves and ice flexure waves.

The SEG advanced modeling (SEAM) consortium is developing a new 3-D earth model representing salt tectonics in the Gulf of Mexico. About mid-year the first set of variable-density acoustic simulations will begin for Phase 1. A more limited set of elastic simulations are also in the plan. Electromagnetic simulations and gravity data will be generated.

3-19

Farnoush

Study of the stationary zone via correlation gathers

I will discuss analysis of the correlation gathers for two cases.

I)cross-correlating the first order multiple at receiver A with the first order multiple at receiver B.

II)cross-correlating the first order multiple at receiver A with the second order multiple at receiver B.

This demonstrates how helpful the correlation gather is for giving the idea about the stationary zones.

At the end, I will talk briefly about the problem of getting the exact Green's function for the case of surface source-receiver distribution (which I just started to work on).

Jia

Anisotropic wavefield separation

This is a continuation of last talk. To obtain separators of anisotropic wavefields, we need to Fourier transform iUx and iUz back into space domain. Ux and Uz are components of anisotropic polarization vectors. Because Ux and Uz's discontinuous feature at Nyquist frequency, we need to approximate them by different weights. It turns out that cosine taper is not a good taper for it. We obtained weights for different orders of finite difference approximations. Using these weights to weight iUx and iUz, we can obtain high order difference operators, and presumably using them to separate anisotorpic wavefields will guarantee good results.

Chuck

Subject 3

Summary 3

3-24

Xiaoxiang

VSP and local anisotropy

Grechka and Mateeva (2007) identified two anisotropic parameters, delta/vsp and eta/vsp, that govern the P-wave slowness-polarization VSP data recorded in vertical boreholes drilled through VTI media. Today I am going to talk about it.

Tony

Analysis for extended imaging condition and image hypercube

I will continue the talk about extended imaging condition and its result,image hypercube. I will use a simple model to show how the hypercube looks like and how it changes with the space-lag and time-lag. Also, I will show that the common analyzing way doesn't work in the examples.

Lennart

Subject 3

Summary 3

3-26

Dr. Evgeny Landa

How correct is a velocity model?

Summary to follow...

3-31

Steve

Time Shifts of Compacting Reservoirs Elastic Coefficient Estimations for Stressed Rock

I discuss linearized time shifts over compacting reservoirs for cases of zero offset data, and its associated issues. I also discuss shifts using non-linear elastic models and derivation of their stiffness coefficients using linerization methods.

Yongxia

Dynamic ray tracing in Anisotropic media

Because Gaussian beam simulation is based on dynamic ray tracing in simulation procedure. I focus on basic information of dynamic ray tracing in anisotropic media and introduce two formulations of the dynamic ray tracing, which are elastic parameter and phase velocity formulations.

Chuck

Independent Component Analysis

Independent component analysis is a relatively new tool that may be very useful for blind source separation and blind deconvolution. In this first of a two part presentation, I will give an overview of independent component analysis (which is not to be confused with principle component analysis).

4-2

Alexandre

Subject 1

Summary 1

Chuck

Passive Imaging in Lossy Media

Traditional passive seismic interferometry techniques fail in lossy media where equi-partitioned coda waves do not exist. In this second of a two part presentation, I will demonstrate the use of independent component analysis and blind source separation in passive subsurface imaging.

Jae

Integration of Seismic and CSEM data with its uncertainty analysis

I continue to discuss the Bayesian Stochastic inversion based on Metropolis-Hastings algorithm. I discuss some questions that arouse - scaling problem, convergence problem, and so on - and possible answers to them.

4-7

Dr. Boris Gurevich

Seismic attenuation and dispersion due to wave-induced fluid flow in heterogeneous porous rocks

Summary to follow...

4-9

Walter Sollner

Multi-azimuth prestack time migration

Conventional PSTM velocity analysis from one predominant azimuth direction allows determining diffraction time functions solely in one azimuth direction. As a consequence, it allows building 3D migration operators only for horizontally layered media, which have with the z-axis an axis of symmetry. Extension of the migration to handle general heterogeneity and anisotropy needs multi-azimuth data. The diffraction time function, in this work, is based on the most general form of Hamilton's principal equation and is approximated by a Taylor expansion in the vicinity of the image ray. The Taylor coefficients are derived from multi azimuth imaging parameters. The second-order coefficients, for example, are obtained from azimuth-dependent time migration velocity analyses in at least three predominant azimuth directions.

Ran X.

Micro-seismic imaging

This presentation focus on the property of wavefront of a single source and how to use wavefield database.

Derek

Flattening: What's Next

I will present two possible improvements to the flattening algorithm. I will show a possible way to push samples from the input image into the output image which may make the algorithm much more computationally efficient. Then ,time permitting, I will discuss the trade-offs of the smoothers used in our regularized least-squares inversion and a possible way to overcome these trade-offs.

4-14

Sponsor's Meeting Reviews Begin

Seminar still meets for reviews M&W @ 4pm

5-5

OVERFLOW

CWP Sponsor's Meeting 5/19-5/22, Golden Colorado

CWP Sponsor's Meeting 5/19-5/22, Golden Colorado

OR

CWP Sponsor's Meeting 5/19-5/22, Golden Colorado

CWP Sponsor's Meeting 5/19-5/22, Golden Colorado

GUEST

CWP Sponsor's Meeting 5/19-5/22, Golden Colorado

CWP Sponsor's Meeting 5/19-5/22, Golden Colorado