This page should help you familiarize yourself with the purpose and operation of the interactive, online section of this activity.
This activity uses frames and layers to allow the user to interactively analyze waveforms and locate epicenters. If your browser supports neither frames nor layers, you will not be able to use this activity. (We apologize that there is no alternate page available.)
When you start the activity, you will see three frames created in your browser window. The image below shows you the main features of each frame, which vary with each of three different screens. All these features are discussed in detail in the text that follows; you can switch between screen imagess by clicking the "SCREEN" buttons at lower right:
When you first load the activity page, the Action Frame will
contain a list of links, and nothing else. This is the
earthquake list, described below.
When you click on an entry in the list, the
Action Frame will reload (we'll
call it Screen 2 now), and the first four fields of data in the
Data Frame will be filled in with information
about the earthquake you selected. Your next step is to
analyze the seismogram traces that fill
the Action Frame in Screen 2.
This frame also contains a menu bar of options
you can use to switch between screens, pull up this help page, or start
over with a new event.
As with the event number and the
date, the number of traces will not change
until a New Event is chosen.
The four options are: New Event, Check Traces, View Map, and Help. In Screen 1, the only one of these you should need to click is "Help". None of the others will function. They are discussed under the "Menu Bar" entries for Screen 2 and Screen 3.
The "Help" option on the menu bar will, at any time, open
this explanation page (but with a black background) in the
Action Frame. You can use this to
check on the function of a feature without exiting the activity
or opening a new window. To return to Screen 1, just click
the "New Event" option on the menu bar.
Each trace should start out roughly flat at the left end, and then deflect significantly at some point toward the right, becoming a seismic waveform. Elapsed time runs from left to right on these seismograms; the time at which the left side begins is given in the Time field of the Data Frame when Screen 2 is first brought up (by choosing an event from the earthquake list).
The three-letter station name on each seismogram is used for identification of each record; it represents the seismometer used to record that trace. This information will be more useful in Screen 3.
The red arrows on the seismograms allow you to become a seismic data analyst; with them, you are to pick the arrival time, on each trace, of the first P-wave arrival. This process is explained more fully under the P-wave arrival arrow section of this page.
You can bring up another arrow on a seismogram by clicking on the "Pick an S Arrival" button. This feature is described in more detail later.
Some technical information about the seismograms used in this activity:
|
|
BAR DAN DGR EDW GPO GSC HEC ISA OSI PAS |
Barrett Dam Danby Domenigoni Valley Edwards AFB Geothermal Program Office Goldstone Hector Isabella Osito Adit Pasadena |
|
PFO PHL PLM PLS RPV RVR SBC SOT SVD VTV |
Pinyon Flat Observatory Parkhill Palomar Pleasants Peak Rancho Palos Verdes Riverside Santa Barbara Solemint Seven Oaks Dam Victorville |
When the seismograms are first loaded (after an earthquake has been chosen from the list in Screen 1), all the P-wave arrival arrows are in a default position on the seismogram. The arrows can be moved, however -- indeed, it is an integral part of this activity that, for every trace, you click on and drag this arrow so that it aligns with the arrival time of the P wave.
How do you find the P-wave arrival? That's actually fairly simple, though some arrivals are more subtle than others. The P wave is the very first seismic wave to reach any given point after an earthquake. Consequently, it produces the first ground motion. So instead of thinking that you're looking for a certain wave arrival, it may help just to think that you're looking for the point in time when the earthquake -- the ground motion that produces the oscillations in a seismogram's trace -- first begins. In other words, look for the first noticeable deflection of the trace from its position of rest; this is the P-wave arrival. The at-rest position for each trace can be found at the far left edge of each seismogram. It is generally a flat, or very nearly flat, line. The record of ground motion should be obvious, in contrast. However, sometimes the first P arrival is weak, so look carefully to determine the exact point at which the trace behavior deviates from its normal at-rest behavior. Then drag the arrow so that the point (and the faint vertical line) aligns squarely with this initial trace deflection.
Depending on your system and browser, you may have trouble dragging the arrows. If you can't seem to get them to drag, try this: instead of clicking on the arrows, try clicking off to either side of them. There are invisible "handles" there that should let you drag the arrows normally. This may take some getting used to. If you find you cannot "let go" of an arrow, try clicking and releasing on the arrow or the invisible "handle." This should "drop" the arrow.
