GG101L Earthquakes and Seismology Supplemental Reading

Transcription

1 GG101L Earthquakes and Seismology Supplemental Reading First the earth swayed to and fro north and south, then east and west, round and round, then up and down and in every imaginable direction, for several minutes; everything crushing around us; the trees thrashing about as if torn by a mighty rushing wind. It was impossible to stand, we had to sit on the ground, bracing with hands and feet to keep from rolling over -March 28, 1868, Mr. F.S. Lyman account of Big Island earthquake Earthquakes are vibrations of the Earth caused by large releases of energy that accompany movements of the Earth s crust and upper mantle usually near tectonic plate boundaries, and volcanic eruptions. Seismologists study earthquakes in order to learn about these subsurface Earth processes with the hope of reducing damage and loss of life. Seismologists also use human-induced seismic energy to prospect below the Earth s surface for natural resources and buried geological structures. Today we will perform an experimental seismology experiment to determine how fast seismic waves travel through soil, and then using this seismic velocity, attempt to locate some home-made earthquakes. I. Determining the Seismic Velocity of Soil setup and data collection Using the SmartSeis Exploration Seismograph we will make and measure earthquakes outside either POST or Sakamaki. Our setup consists of an array of 12 geophones (simple seismometers) connected by a cable to a SmartSeis recording device (a seismograph). The SmartSeis produces an image of seismic vibrations (a seismogram). We will simulate earthquakes by whacking a metal strike plate with a sledgehammer. Figure 1 below shows the paths of seismic rays as they emanate from the strike plate. They initially start downward because the initial vibration is downward (the way the strike plate was hit), but then curve more horizontally and eventually up to the surface. In reality they radiate outward in all directions, but the diagram only shows those that go in the direction of the line of geophones. And of course there are also seismic waves that come to the surface at places where there are no geophones, but they aren t measured so they re not shown. In Figure 1, the little tic marks along the seismic rays represent equal intervals of time, 10 ms, for example. So in this example it would take ~65 ms for the seismic energy to get to geophone #1, ~120 ms to get to geophone #2, and so on. As long as seismic noise at the area is low, geophone #1 wouldn t show any vibrations until 65 ms after the strike plate was hit, geophone #2 wouldn t show any until 120 ms after the strike plate was hit, and so on. If you know the spacing between Figure 1: seismic ray paths geophones and the time difference between arrival of the seismic energy from the hammer blow, you could figure out the seismic velocity of the subsurface material. This is because velocity = distance/time. This is precisely what you will be doing in the first part of today s lab. 1

2 Notice that in order to reach geophones that are farther from the strike plate, the seismic rays have to pass deeper and deeper under the surface. Notice also that the rays that travel to geophones #5 and #6 pass deep enough so that they get into an underlying layer that has a faster seismic velocity. This is represented by the 10 ms tic marks being farther apart (if the same time represents a longer distance, the velocity must be higher). This difference between the behavior of seismic waves traveling to close geophones vs. those traveling to far geophones will be important to remember later. You might say Now wait a minute, the distance that the seismic energy is traveling isn t just that measured across the ground from one geophone to another, it includes the part heading into the ground and then up to the surface! You would be right. However, the curviness of the ray paths in Figure 1 is exaggerated so that you can see what is going on. The real situation is more like that in Figure 2. The ray paths still aren t exactly horizontal from the seismic source to the geophones, but they are not too far from horizontal, and we are going to ignore this slight complication. Figure 2: seismic ray paths in a slightly more realistic orientation Your TA will be busy setting up the SmartSeis seismograph. Meanwhile, this is what you will do 1. Stretch out the 50 meter tape in a straight line 2. Use the Brunton compass to determine the azimuth of this line, and record it in Worksheet 1 3. Put the receivers in the ground every 2 m along the tape (but not at zero) 4. Place the metal plate (which will be the earthquake location) at the near end of the line. 5. Record the geophone spacing in Worksheet 1 6. Once your TA has everything set up correctly 7. Generate an earthquake by hitting the metal plate with the sledgehammer. Make sure you hit with the green side up, and that you don t hit the hammer trigger wire! By hitting multiple times (separated by ~5 seconds), the SmartSeis will stack the data. The result is that the seismic signals that are the same each time (i.e., the waves generated by the hammer) add up, whereas those that are different (random noise) tend to cancel you get a cleaner trace. 8. The seismogram is kind of like a graph, with geophone number on one axis and time (measured in milliseconds ms) on the other. One ms is 1/1000 of one second. Figure 3 on the next page is an example of a seismogram from the SmartSeis. 2

