GPH 492/692 - Refraction Velocity Lab

John Louie, January 20, 2013
Copyright © 2012 John N. Louie, all rights reserved.
The Resource Geology Seismic Processing System for Java (JRG) Home Page

Contents:

Introduction - Getting Java, JRG, Viewmat - Preparing Your Data - Running Viewmat - Loading Data
Displaying Seismic Records - Setting Survey Geometry - Picking First Arrivals - Assignment

Velocity Estimation from First-Arrival Times

Objective and Due Dates -- This exercise allows you to pick first-arrival times from a refraction survey and try two methods of estimating a velocity section. First you will make a time-distance plot and use simple equations from the notes to find refractor velocity, depth, and dip. Then you will send your picked travel times to a local company, founded by Mackay alumni, that has developed the most advanced refraction-time inversion method available. They have agreed to invert your data and send a velocity section back to each of you so you can compare the two results.

Target Dates:

This write-up mostly gives you detailed instructions on how to run the Viewmat application. The actual assignment, telling you what to do and what to turn in, is at the bottom.

You are free to use the JRG package after the class is over (that's why the instructions here cover problems that may not occur with the assignment's data set). The package and its source code are available for use and open-source modification for free under license from the code's author. The entire code you use here is Copyright 1998-2013 by John N. Louie.

Key Skills

When your team completes this lab exercise, you should have gained experience toward the following skills.

Getting Java, JRG, and Viewmat

For this exercise you will pick first-arrival times from seismic traces using the instructor's RG Seismic Processing Package for Java (JRG). To complete the exercise you will need three resources:
  1. A computer onto which you can install or use Java software: Windows; MacOS; Solaris; or Linux. The computer lab downstairs in the DeLaMare Library is available for your use in this course. You can use the lab any time it is not reserved for a class.
  2. An Internet connection so you can download the software and data, and email your picks to Optim Inc.
  3. A spreadsheet application like MS Excel to plot your time-distance curves; not essential but maybe easier than plotting them by hand.
If you have trouble getting access to any of these resources, please see the instructor.

Getting the Java Platform

You computer probably has Java installed already. Go to java.com to find out, and install the latest free Java Runtime Environment (JRE). You will want to have Java version 1.4 or later.

Click here for additional details on how to see if your computer has Java installed, and where to get Java versions for older computers.

Getting JRG and Viewmat

Download and unzip the 492-refraction.zip ZIP archive. Put it in a convenient place such as your desktop since you will be using it constantly for this assignment.

If you will be using a UNR computer that you log into with your NETID, you will need to copy this folder to a memory stick before you log off. If you don't, you WILL lose your work!

The ZIP archive contains:

  1. jrg.jar and jrg500.jar, the JRG/Viewmat application. Use jrg500.jar when you are working with a large data set, over 5 Mb.
  2. dixie-refr.jrg, a JRG Pack with the seismic data you will pick, and the other information below, all ready to load into JRG/Viewmat.
  3. All these web pages for the assignment, for your convenience.
  4. dixie-refr.sgy, the seismic refraction is also contained in a 1.0 Mbyte binary SEG-Y format file. SEG-Y is a standard defined by the Society of Exploration Geophysicists for the interchange of seismic survey data. These data were obtained in southern Dixie Valley, Nevada, by the Spring 1998 Geol 453/653 class. A journal paper describing our results is R. E. Abbott, J. N. Louie, S. J. Caskey, and S. Pullammanappallil, 2001, Geophysical confirmation of low-angle normal slip on the historically active Dixie Valley fault, Nevada: Jour. Geophys. Res., 106, 4169-4181. (Note: the SEG-2 format is commonly written by engineering seismic recorders, and the data has to be converted to SEG-Y).
  5. dixie-med-vp.txt, a text file of source and receiver station locations.
Click here for additional details which files are in the ZIP archive, and how to download them separately.

Preparing Your Data

Viewmat makes exporting your picks to Optim's SeisOpt®@2DTM package quick and automatic, especially if you follow good recording practice. For this exercise your seismic traces are already correctly labeled with source and receiver geometry. If you were employed in contracting or supervising seismic refraction surveys, you would always ask your recording observer to enter complete survey geometry info into the instrument during the survey. If the original field records have at least the correct shot numbers, receiver numbers, and an in-line distance entered for each channel, you have enough information for SeisOpt®@2DTM. Keep in mind that geometry information is best entered and checked in the field, not later.