When you have picked the arrival times for the P wave on every trace,
you still have one more arrival to pick before you can locate the
earthquake: an S-wave arrival. How to pick
this arrival is outlined in the entries below.
Once this button has been clicked and the blue arrow appears, the button should then read "Remove the S Arrival." Clicking on this same button again will cause the blue arrow to vanish. The arrow can be brought back and removed as many times as you like. Since there can only be one blue S-wave arrival arrow visible in the Action Frame at any time, you will need to remove an S-wave arrival arrow from one seismogram before you can reveal the arrow on a different one.
If any S-wave arrival arrow is visible (i.e.
this button has been clicked accordingly), the button labelled
"Find Origin Time and Epicenter"
will also be visible at the bottom of the page, below the last trace.
If you remove this S-wave arrival arrow, that
button will no longer be visible, because without an S-wave arrival
time chosen, the computer cannot calculate the origin time of the
earthquake.
To locate the earthquake you've chosen, you will need to pick an accurate S-wave arrival time. This can be very difficult to do, even for those with experience analyzing seismic data. This is the reason only one S arrival can be picked per event; asking you to pick the S-wave arrival on every trace would be asking too much. If you only need to pick one arrival, you can afford to be more selective, and try to pick the most obvious S-wave arrival in the bunch.
Recall that the arrival of the S wave on a waveform is most
obviously noted by a sudden increase in the amplitude
(deflection) of the trace oscillations. There may also be an
associated increase in the wavelength of the oscillations on
the waveform. To choose an S-wave arrival, look for the trace
on which these clues appear most evident, click the
"Pick an S Arrival" button there, and
then drag and place the blue S-wave arrow that appears. It
should drag in the same manner as the red P-wave arrows.
If you have trouble dragging it, remember the tips
we suggested to help.
Clicking this button will cause the Action Frame to
reload, bringing up Screen 3 -- a map with
travel-time circles superimposed on it.
In Screen 3, you will have the chance to mark
the location of the earthquake's epicenter on this map.
While Screen 3 is loading, the Time field
in the Data Frame will be updated. Whereas
before it had shown the time at which the left-hand edge of the
seismogram records began, it will now display the
origin time of the earthquake in Universal Time (UTC).
This origin time will be calculated automatically be the computer
based upon the trace on which you chose both P-wave and an S-wave
arrival times. Once calculated, this origin time is used to
compute the travel-time circles you will
see in Screen 3.
Once the initial values are placed in the data fields, only one will change during the course of your work on Screen 2. That value, the time, will change not only in a numerical sense, but in terms of its definition. When Screen 2 first loads, "Time" stands for the time at which the traces in the Action Frame begin -- the value of the time at their left-hand cut-off. This lets you know you're looking at records made across southern California starting at that time.
When you are done picking wave arrivals, and you click on the
button labelled "Find Origin Time and Epicenter,"
however, both the value and the definition of the
Time field will change. The field will
then give you the calculated origin time of the earthquake,
which may actually be earlier than the start time of the
seismograms. Once an initial origin time is calculated, it
doesn't matter how many times you switch back and forth between
Screens 2 and 3; the Time field will represent
origin time from then on... until you clear all the fields
by choosing a New Event with the menu bar.
The event number will not change while you are working with or
when you leave Screen 2, unless you choose a
New Event.
The date will not change while you are working with or when you leave
The number of traces will not change while you are working
with or when you leave Screen 2, unless you choose a
New Event.
When Screen 2 first loads, "Time" stands for the time at which the traces in the Action Frame begin -- the value of the time at their left-hand cut-off. This lets you know you're looking at records made by seismometers across southern California, starting at that arbitrarily chosen time (the instruments run continuously, but you can choose when to sample the data they gather).
Once you are satisfied with the wave-arrival picks you have made, and
click on the "Find Origin Time and Epicenter" button,
the computer will calculate an origin time for the earthquake, and
place this value into the Time field in the Data Frame.
From that point on, until a New Event is
chosen, the Time field will contain the most recently calculated
value for the earthquake's origin time.
One new option is functional once Screen 2 has first been loaded: "New Event." Though it is unlikely you will want to immediately clear the data from the earthquake you just selected from the list in Screen 1, you may do just that by clicking "New Event" on the menu bar. This will erase all the data fields and bring Screen 1 back into the Action Frame, so that you may choose a new earthquake from the list there.