3 Figure 3. Example seismogram from outside POST: geophone spacing = 2 m. First arrivals have been circled. 9. The numbers up the left-hand y-axis (1, 2, 3 ) are geophone number, not the distance. To get distance, you would have to multiply these numbers by the geophone spacing, which in the example of Figure 3 was 2 m. 10. The x-axis is time in ms, increasing to the right, and each dashed line is 5 ms. The left-hand y-axis represents time = zero, and the timer started the instant that the hammer hit the strike plate (that s why there s an electronic trigger attached to the hammer). Notice that the only labeled line happens to be 50 ms, so when you get your hardcopy printout, it will probably be useful to write in values every 5 or 10 ms. 11. Extending to the right from each geophone s number is a line that starts out straight (ideally, anyway), and then starts to wiggle. This is the seismic signal recorded at each geophone. If you look at the signal at geophone #1 above, you notice that there is a nice straight (no noise) line for about 5.5 ms. Then there is a slight dip downward, and the maximum of this downward wiggle is reached just after 7 ms. Right where the first deflection occurred, at 5.5 ms is the first arrival, and this is what you want to look for. Unfortunately, it is usually very subtle. 12. That initial body wave pulse generates surface waves which have a much larger amplitude, and as you follow the trace for geophone #1 to the right (increasing time), you can see that the ground at geophone #1 vibrated pretty significantly and continued to do so until the recording ended. 3

4 13. Notice that the first arrival at geophone #2 is at time = 10 ms, later than when it arrived at geophone #1. This, of course, makes sense because geophone #2 is farther from the strike plate than geophone #1 is. 14. As you go from geophone #1 to geophone #6 there is a pretty regular pattern of the first arrival and subsequent surface waves occurring farther to the right (although geophone #6 has noisy data). This progression to the right makes sense too, because it takes longer for the seismic energy to get out to these farther geophones. Beyond geophone #6, the signals are unfortunately quite noisy. 15. It is important to notice that if you were going to draw a line connecting the first arrival positions, there would be a kink at geophone #3, with one line connecting 1, 2, and 3 (and 0), and another line connecting 3, 4, 5, and 6. This kink is an indication that there is a boundary beneath the surface that is separating slower material above (the soil) and faster material below (probably the ~70,000 year-old Tantalus lava flow). Seismic rays that travel beyond the distance of geophone #3 (which is 6 m because they re 2 m apart) has to pass through this lower layer. Because sound travels faster in this deeper layer, the arrival at geophone #4 is earlier than you would expect from just looking at the arrivals at geophones #1, 2, and OK, where were we? Once you get a good, clean-looking seismogram from stacking 4-5 hammer whacks, have your TA print it out, and write on it something to the effect of Distance Determination Seismogram S/he will xerox this so that each of you will have a copy. Instructions for actually determining the velocity are presented in section III of this lab handout. II. Determining the Location of an Unknown Earthquake setup and data collection 17. Select a starting point somewhere in front of POST or Sakamaki, and extend the two 50 m tape measures outward from this point at 90º from each other. Use the Brunton to make sure that it is a 90º angle and to record what the direction of each tape is in Worksheet 2. These two tapes will form the coordinate system for your seismic area, and will be analogous to latitude and longitude or Easting and Northing in the UTM system. Record in Worksheet 2 and on the graph/map on p. 11 the azimuth of each axis with respect to the place where they intersect. 18. Next, pull up the geophones (carefully) from the straight line and re-insert them roughly in a circular loop within the angle produced by the two tape measures. Plant flags next to each geophone so that they can be located easily. It does not have to be a perfect circle, and in fact it can actually be any shape you want. Make sure each geophone is still connected to the geophone wire. 19. Now you need to use the tape measure to determine the location of each geophone with respect to your coordinate system. Probably the easiest way to do this is to use the piece of rope. Have one person hold one end of the rope at a geophone and another person pull taut it so that it is perpendicular to one tape measure. You can use the Brunton to make sure that it is actually perpendicular if you like. Record the distance along that particular axis for that particular geophone, and record it in Table 2. It will probably be quicker to measure the position of all the geophones with respect to one axis first, and then measure their positions with respect to the second axis, but it will require care not to get the numbers mixed up. Having a team of 3-4 people working on this will help. 20. Next, the TA will send half the class around the corner so that they can t see the seismic area. The folks left behind will then select an earthquake location somewhere in the area defined by the tape measures. Place the strike plate here, and determine its location relative to the two tape measures the same way you determined the positions of the geophones (use the rope). Report this position to your TA, and record it in Worksheet 2 where it says Unknown Earthquake Location # With half the class still out of sight, make sure your TA has the SmartSeis all set and ready to record, and then whack the strike plate 4-5 times as before. Remember that you hit with the green side up, and that you need to take care not to get tangled up in the hammer trigger wire. 22. After your TA makes sure that the seismogram is good, pick up the strike plate, and bring it and the hammer back to where your TA is. This is so that the other students don t get any hints about where the unknown earthquake has taken place. It might even be a good idea to put a few leaves on the spot where the strike plate was, especially if it has made a tell-tale square indentation into the ground. 4

