|
|
The Distribution
|
Activity #1: Where Do Earthquakes Happen?
Epicenters seem to be located almost everywhere on the map, but they are concentrated in some places. These clusters of epicenters are often elongated, or appear to occur next to other clusters in somewhat linear arrangements.
There appears to be some correlation between the two (epicenters and fault traces), since some of the obvious "holes" in the seismicity distribution occur in areas devoid of major fault traces. On the other hand, not all major fault traces seem to be associated with large numbers of epicenters, and some clusters of epicenters do not lie along any major fault trace.
There is a noticeable correlation between the two sets (epicenters and fault traces), but that correlation is far from one-to-one.
Since earthquakes occur when a fault slips, the fact that epicenters can be found almost anywhere on this map suggests that there must be a very large number of small faults all across southern California.
There are a number of epicenters closely associated with the San Andreas fault zone, but not nearly as many as can be found near some of the other major fault zones. Perhaps this suggests that the San Andreas fault zone ruptures primarily in large events, or perhaps the level of stress near the fault is, in general, too low to cause minor earthquakes along tiny, nearby faults.
Activity #2: Where Do Large Earthquakes Occur?
The epicenters on this map are distributed similarly to those on the map from Activity #1, and possibly are even more concentrated in bands and zones; they certainly are not randomly distributed.
Yes, the epicenters of large earthquakes seem correlated with the locations of major fault traces, though a few examples do not match the set of faults shown on this map.
The correlation between the epicenters and the fault traces appears fairly strong.
There seems to be a greater correlation between the location of fault traces and the epicenters of only large earthquakes than there was between those traces and the set of all epicenters. Given that the size of an earthquake depends on the amount of fault rupture involved, in makes sense that large earthquakes must occur along large faults, so there should be a strong correlation, as only the largest fault zones are shown on the map.
There seems to be a much weaker correlation between the fault trace locations and the epicenters on this map than there was on the southern California map. This does not necessarily support the conclusion made above.
Some of the epicenters that did not correlate with a large, mapped fault zone were at the lower end of the magnitude scale, and perhaps could have ruptured a moderately small, unmapped fault. Also, the Los Angeles Basin is an area of compression and thrust faulting, some of it blind -- so no surface trace would show up on the map for such faults. Even thrust faults that reach the surface have a very shallow dip, so epicenters would not necessarily line up along the trace of the fault (its intersection with the surface), like they do for nearly vertical faults.
Activity #3: Does Topography Signal Earthquake Potential?
Many, but not all, of the most obvious sudden topographic changes do show a correlation with seismicity.
Most of the more obvious linear features (but those not necessarily associated with a marked change in topography or elevation) do seem to correlate with increased levels of seismicity that follows a similar linear trend.
The areas of definite topographic change may have been created by previous active faults or other earth processes (e.g. volcanism) that are no longer active, and thus, no longer generate seismicity. Some linear features (e.g. canyons) may also be related primarily to erosion, and thus would not correlate with seismicity.
Activity #4: An "In-Depth" Look at Earthquake Distribution
There are definitely some fairly systematic large-scale variation in the depth of the hypocenters shown on this map; the entire image is not uniform in color, though some smaller regions have a rather uniform distribution.
The maximum depth given on the scale is 34.0 kilometers. Earthquakes in southern California below about 20 to 24 kilometers seem to be very rare. This depth is probably indicative of the maximum depth of the brittle crust across most of southern California. Based upon this alone, it is sensible to rule out that this area is the site of major tectonic collision, because that sort of tectonic environment generally creates a much thicker crust.
The Transverse Ranges, the Los Angeles Basin, the area near the intersection of the San Andreas and San Jacinto fault zones, and the extreme southern end of the San Joaquin Valley all seem to have greater than average concentrations of earthquakes deeper than 10 km. The Mojave, the Salton Sea area, and the region east of the Sierra Nevada (the Basin and Range province) seem to be lacking in deep earthquakes.
Most of the deeper hypocenters seem to be associated in some way with the Big Bend of the San Andreas fault zone. This is a restraining bend -- a zone of compression -- and there are many reverse faults and thrust faults in this area.
Extension and shearing, carried out by normal and strike-slip faults, tend to dominate regions where the crust is shallow.
The simplest explanation for the large-scale variations in depth has to do with the thickness of the crust. In areas where the crust in thin, deep earthquakes will be absent; in areas with a locally thicker crust, deeper earthquakes are possible.
Answer dependent upon current seismic activity.
Answer dependent upon current level of seismic activity and user's expectations.
Answer dependent upon current level of seismic activity.