If you have come back to Screen 2 from Screen 3, the "View Map" option is also a valid choice. This will take you to a background map of southern California overlain with travel-time circles with which you can assess the location of the earthquake's epicenter. This can only be done once the computer has calculated these circles, however, so until you have clicked on "Find Origin Time and Epicenter" at least once, "View Map" is not an option.
Ideally, all the circles on the map should intersect at a single point, if you chose the wave-arrival times correctly. In reality, you can expect a single-point intersection for three circles; if there are more than three circles on the map, they aren't likely to ever intersect in exactly one point, no matter how well you do in picking the P-wave and S-wave arrivals on the seismograms.
If the circles on the map don't intersect very closely,
or don't intersect at all, you should probably go back and
check the traces in Screen 2.
Revise your picks if necessary, and then click the
"Find Origin Time and Epicenter" button
to have the computer calculate new travel-time circles (as
well as a new origin time.
If you would like to reconsider the location you selected, just click on the "Unlock Epicenter" button to release the crosshairs, so that you can drag them again. This will also erase the entry in the Epicenter field of the Data Frame.
Should you decide to review your arrival picks in Screen 2
by using the "Check Traces" option on the menu bar,
the old epicentral location will be marked on the
background map when you return to Screen 3,
but it will be marked with a set of "shadow" crosshairs.
They will appear yellowish and cannot be dragged, as can the
real crosshairs. This "shadow" crosshairs is put there to
remind you of exactly where you last located the epicenter,
before the arrival time revisions. Locking the real crosshairs
will cause the "shadow" set to disappear; a new
epicentral location
will appear in the Data Frame.
Should you exit Screen 3 with the crosshairs locked,
a set of "shadow" crosshairs will appear on the
map when you return. If this button is
in unlocked mode (i.e. it reads ""Lock Epicenter"), however,
no such record of the previous location of the crosshairs
will be kept. Clicking this button to lock the active set of
crosshairs will make the "shadow" set
disappear.
You can clear this field by unlocking the crosshairs with
the "Unlock Epicenter" button. If you
switch back to the Screen 2 to "Check Frames"
without the epicenter locked, however, the
"shadow" crosshairs will not exist on
the map when you return.
"Help" still functions as it did in Screen 1 and Screen 2. To return to Screen 3 from the Help page, click the "View Map" option on the menu bar.
"Check Traces" allows you to go back to Screen 2 to reassess, and if necessary revise, your wave-arrival picks. If you have locked the epicenter before you leave Screen 3 to check the traces, that location will remain in the Epicenter field of the Data Frame, and a set of "shadow" crosshairs will show up on the background map when you return to Screen 3. Otherwise, no data about the epicenter will be retained when you go back to Screen 2 to "Check Traces."
If you are satisfied with the location of the epicenter as you have it marked and locked on the map, you should write down the time and location you found for that earthquake, and, if you'd like to locate another earthquake, click on the "New Event" option on the menu bar. This will erase all the fields in the Data Frame, and bring up Screen 1, with the earthquake list in the Action Frame. You can then start the locating process all over again with a new data set.
If instead you've located your last earthquake for the time being,
simply click on the Quit button to go back
and answer the review questions at the end of the
Activity #3 introduction.
Click here to load
the online activity now!
Calculating the origin time for an earthquake is accomplished using a single seismogram for which both a P-wave and an S-wave arrival time have been picked. In this example, let's say that seismogram was recorded by station ISA (Isabella). Because the seismograms are shown at a known, fixed resolution -- 10 pixels per second of elapsed time (for most; 5 pixels per second for earthquakes given in "compressed" time) -- the number of pixels between the two arrow layers on the seismogram can be translated directly into a difference in time of arrival. For instance, a 100-pixel gap between the arrows would equate to a 10 second S-P travel time.
Using the conversion factor of
On page 11 of this section you saw a P-wave velocity profile, illustrating that the velocity of seismic waves is by no means uniform through different parts of the Earth's interior. However, because of the simplicity of this activity, we chose a "zero-dimensional" model for seismic wave velocities. That means the computer making the calculations for origin time and travel-time circles treats both the P-wave and S-wave velocities as constants, independent of location. This limits the accuracy of the calculations, but it makes them much simpler and faster. The P-wave velocity used in the calculations is 6.3 kilometers per second; the S-wave velocity used is 3.7 kilometers per second.
