microseisms The first thing you find after overcoming the technical problems of building and installing an amateur seismograph is that it constantly records a low-amplitude signal with a period of between 3 and 6 seconds (at least that's what I find in the UK: elsewhere it may be longer). This is microseismic noise as shown in the image to the right (using a fairly coarse ADC). Although these are a nuisance because they prevent detection of faint seismic waves from earthquakes, they are also useful in checking that the seismograph is functioning. Professional seismologists have also obtained useful information from them.

My Lehman seismograph has been running now for over a year. In that time it has detected more than one hundred earthquakes, most of them very far from my house. One of the clearest sets of signals is the following:

Japan earthquake seismogram This is the magnitude 7.0 quake that occurred on Sunday, July 10, 2011 at 00:57:12 UTC offshore near Japan. That is 9450 Km from my house. The display shows one hour of data in five successive 12-minute bands from top to bottom, a so-called helicorder plot, although it is no longer helical. :-) The P-waves (top line), S-waves (second from top line), and surface waves (fourth and fifth lines) are clearly visible.

That's a most rewarding signal to have recorded, but by no means the furthest away that I have detected.

You may be wondering how I know what that event was. Thanks to the miracle of the Web, I sometimes know about large earthquakes even before their seismic waves have reached my detector. The principal sources of this information are the American USGS and the European EMSC. Both of these list many quakes down to magnitudes that are too small for me to detect. The USGS has a particularly useful page that allows you to calculate arrival times of the waves at your seismometer. In my experience this is very accurate. They also predict the value of the ground displacement and velocity at your site, and I plan to investigate the use of this as a means of calibration.

Channel quake Here's another interesting event: the magnitude 3.9 earthquake in the English Channel on 14 July 2011, about 150 Km away from me, a rare event. This was originally published as M4.3, but, as often happens, it has been downgraded. It looks very different from a teleseismic signal like the Japanese example above: obviously it's much shorter in time, and it has higher-frequency components. This is the smallest magnitude earthquake that I have detected so far.

detectability curve This raises a very interesting question: what magnitude of earthquake can I detect and how much does that vary with distance across the world? The first thing to say, perhaps, is that everyone knows about the 'Richter scale' but in fact no-one uses it. Most quakes reported by USGS or EMSC use the Mw scale: moment magnitude. There are other scales too. However, for the amateur these are all very similar and all are logarithmic. The British Geological Society has an interesting page that shows how to calculate magnitude from signal size and earthquake distance. This can be used in reverse: for a given seismometer signal sensitivity, how does detectable magnitude vary with distance? The graph on the right shows the result for a device with a sensitivity of 1 micrometre/second. Earthquakes that fall on or above the line are detectable. Even at a distance of 20,000 Km, magnitude 6.2 events can be detected. Closer to the seismograph, smaller magnitudes are detectable. However there is a geographic effect: the seismograph is strongly directional and the nominal sensitivity must be multiplied by the necessary trigonometric factor (a value between zero and one) for each earthquake's position. Fortunately, earthquakes have both longitudinal and transverse waves, so the seismograph can not be blind to both.

hidden earthquake

As I mentioned in the analysis page, the Fast Fourier Transform (FFT) is invaluable for digging out seismic signals that are hidden by the microseismic background. For example, the image to the right shows an earthquake, but it is far from obvious:

revealed earthquake

You could easily pass over such an image, but the one on the right shows exactly the same data, but with an FFT low-pass filter used to remove the microseismic background.





Back To Top




Valid XHTML 1.0 Strict