The region covered is the rectangle defined by 32° to 36.25° N latitude, and 114.75° to 121° W longitude. The intervals were days, divided exactly at midnight (12:00 am) Pacific Standard Time, which is equivalent to 08:00 Greenwich Mean Time (Coordinated Universal Time).
Answers will vary to the first two questions. The monthly totals for 1996 are generally higher than those from 1998, especially when one considers that the area covered by the 1998 animations is larger than that covered in 1996.
A few moderately large earthquakes (green in symbol color) appeared near the northern edge of the map, and the daily count jumped dramatically -- into triple digits -- before gradually falling back to a lower level by the end of the month. The graph shows this sudden peak and gradual decay very well.
The seismicity rate for the rest of the map (the part nowhere near the magnitude 5 earthquakes) would probably not change much over the course of the month.
One; the Hector Mine earthquake of October 1999. This is a rate of one magnitude 6 or greater earthquake every four years, or 0.25 per year, so the rate from 1992 of 3 per year was much (12 times) higher than the average from 1996 through 1999!
The aftershock sequence was still going at year's end; the map shows a very clear line of epicenters along the same line that "lit up" when the Landers earthquake first struck.
Answers will vary, but may resemble the following: The rate of earthquakes in the latter half of 1992 was most likely much higher than the monthly animation chosen for Stop 2, and probably even higher than the rate from March 1998.
Yes; they seem to be associated with the largest earthquakes, and so are probably large aftershock sequences.
The two largest earthquakes seem to have occurred in 1952 and 1992.
The rate of sporadic earthquakes did vary noticeably, though not in any obvious pattern, but the large clusters definitely dominated the rate. No year of sporadic earthquakes came close to matching the number of earthquakes generated by the largest clusters.
Answers will vary.
Answers will vary.
Answers will vary.
Answers will vary.
Answers will vary.
Answers will vary.
Answers will vary depending on the rates of seismicity determined by each student.
Activity #6: Is Earthquake Timing Influenced?
Answers will vary, but the two most obvious peaks are in the early morning and the mid-afternoon. There is a low point in the late morning hours, and one at 1900 hours.
No; in fact, according to the data, the most common hour for a large earthquake to strike is between 3:00 and 3:59 pm.
Answers will vary.
Answers will vary.
Answers will vary.
Answers will vary.
Answers will vary.
Answers will vary, but should generally affirm that a random process could produce the distribution seen in the real histogram.
Answers will vary.
Answers will vary.
Yes; nighttime.
Answers may vary; an example follows: Just as AM radio signals are often more easily received at night due to lower levels of noise and interference, the same thing may apply to the seismic network that "listens" to earthquakes -- lower noise may mean that weaker signals can be received at night. If this is the case, raising the lower limit on magnitude of the earthquakes we use in our study to something large enough that the network would rarely, if ever, miss regardless of normal amounts of daytime noise, could give us a better data set -- one not subject to the same artifact.
Answers may vary; an example follows: The histogram of earthquakes of magnitude 2.0 or greater showed a much more random and, overall, even distribution with respect to the time of day than did the plot of all earthquakes. The nighttime peak was not at all present in the refined data set, nor was there any obvious bias toward any other time of day.
Answers may vary, but the refined data set is probably a more reliable one.
The plot of earthquakes no larger than magnitude 1.9 would probably have a very distinct nighttime peak. Since they are near the limit of detection, it would make sense for the number of them detected to be quite susceptible to any noise or interference.
Activity #7: Landers Shakes Things Up
(Answers will vary for all questions; suggested responses are given below.)
Drawing the aftershock zone boundary using the one rupture-length rule would change little on this diagram.
Yes; for example, there seemed to be many more earthquakes above magnitude 3 that occurred outside of the aftershock zone in the two weeks after the Landers earthquake than occurred outside that zone, and above that magnitude, in the two weeks before it.
Imposing a cut-off magnitude would probably make the comparison more obvious; that cut-off should be set somewhere around magnitude 2, since earthquakes smaller than that are not very common in the "After" frame.
It looks like the rupture propagated from south to north, given the distribution of triggered seismicity. This supports (though falls far short of proving) the idea that triggered earthquakes may be a response to the shaking produced by a large earthquake.
Activity #8: The Gutenberg-Richter Relation
[No answer required.]
The slope of the line should be negative.
Answers will vary.
Answers will vary.
Answers will vary, but should be roughly equal to 1.
Answers will vary.
Answers will vary.
Answers will vary.
Answers will vary.
Answers will vary.
Activity #9: Aftershock Sequences
The number of aftershocks, N(t), at
Answers will vary.
Answers will vary, but the new method of analysis should
have made it more evident that the North Palm Springs aftershock
sequence decayed more rapidly, and thus had a higher
The rate of decay (p) for this sequence is roughly equal to 1.3, which makes is a relatively high rate of decay for a southern California aftershock sequence.