Because of this, travel-time circles will sometimes be inaccurately sized. If a line drawn between the earthquake source and a given station goes primarily through sedimentary basins, for instance, the circle drawn for that station will likely be too large, because the computer will overestimate the distance based upon the travel time. Similarly, if the line between station and source runs primarily through mountains, the circle centered on that station may be too small. The effect will usually be most pronounced for stations located in low-velocity sedimentary basins, however. "Slow" velocities through sediments fall farther below the average velocities used than "fast" velocities through hard rock rise above them.
Using the conversion factor, then, the computer solves for a distance from the source to station ISA. Using the figures given above, that distance is 89 kilometers.
The P-wave arrival must have reached station ISA in the time it takes a P wave to travel 89 kilometers. The computer calculates this time by dividing the distance by the P-wave velocity, and comes up with an answer of 14.1 seconds. Now it is possible to find the origin time of the earthquake, because it has been determined that the earthquake must have occurred 14.1 before the P wave first arrived at station ISA.
But to make this origin time calculation, the computer first needs a frame of reference. For this it uses the initial value given in the "Time" field of the Data Frame: the time corresponding to the left-hand edge of the seismograms. Let's assume this time is 05:54:20.0 UTC. From this, you can convert the distance between the P-wave arrow and the left-hand edge into a time. Assume the distance is 80 pixels; the time the P wave arrived at station ISA would be 8.0 seconds past 05:54:20.0 UTC, or 05:54:28.0 UTC.
Now it is possible to calculate the absolute origin time. The computer subtracts 14.1 seconds from the time of the P-wave arrival (05:54:28.0 UTC) to arrive at an origin time for this earthquake of 05:54:13.9 UTC.
Once the computer "knows" the origin time of the earthquake, it only needs P-wave arrival time and the locations of the stations to plot the travel-time circles.
The computer already has found the distance from station ISA to the hypocenter: 89 kilometers. Now it needs to convert that to an epicentral distance, based upon the depth of the hypocenter. For this activity, all earthquakes are assumed to occur at a depth of 6.0 kilometers. Solving for the long leg of a right triangle with a short leg of 6.0 kilometers and a hypotenuse of 89 kilometers will yield the "map distance," the distance to the source when projected to the surface (as on a map). That map distance is 88.8 kilometers. You can see then that it's not entirely crucial for the computer to "know" the exact depth of the earthquake; a 6.0-kilometer difference in depth translates to only a 0.2-kilometer difference in map distance.
To draw a travel-time circle of the appropriate size, the computer must convert kilometers to pixels, given the scale of the map atop which the circles will plot. That conversion factor is 1.05 pixels per kilometer. Thus, the travel-time circle centered on station ISA, the map coordinates of which are read by the computer from a list, would need to have a radius of 93 pixels. A script writes the circle dynamically, scaling it correctly. The first travel-time circle has been plotted.
The trace from station ISA had two arrival picks on it. Can a travel-time circle be created from a trace on which just a single arrival time is marked? The answer is yes; now that the origin time of the earthquake has been calculated, all the computer needs to calculate a travel-time circle for each trace is the arrival time of the P wave. Assume that station GSC (Goldstone) also recorded this earthquake. If the P-wave arrival arrow you placed on its trace is 115 pixels from the left-hand edge, that means station GSC first experienced the P wave at 11.5 seconds after 05:54:20.0 UTC, or 05:54:31.5 UTC. Since the earthquake's origin time is 05:54:13.9 UTC, this means it took 17.6 seconds for the P wave to reach Goldstone. At a velocity of 6.3 km/sec, that translates to 111 kilometers. Again, a travel-time circle is calculated based upon this distance, the depth of 6.0 kilometers, and the conversion factors to turn kilometers into map pixels. This circle, 116 pixels in radius, is then centered on the coordinates for station GSC.
Similar calculations are performed using the arrival picks from every station until all the travel-time circles have been drawn. Then it's up to you to find the most likely epicenter by studying the intersection of these circles.
That's how it's all done. Hopefully you followed this walk-through without getting confused. It's a long and tedious process (fortunately the computer does it for you!), but fundamentally, it's pretty simple.
Return to the first
page of Activity #3