Traffic Noise vs. Earthquakes

Events other than earthquakes are recorded by the ANZA seismic network. Traffic noise is often recorded on our sensitive broadband instruments, especially those located close to busy highways. One of our stations close to UCSD on Mt. Soledad (station SOL) is sensitive enough to pick up the vibrations of heavy traffic throughout the day Monday to Friday. The images below compare what a 'quiet' period, a 'noisy' period and an earthquake look like on the vertical (Z) component waveforms we record at the Mt. Soledad station.

Note: The times along the base of the waveform images are in UTC, which, during the summer, is 7 hours ahead of the west coast of the United States.

Waveforms recorded during a 'quiet' period

Waveforms recorded from approximately 01:50 AM to approximately 02:50AM on June 19, 2003 are illustrated below. This is a period when traffic close to station SOL would be at a minimum. The important points to notice are the amplitude of the waveforms (shown in the red box on the left-side of the image), which are a maximum of slightly greater than +6500 nm/sec and a minimum of slighty less than -2000 nm/sec. Most of the waveform amplitudes lie between +3000 nm/sec and +2000 nm/sec. The high amplitude peaks on this trace may be the result of a truck passing by.

Quiet Seismogram

Waveforms recorded during a 'noisy' period

Waveforms recorded from approximately 07:45AM to approximately 08:45AM on June 19, 2003 are illustrated below. This is a period when traffic close to the station would be heavy during the morning commute across Mt. Soledad, and would consequently cause vibrations in the ground, effecting what is recorded by our instruments at station SOL. Note the amplitude of the waveforms (again in the red box on the left-side of the image), which reach more than +30,000 nm/sec. This is over 4 times what is recorded during the quiet period shown above! Also, note that most of the waveform amplitudes lie between +8000 nm/sec and -2000 nm/sec which is a much greater range than during the quiet period. Our ability to accurately record waveforms generated by very small earthquakes at this station is slightly compromised during these noisy periods of the day. Note that large earthquakes would still appear clearly on the waveforms.

Rush hour image

Waveforms recorded over two days

To better understand the comparison between 'quiet' and 'noisy' periods of 'human-created' seismicity it helps to observe what is recorded over a prolonged period (image below).

[Click here to see a larger copy of this image]

48hrs image

The vertical red lines show midday and midnight, while the green areas represent approximate times of heavy commuting traffic. You will notice that in the early morning hours (02:00 AM to 05:00 AM) the waveforms recorded are very low amplitude. This reflects the low volume of traffic close to station SOL.

Waveforms recorded during a real (small) earthquake

So now that we understand the difference between waveforms recorded at 'quiet' and 'noisy' times, what do the waveforms look like when a real earthquake occurs? Waveforms recorded by the Mt. Soledad instrument during a recent earthquake near San Clemente Island on June 9th, 2003 (more information on this event) are illustrated below. The earthquake was small (a magnitude 3.03 event) and occurred at 06:32AM on a Saturday morning, when traffic volume would be small. You can see how the waveforms clearly record the earthquake even though the epicenter was more than 80 km (50 miles) away from the station! (The red 'P' flag was inserted by one of our analysts to show when the primary wave occurred). Note the amplitude of the waveforms (the red box on the left-side of the image) - the maximum amplitude was only just above +13000 nm/sec. If you compare this with the amplitudes measured during a noisy period such as the morning and evening commute times (> +30,000 nm/sec) you will see that this small event would be indistinguishable from the surrounding seismic noise generated by traffic!

Event image

Summary

Although instruments at stations like Mt. Soledad in the Anza network may have the recorded waveforms obscured by seismicity generated by heavy traffic, analysts never use just one station's records to determine if and when an earthquake occurred. We currently have 17 broadband stations in our Anza array and an event is registered (and considered as generated by an earthquake) only when 6 or more stations are triggered by a notable change in amplitude.

Telling the difference between seismicity caused by an earthquake and by noise such as heavy traffic is just one of the many skills our analysts have.

URL: http://eqinfo.ucsd.edu/faq/traffic.php [Last updated: 2015-10-22 (295) 23:07:44 UTC]