Of course, if you are asked to work with legacy seismic data that was recorded before your involvement with a project, you will often have to add geometry information yourself. If you have a data file without enough geometry included, do not worry. You can still display the records quickly, and enter the geometry info for the survey while you are using Viewmat.

In this course, after we have recorded seismic data during our Spring Break field exercise, we will need to export our seismic records as standard SEG-Y files for Viewmat. Most recorders store records internally in SEG-2 format. Every instrument manufacturer provides routines that run on their equipment and produce SEG-Y files. If possible, convert all the records from one survey into one large SEG-Y file, and transfer that to the computer where you will interpret them with Viewmat. The dixie-refr.jrg JRG Pack you are given for this exercise contains all the correct geometry information, but the original dixie-refr.sgy data file in SEG-Y format does not.


Running Viewmat

Double-click on the 492-refraction.zip archive to unpack it and open the folder. Start the JRG/Viewmat application by double-clicking on the jrg.jar file. If you have trouble with this, try these detailed instructions for older computers.

If while working with a large data set Viewmat stops running, or you can still make menu selections but it doesn't seem to be actually working on the data, then the application is probably in an Out of Memory condition. Quit Viewmat and start it again by double-clicking on the jrg500.jar file, which will allocate most of your computer's RAM to the application.


Reading Data

Viewmat starts up with a default blank data set. Locate the menu bar at the top and select Load JRG Pack... from the File menu. A dialog box comes up asking you to navigate to your JRG Pack folder dixie-refr.jrg. If you are not able to select the ``dixie-refr.jrg'' folder directly, select any of the files inside it and click OK or Open.

For reading SEG-Y data files, see these detailed instructions. Once you have opened and worked with a data set in Viewmat, you can save your work at any time as a JRG Pack from the File menu, so later you can load it in easily to continue working. It is always a good idea to save your work every 15 minutes or so as a JRG Pack, onto your memory stick if you are on a public or lab computer.


Displaying Seismic Records

To easily pick first arrivals on some data sets, you will probably want to change how Viewmat initially shows your data. The dixie-refr.jrg JRG Pack has everything set up for easy first-arrival picking, but see these detailed instructions if you want to learn how to edit the display parameters.
  1. First, if you have loaded a multi-record file, press the Animate button. The first record of a survey is often the worst!

  2. Apply trace balancing by selecting tegain from the On Each Vector sub-menu of the Methods menu on the new display. Press the tegain button in the tegain dialog. Trace balancing changes the actual data amplitude values in a way that can't be undone, so make sure you save your results as a new file or JRG Pack, and don't over-write the original data.

  3. tegain changes the overall amplitude of the data, so it may appear gray or clipped. To pick first arrivals, you want to set a low clip value that emphasizes when the trace amplitudes go from positive to negative. This will oversaturate the large-amplitude waves, but we are not looking at those in this lab. From the Edit menu, select Clip at RMS for a better display. Changing the clip does not alter the data, only how it is displayed. Press the Animate button to see all the records of a multi-record file.

  4. To enlarge the data display, select 200% or 500% from the Zoom Image To: sub-menu of the View menu. Changing the zoom or vertical exaggeration do not alter the data, only how it is displayed.

  5. This exercise does not benefit from filtering the data, but with other data sets you may need to apply a bandpass filter to your traces to mitigate source-generated noise. Check these detailed filtering instructions.

  6. Save another JRG Pack with a new name after making your adjustments (again, to your memory stick if you are on a public computer).

Setting Survey Geometry

  • For your first-arrival picks to export easily to SeisOpt®@2DTM, and to easily make time-distance plots of your picks, you must have correct geometry information loaded from your data file into Viewmat. The geometry describes the map locations of the seismic sources and receivers for each record. This information usually comes with the SEG-Y file from your acquisition contractor. Our Dixie Valley refraction data geometry is correct when you load the dixie-refr.jrg JRG Pack.

    You will notice on the display that each record (each plane) shows the seismograms from 48 receiver geophones recording the same seismic source, or shot. Each vertical strip of the color image shows the seismic amplitude with time after the shot increasing from zero seconds at the top to one second at the bottom. The time axis is reversed like this because structures that are deeper in the cross section below the survey line will show seismic waves that arrive later. So the signals from deeper structures will this way plot below the signals from shallow structure.

