1. Geology 5660/6660 Applied Geophysics 23 Feb 2016 Lab 3 © A.R. Lowry 2016 Seismic Reflections Next assignment due one week from now Due noon Mar 1

2. Your lab assignment (due at noon on 2 Feb): Install the Burger et al software on a pc/laptop you use Model using Refract the following, using 20 stations, 5m shot offset, 5m spacing:  a 9.4 m layer ( V = 2 km/s) over a 5 km/s infinite space  a 3 m layer ( V = 2 km/s) over a 3 m layer ( V = 1 km/s) over a 5 km/s infinite space. Compare and contrast the travel times derived from these models. How would you distinguish these two Earth structures based on the observed data (i.e., the circles)? Model the same two models using Reflect (be sure to check “show reflections”) What, if any, additional information does this give you? Lab write-ups should include plots, numerical data if relevant, and explanatory background where appropriate!

3. Visually, the observations (i.e., the circles in these plots!) appear very similar, insomuch as one can tell by looking… The lines are model representations (& the first “direct arrival” lines are a bug!)

4. In fact they’re almost identical! And if you look at times, they are identical to “standard” sig-dig

5. The reflections on the other hand are slightly different (as one would expect, because the reflection seismic method does not have the same limitation on sensing low velocity layers). But differences are subtle and may not be clear how one would use, because layers are thin and reflection wavelets superpose

8. However it’s much easier to pick out the differences if we use a shorter wavelet pulse! Because now the wavelets are not superposed. Hence, resolution of the seismic reflection method depends on frequency content of the returns!

9. (1) Seismic data above were collected on a farm near Gosport, IN. The faint lines demark 0.01 s intervals; darker lines are at 0.05 s. (a) Pick (as best you can) the time of the first arrival on each trace. (Hint: The furthest first arrival is at < 0.1 s). (b) Geophone distances are (0,4,8, 12,16,20,26,32,38,44,50,56,62,68,74,80,86,92,96,100,104,108,112, 116) m from the source. Enter the times & distances into Refract and model the velocity structure assuming 0° dip on all layer boundaries. Do any features of the data suggest nonuniform layer thickness? Lab 2: Part 1. Travel-Time 0.01 s 0.02 s 0.03 s 0.04 s 0.05 s

10. One subtlety : The positive deflection at 4 m would imply a (nonphysical) velocity of 200 m/s (< 330 m/s for sound in air!) The true first arrivals are negative (polarity problem) so hard to see. Picking the positive gives wrong V, so wrong thickness, in the first layer, and hence wrong thickness (but not wrong V ) in deeper layers. Hints : Helps to blow up to large scale; more accurate if you create and use a ruler…

11. Gosport Best FitRMS = 1.08 ms

12. Lab 2: Part 2. Amplitudes in Reflect (2) This plot shows output from the program Reflect. Assume the distance between traces corresponds to a displacement amplitude of 1. Do amplitudes reflect expected relations for geometrical spreading? For attenuation? Why or why not?

13. 1/r Direct Spherical Spreading: 1/r Attenuation: Cylindrical Spreading: Refracted Air Reflected Different arrivals have different amplitudes within each trace, but do not exhibit expected distance dependence…

14. (3) The plot @ left shows only the reflected wave for the same model. (a) Derive a relation for angle of incidence  of the reflected P-wave as a function of distance. (b) The image at right (from a website application) shows Zoeppritz amplitudes for the reflected P. Are these amplitude relations exhibited in the model at left? Why or why not? Lab 2: Part 3. Amplitudes in Reflect

15. x h 

16. 14° 27° 36° Trace at 9 m should be ~ 5 times the amplitude of trace at 6 m!

17. Must be doing something though! Probably uses reflection and transmission coefficients, but not spreading, attenuation or Zoeppritz.

18. (4) Amplitudes shown below were measured from first arrivals in the Gosport data (Part 1). (a) Are these consistent with what you might expect from your travel-time analysis? Why or why not? ( Hint : Think about the propagation path of each arrival!) (b) Estimate frequency of the first-arriving wave for each arrival from the plot (Part 1). (c) Correct the amplitudes for the effects of geometrical spreading. Then, separately for each layer, estimate the quality factor Q of anelastic attenuation. Lab 2: Part 4. Amplitudes from observations (Hint: Don’t correct for geometrical spreading if r = 0 !) (Hint: Don’t try to compare refracted arrival amplitudes to r = 0 amplitude!)

19. Must first recognize that there are three different arrivals with three different values of Q ; two types of spreading Layer 1 spreading correction is Layer n (= 2, 3) spreading correction is Invert the relation:  (Since amplitude measurements are noisy, I used all arrivals from a layer & averaged) 1 23 V = 1330 m/s f = 250 Hz Q = 9.5 V = 3510 m/s f = 125 Hz Q = 6.6 X X 465 100 15.1

20. (d) Assuming that both your velocity and your attenuation structure estimates are correct, look through your text (and whatever other literature you may be able to find, but be sure to include references!) for examples of soils/sediments/rocks that match your estimates of V P and Q. Does the combination of these two properties reduce the ambiguities in possible interpretations? What interpretation of the data would make the most geological sense?

21. Crustal Rocks 123 The velocities are consistent with a wide range of possibilities but Layer 1 is clearly dry soil, alluvium or dry sand; Layer 2 is either the same material as Layer 1 (but below the water table) or a change to a different type of alluvium or shale; and Layer 3 is shale, sandstone, limestone or dolomite. ALL the quality factors are extremely low (soil Q = 2–20!), ruling out consolidated sed rx in Layer 3. Given the geologic environment, L1 = dry soil; L2 is most likely water table, & L3 is either shale or (more likely) glacial-derived clay.

22. Lab 3: Seismic Reflection 02000 40006000800010000 0 0.5 1.0 1.5 2.0 2.5 3.0 (m) (sec)

23. (1)These are synthetic (idealized) data generated using a numerical (finite element) approach. There are four primary reflections present in the data. Pick ten travel-times (i.e. at ten different distances) from each of at least three of those reflections. Using Excel (or some similar software that allows you to regress a line fit– you can do this by hand calculator or in Matlab, or on graph paper if you prefer) plot x 2 versus t 2 for each of your reflections, and calculate the velocities and thicknesses these would imply assuming each reflection is from a single-layer interface over a half-space. (2) Now do the same for one of the multiple reflections.

24. (3) Again, without using Reflect, use the intercepts and slopes of lines on your x 2 – t 2 plots with Dix equations to calculate velocities and thicknesses. (4) Do similar calculations for velocities and thicknesses using Dix equations, but now instead of using taken from the inverse slope of x 2 – t 2, use the first- order T NMO relation (i.e., use the inverse slope of a line fit to a plot of T NMO versus x 2 /2t 0 ). How does this change the model? Which is likely to be more accurate and why? (5) Now plug your travel-time picks into Reflect and calculate a model of velocities and thicknesses. How does this compare to your “hand-calculated” models?

25. (6) Create a record-section plot for all of the arrivals using Reflect and compare/contrast with the synthetic record section from which you derived your travel-times. Why do you suppose they are different?