5 23. Have your TA print out the seismogram, and label it Unknown Location #1 24. The two groups next trade places, and those who were out of sight will now repeat steps 21-23, filling in Unknown Earthquake Location #2 in Worksheet 2. This time have your TA label the seismogram Unknown Location # At this point you are pau with data collecting. Carefully pull up all the geophones and disconnect them from the geophone cable. Pull up the flags also. Put all the geophones and other equipment back in the buckets and on the cart, and return them to POST 704. Please scan the experiment area to make sure that nothing has been left behind. III. Determining the Seismic Velocity of Soil data reduction and interpretation 26. While your TA is making xeroxes of the seismograms, you can plot geophone positions from Part II on the graph/map provided on p. 11 of this lab handout. Make sure to label the azimuths of the axes with the same numbers you recorded with the Brunton compass. 27. For the velocity determination, you will be looking at the seismogram that was produced when the geophones were all in a straight line. It will be labeled Distance Determination Seismogram or something similar. Try to find the first arrival for each geophone s trace, and circle it. These are pretty subtle, and even trained seismologists sometimes argue about them. Figure 4 shows some first arrivals selected for the sample seismogram. Use the time scale to determine the time of each first arrival, and fill in the last column of Table 1. If the first arrivals are Figure 4: example seismogram showing first arrival picks. too hard to see or are all messed up, you can instead use the first full trough or peak in the seismogram trace. But you have to be consistent. 28. For the data from the example seismogram, the first arrival times for the first 6 seismograms would look like the following: Geophone # Distance from strike plate (m) Time of first arrival (ms)