    From left to right across the top of the color image, the receiver traces are arranged according to the geophone X coordinate. The X coordinate increases eastward, so left to right is east, making a north-looking section. Traces with earlier first arrivals, nearer the top of the color image, are closer to the seismic source. The first record has a source off the east end of the receiver line, the fifth record has its source at the left end, on the trace of the 1954 Dixie Valley magnitude 7.2 fault rupture, and the other records have sources on the line, each time only 3 m or so from one of the receiver geophones.

    For more details on how to correctly apply geometry information when it does not accompany your data file, see this detailed explanation of how to load the Dixie Valley survey geometry.


    Picking First Arrivals

    The heart of this exercise is picking the times of the first-arriving waves on each trace on each record. See the detailed picking instructions to find out how to make your time picks, edit them, save and import them into Excel, export them for Optim, and see a data plot.

    Assignment Tasks

    You should complete all but the last question by the thrid week of the semester, and send your team's picks to Sathish Pullammanappallil by then. He will send the optimized velocity result back to you in several days, and you should answer the last question and turn in all the answers by the fifth week of the semester.
    1. Install Java and JRG.

    2. Download the seismic and geometry data.

    3. Run Viewmat, load the seismic data JRG Pack, and check the geometry.

    4. (10 points) Pick the first arrivals, correct them, and
      1. send the four SeisOpt® files to satish@optimsoftware.com.
      2. Also turn in a correctly labeled plot of the first record with the picks superimposed. Email the ''.ps'' file out of your team's saved JRG pack to the instructor for comparison against the other teams' results.

    5. (10 points) Save your picks, import them to a spreadsheet, and make a t-x plot of the off-end and reversed shots (the first and last records only). See the hints above on importing saved picks to Excel; or make a plot on graph paper by hand.
      1. Print out your team's plot and give it to the instructor, with all axes completely and correctly labeled, and indicating which picks are from the forward and downdip (source is downdip of the receivers) shot, and which are from the reverse and updip shot.

    6. (20 points) Make a dipping-layer-over half-space interpretation of the picks from the two end-shots, the first and fifth records. Draw straight-line fits of apparent velocity and intercept time on your plots by eye (only an experienced spreadsheet user will know how to have the fits done automatically, and hand fits are just as good).
      1. First draw the V1 line from the source location at zero time (the (0,0) point) through the near-offset picks. Then draw another line at the higher velocity V2 through the rest of the points. (Always draw the shallowest layer's line first, then the next layer down, etc.)
      2. Now draw lines for the reversed shot's times. As you notice, the equations all assume you get just one V1. Most analysts, you should know, average the V1 slopes drawn on the forward and reverse. So you can do the same.
      3. But make sure your (0,0) constraint is really at zero source-receiver offset. Check the "sx" values in the spreadsheet for the x-coordinates of the shots.
      4. Constrain the higher-velocity fits with the reciprocal time, and mark that on the t-x plot.
      5. You will have to calculate the velocities of your fit lines by hand from your marked-up plot; this is easy if you have included many grid lines at fine x and t intervals.
      6. Use the appropriate equations from the text or lecture notes to compute layer velocities, refractor depth, and dip.
      7. Draw a cross-section showing refractor geometry.
      8. Turn in the section and the t-x plot with the fit slopes and intercepts. You can plot everything by hand on graph paper if you need to; just make sure the plots are neat and completely and correctly labeled.
      9. Report the dip. Which direction does the refractor dip? (Note: the higher station X-coordinates are to the east while the higher station numbers are to the west.)

    7. (10 points) Using another copy of your t-x plot, make the highest-velocity interpretation you can of the forward and reverse times, especially considering the error inherent in each pick (particularly the far-offset picks). Don't forget to fit a reciprocal time.
      1. Draw this high-apparent-velocity interpretation on the t-x plot and turn that in.
      2. Recompute the refractor velocity, dip, and depths, again for a single refractor.
      3. Report depths, velocities, and dip; drawing another section is not neccessary.

    8. (10 points) Now make the lowest-velocity interpretation you can.
      1. Turn in another copy of the t-x plot with the low-apparent-velocity fits.
      2. Report depths, velocities, and dip.
      3. How do these compare with the high-velocity model?
      4. What can you say about the accuracy of your original interpretation?