For a value of
This aftershock sequence could be expected to last about 63 days. (i.e. At day 63, the rate would be about 3 per day, similar to the background rate of seismicity before the mainshock.)
The Oceanside aftershock sequence has the larger
The North Palm Springs sequence has a
An aftershock sequence with a low
Both of these sequences are very "productive" compared
to the average southern California aftershock sequence. The North
Palm Springs sequence has an
It is unlikely that an aftershock sequence would have
an
There is a difference in the apparent amount of "decay" in the magnitude of the largest aftershock over time.
A sequence with a high b value will appear to decay rapidly in magnitude, since there is much less chance of a large aftershock occurring as the total number of aftershocks decays over time (according to the p value of the sequence). Hence, the Oceanside sequence appears to drop quickly and then level off in maximum magnitude quite early in the sequence, while the North Palm Springs aftershock sequence displays a more prolonged decay in maximum magnitude.
The a value of each sequence probably has little effect upon the apparent decay curve of each graph, though the high a value of each sequence helps "fill out" the graphs, since it means each sequence produced a lot of aftershocks (a large data set). The p value of each sequence determines just how obvious the decay effect appears, since it influences the number of aftershocks produced over time; as this number decays, the effect of the b value (on the maximum magnitude of all aftershocks on any given day) becomes more pronounced.
The a value of the Whittier Narrows aftershock sequence appears to be smaller than the a values of the Oceanside and North Palm Springs sequences, since this sequence produced fewer earthquakes for a mainshock of larger magnitude. However, it also produced a very large aftershock for its size, which tends to indicate a higher a value.
The p value of the Whittier Narrows aftershock sequence seems to be significantly higher than that of the Oceanside and North Palm Springs sequences, based upon the method of evaluation from Exercise 1.
The a value found for the Whittier Narrows sequence was roughly -0.5, which is even higher than the two previous aftershock sequences studied. The b value of this sequence appears to be roughly 0.7, a low value, which explains why there were not many small aftershocks of this mainshock despite the high a value of the sequence.
This sequence, in a word, is weak. Its a value must be very low, near the extreme of the range of a values possible in southern California, somewhere between -2 to -3, perhaps.
After plotting the data set, its seems the a value of this sequence is roughly -2.5 (very low); its b value appears to be something like 0.9 (about normal).
Activity #10: Recognizing Foreshocks
Answers will vary.
Activity #11: An Earthquake Deficit?
Answers will vary.
Part III: Analysis and Interpretation
Removing these two earthquakes resulted in a loss of
Answers will vary.
Answers will vary.
Answers will vary, but in general this addition should more than make up for the energy deficit, so it should seem pretty important in your calculations.
Answers will vary: divide the number of years in your study by 200, then multiply by 100 (i.e. just divide by 2) to get the correct percentage.
Answers will vary.
Answers will vary.
Activity #12: Slip Rates vs. Seismicity Rates
Exercise 1
There is a lot of blue, which represents a rate of about 1 mm/yr. Yellow, representing roughly 4 mm/yr, is prominent in the lower right corner of the map.
There may be some correlation between the seismicity rate of an area and the slip rate of the nearest major fault, but this correlation seems pretty weak.
By comparing the ratio of area versus pixels, it seems like there is more correlation than originally perceived -- high slip rate faults don't cover much of the map area, but they do contribute significantly to the number of epicenters. On the other hand, the correlation is still, obviously, not one-to-one; some faults with very high slip rates simply do not show much associated seismicity. As for the comparison itself, this kind of calculation could be skewed by the choice of major faults presented on the map, but more likely to affect the results of such a surficial comparison is the fact that there is no way of noting the number of epicenters that plot atop each other in the more "crowded" regions of the map. Knowing this would allow a person to better estimate the total amount of seismicity experienced in those regions.
The ratio would be 5 fast-fault ruptures to every slow-fault rupture.
There seems to be about 5 to 10 times the length of low-slip faults, compared to the total length of faults with high slip rates. The ratio of low-slip to high-slip earthquakes, however, is about 1.
Yes, there seems to be a stronger correlation between the rate of large earthquakes and the slip rates of faults than there was between slip rate and seismicity of all magnitudes. Before the issue of ratios was noted, the correlation did not seem quite so strong.
Yes, some fault sections with very high slip rates were lacking in large earthquakes. This could simply mean that the recurrence interval of those faults is longer than the period of time (about 100 years) covered by the map.
No, none of the earthquakes of magnitude 7 are yellow, orange, or red. This finding may not be significant because the sample set is so small.