6 29. You don t actually have a geophone #0, but if you did, its travel time and distance would both be zero. 30. Next, create a travel-time plot from the second and third columns of Table 1. Graph paper for this can be found on page 11 of this handout. You want to plot travel time on the x-axis and distance on the y-axis. Join your points with one or more straight, best-fit lines, and if you have noisy data for your farther geophones, extrapolate the line out to those distances. 31. Is there a break in slope for your points? In the example data (Figure 5) the answer is yes because there is an obvious bend in the line connecting the points. One straight line fits the data for out to 6 m (geophones 1-3 and zero), and another fits the points from 6 m outward (sort of). This means that in order for seismic waves to travel as far as geophone #4 (8 m), they must pass through a layer with a higher seismic velocity that is beneath the soil. The slopes of lines on travel-time plots are equal to seismic velocity. Remember, the slope of a line is the vertical distance divided by the horizontal distance. For the upper layer, which seismic waves traveling outward as far as 6 m travel through, the slope would be: (6-0)m / (14.9-0) ms = m/ms Figure 5: example travel-time plot for seismogram shown in Figure 4. Note the kink at 6 m (geophone #3). The more gradual slope from 0 to 6 m is an indication of a slow seismic velocity whereas the steeper slope beyond 6 m is an indication of a higher seismic velocity. Note that the upper slope has been extrapolated out to 18 m. There are 1000 ms in 1 second, so to get the seismic velocity in m/s, you multiply by 1000, and get 402 m/s. Between geophones #3 and #6, the slope would be: (12 6)m / ( ) ms = m/ms = 2307 m/s. Thus, there is quite a difference between the seismic velocities of the soil and whatever that underlying layer is. 32. Keep track of where the break in slope between the two best-fit lines is because you ll need this information. 6

7 IV. Determining the location of the unknown earthquake If we have a travel-time plot (or we know the velocity) in a region such as the POST yard, and we know how long it took a seismic wave to travel from an earthquake to each geophone, then we can calculate the earthquake-to-geophone distances. Because there was only one earthquake location for each unknownearthquake experiment, the earthquake-to-geophone distances should theoretically all place the earthquake in the same location. 33. The idea is that each geophone records nothing until the seismic energy arrives. The time between when the hammer hits the strike plate (and starts the recording) and when the geophone records the arrival of seismic energy tells you how long it took for the seismic energy to travel from the strike plate to that particular geophone. If you can measure how long that time is, then you can look up the corresponding distance on the travel-time plot. 34. All this tells you is that the earthquake was a given distance from that particular geophone; what this does is narrow down the possible earthquake location to a circle that is centered on that geophone and has a radius equal to the geophone-earthquake distance. This is because from the point of view of that geophone, any location on that circle satisfies the calculated earthquake-geophone distance. 35. If you have a signal from a second geophone, it will allow you to calculate a second earthquake distance, and plot it as a circle centered on that second geophone. Every point on that second circle will satisfy the calculated earthquake-geophone distance for the second geophone. 36. Two circles intersect in two places, so you have narrowed down the possible earthquake position considerably. What you need is a signal from a third geophone, and a corresponding circle centered on that third geophone. Three circles can only intersect at a single point, and that point will be the only location that satisfies the data from all three geophones. 37. Of course, no data are perfect, and it may be that the three circles don t quite intersect at an exact point. Instead, they may define a small triangle. All you can say is that the earthquake was somewhere in that triangle. Better yet, if you have data from >3 geophones (which you do), you can make up for some of the noisiness of the data, and get a better earthquake location determination. 38. Figure 6 illustrates how this would work for data from 6 geophones (the filled squares). A distance to the earthquake is calculated for each geophone, and a circle with that distance as its radius is drawn around each corresponding geophone. The point where all 6 circles intersect is the only place where the data from all 6 geophones can be accommodated, and that must therefore be the earthquake location. 39. From the seismogram you have for an unknown earthquake, determine the first-arrival time for each geophone. Remember that the first arrival is indicated by the first, usually very subtle, downward deflection of the seismic trace. Figure 6: diagram illustrating how data from 6 different geophones allow you to locate an earthquake. 7

8 40. The hardest thing to do is pick out the all-important first arrivals for each geophone. Figure 4 shows examples, and admittedly they are often pretty subtle and/or ambiguous. 41. Once you have determined the earthquake-geophone distances for each geophone and entered them in Table 2, use a compass (the circle-drawing kind, not the which-way-is-north kind) to draw circles around each geophone on the final map. Remember, the radius of each circle will be equal to the earthquake-geophone distance you calculated for that geophone. 8