    9. (10 points) Make a three-layer, three-velocity interpretation of our data. The equations for multiple dipping interfaces, which you can get from Burger's text (on reserve in the DeLaMare Library with its software CD) on p. 89 and implemented in his RefractSolve program, are too complex to use here by hand.
      1. Just use the picks from the reverse shot (the last record) and assume no dip, unless you want to use RefractSolve as in previous offerings of Geol 492/692.
      2. What evidence do the reverse pick times show for an intermediate velocity? Do the forward pick times show any?
      3. Remember to start drawing V1 from (0,0), then V2, then V3.
      4. Report velocities and thicknesses from your best line fits. Of course refractor velocity interpreted from the reverse record assuming no dip will be lower than in the 2-layer dipping interpretation.
      5. Turn in another copy of the t-x plot with the 3-layer fits on the reverse times.

    10. (10 points) Supposing the intermediate layer is really a hidden thin layer,
      1. draw your 3-layer fits of the reverse picks on another copy of the t-x plot, but move your V2 lines back to just intersect the V1 to V3 crossover.
      2. Report velocities and thicknesses for this minimum amount of the intermediate thin layer, and
      3. turn in the fits on the t-x plot.
      4. How do the depths to the deepest V3 layer compare with those from your best-fit 3-layer interpretation above?
      5. What constraints do you have on the depth of the deepest refractor? Your inferred hidden-layer errors may be applicable to the dipping-refractor interpretations as well.

    11. (10 points) Evaluate the effect of the time picks not landing exactly on your fit lines.
      1. Do the data show any delays that seem consistent on both forward and reverse shots?
      2. Starting with your 2-layer interpretation, measure the maximum delay by fitting a line to the slope of the longer-offset refraction arrivals, and then moving it (while keeping it at the same apparent velocity or slope) to a) just graze the earliest times from that layer, and b) just graze the latest times.
      3. Turn in a copy of the t-x plot with the maximum and minimum times.
      4. Subtracting the two intercept times should give you an estimate of the maximum .

    12. (10 points) Assume all your delay results from structural deflections h in the refractor.
      1. Using the equations in the notes, compute this maximum deflection. Ignore the dip, or intermediate layers.
      2. What proportion of the average refractor depth is this deflection?
      Later we can look into assuming the refractor is also a density difference of, say, 0.4 g/cc, and use the simple infinite-plate formula from gravity to estimate what is the maximum number of milligals of gravity anomaly these structural deflections might produce; and whether we could observe such structure with our gravity instruments.

    13. (10 points) Next assume all your delay results from lateral velocity changes in the shallowest layer (again ignoring dip and intermediate layers).
      1. Compute the new surface velocity V0.
      2. What proportion of V1 is the velocity change V1 - V0?

      Assuming that resistivity in this layer changes in proportion to changes in velocity due to changes in porosity, it is possible to estimate the resistivity changes if you have some locations where resistivity and seismic data overlap in the same environment. We may look at this later.

    14. (20 points) (Turn in Feb. 24) When you have the results of Optim's SeisOpt®@2DTM on your picks of all 5 records,
      1. compare them against the information on dipping structure and errors you derived from the questions above.
      2. Turn in both the SeisOpt® velocity section and the hitcount section (black and white printouts of color plots are OK).
      3. Write a paragraph comparing SeisOpt® results against the simple layer calculations. Address the following:
        1. Based on the velocities above and below your dipping refractor, draw the location of the refracting ``interface'' through the SeisOpt®@2DTM velocity section, on the copy you turn in.
        2. Compare the depth and dip of the ``interface'' between your layer calculations and the SeisOpt® interpretation.
        3. Compare your intermediate-layer calculations against the velocity gradients you see near the ``refractor'' in the SeisOpt® velocity section.
        4. Compare the depth deflections in the ``refractor'' as seen in the SeisOpt® section against the estimates of h you computed from for question 12.
        5. Compare the velocity variations above the ``refractor'' as seen in the SeisOpt® section against the estimates of V0 you computed from for question 13.

    Contents:

    Introduction - Getting Java, JRG, Viewmat - Preparing Your Data - Running Viewmat - Reading Data
    Displaying Seismic Records - Setting Survey Geometry - Picking First Arrivals - Assignment