Not all ground-motion sensors are created equal
October 29, 2019

You may be surprised at the diversity of sensors that are able to detect seismic waves. Everything from the accelerometer in your cell phone to multi-million dollar satellites orbiting at altitudes of over 12,000 miles are capable of recording seismic vibrations. Surfers who visit the tsunami-prone Mentawai islands in Sumatra often invert an empty glass bottle on the floor before retiring, creating a very rudimentary ground-motion sensor in the absence of cell-phone coverage in that remote Indonesian region! This article will focus on three of the most ubiquitous seismic sensors in use today: broadband seismometers, strong-motion seismometers, and the Global Navigation Satellite System (GNSS).

Not all seismometers are created equal because not all ground motions are the same. When a seismic observatory such as the center records a ground motion, that recording may be processed in a variety of ways to achieve different purposes, and different types of seismometers excel at different tasks. The workhorse sensor for the center is the broadband seismometer. These extremely sensitive devices record seismic energy across a large range of frequencies (thus the “broadband” term). They usually record seismic velocity on three channels: up-down, east-west, and north-south. This allows them to capture the complexity of seismic waves as they move through the three-dimensional earth. These sensors excel at “listening” to the earth for weak motion—very small ground motions that are induced by small local or regional earthquakes, or very distant earthquakes. When the center’s website reports, for example, a magnitude 1.1 earthquake in the Brooks Range, that detection was made possible by our network of broadband seismometers. Broadbands are also very astute at recording seismic noise. This may not seem very useful, but this ambient, or background, noise is special. It is the “hum of the earth,” induced primarily by wave action at the coasts, and seismologists use it to study the geological structure of the rocky crust. Broadband seismometers form the backbone of the center’s monitoring capability.

From the above description, it may seem that broadbands are the only sensors that a seismologist would need. However, they do have a crucial limitation. Because they are optimized for recording faint, weak-motion signals, they do not do well in the large ground motions induced by strong or nearby earthquakes. Strong shaking will cause broadband seismometers to “clip”, or go off scale (see Figure 1). This phenomenon is similar to what happens if a stereo sound-system is too loud for its speakers, causing them to buzz and not properly reproduce the music. This is unfortunate, because it means that broadband sensors perform poorly, at least locally, at recording big earthquakes, and big earthquakes are the most societally impactful seismic events. After all, the definition of strong motion is ground motion that impacts people or their environment. Fortunately, the center operates another class of sensor called a strong-motion seismometer. As the name implies, while these do not excel at detecting the very faint seismic signals that are captured by broadbands, they have a large dynamic range, and can record even very large ground motions. This makes them extremely well-suited to engineering applications; dense networks of these sensors will often be located in urban environments such as Anchorage (see Figure 2) or around critical infrastructure such as power plants and the Trans Alaska Pipeline. In fact, although there is overlap, you can roughly think of broadbands seismometers as used for scientific purposes, while strong-motion sensors are for engineering. 

To get an idea of the huge dynamic range of strong-motion sensors, they are able to record accelerations up to 2 times the earth’s acceleration due to gravity, or 2g. Items will often be ejected from shelves in buildings experiencing just 1g! By contrast, if you stand a few feet from a broadband seismometer and stomp the ground, that instrument is likely to clip the signal. Following a significant earthquake, strong-motion sensors provide critical information to engineers such as if a structure experienced an acceleration large enough to have caused damage. Historical strong-motion records compiled over a period of years are used by engineers to design structures that can withstand likely accelerations in an area. At the center, we use strong-motion records as input into a model, called a Shakemap (click here to see our Shakemap page), that is able to estimate shaking intensity over a large region.

The final class of sensor is a relative newcomer to the science of seismology, but for smartphone users, it may be the most familiar. The GNSS, of which the Global Positioning System (GPS) is the US version, consists of a constellation of satellites in medium earth orbit (around 12,000 miles in height) that transmit a very accurate clock signal that can be decoded to obtain a precise 3D estimate of position. GNSS was originally designed for military applications, but now is used for everything from guiding bombs to finding the nearest coffee shop. In the earth sciences, GNSS has been used to examine how a precise position on the earth moves over a period of years to decades to study plate tectonics. GNSS receivers located near plate boundaries move at higher rates than receivers in continental interiors. This is because tectonic plates deform as they grind against each other at the boundaries. Improvements in technology now allow GNSS to record positions with much less than one inch of accuracy, and for the positions to be processed in near real time. This level of sensitivity allows for the recording of ground motions at a GNSS receiver for larger earthquakes, turning the GPS receiver into a space-borne seismometer. 

Using satellites to record earthquake ground motions is undoubtedly novel, but why is this useful when we already have access to broadband and strong-motion seismometers? Particularly for large earthquakes, GNSS seismometers have some notable advantages over traditional broadband or strong-motion seismometers. They do not clip in the presence of large ground motions, so like a strong-motion seismometer, they are useful for recording large, nearby earthquakes. However, the most unique characteristic of GNSS seismometers is that they record the full spectrum of earthquake ground motion.

Earthquakes cause the ground to move in two distinct ways. Seismic waves cause the earth to vibrate in an elastic manner. That means that although the earth moves as the seismic wave passes, it returns to its original position, much like an elastic rubber band will snap back to position after an external stress, such as a pulling finger, is removed. All three types of sensors—broadband, strong motion, and GNSS—record this type of elastic ground motion. What is unique about GNSS sensors is that they also record the non-elastic ground motion that occurs during a large earthquake.

Earthquakes occur as a result of permanent motion on a fault. This motion actually travels some distance from the fault at the same speed as the seismic S-wave (see types of seismic waves) and is recorded at nearby GNSS receivers as a one-time, permanent movement called a static offset (see Figure 3). This is extremely useful, because the distance a GNSS sensor travels as a result of the static offset is proportional to the size of the earthquake. To compute the magnitude of a large earthquake, scientists essentially use the Pythagorean distance formula that you may have learned in high school to find the distance the GNSS sensor moved, and then compare that distance to historical events to estimate the magnitude. This technique is not only very simple, but it can be done rapidly. In contrast, traditional seismometers take much longer to estimate the size of a large earthquake, because it takes time for the seismic waves to reach sensors that are far enough away to avoid clipping, and it also takes time to record the low-frequency repeating seismic motion that indicates a large earthquake. 

Each type of seismometer we have discussed has its strengths and weaknesses, which is why we utilize a variety at the center, and why we are looking forward to incorporating new technologies such as GNSS. The more types of sensors we have, the better we can characterize the different components of earthquake-induced shaking, enhancing our monitoring mission of providing timely and accurate earthquake information to the public.

Figure 1: Seismic station RC01 contains both a broadband seismometer and a strong-motion seismometer, which allows us to compare the performance of these different classes of sensors. In the presence of the strong ground motion induced by the large M7.1 Anchorage earthquake, the broadband sensor clips, causing it to look very different from the accurately recorded signal on the strong-motion sensor. However, the lower two seismograms show that both instruments are able to accurately record the weak motion from a much smaller aftershock.

Figure 2: This image shows the locations of the sensors comprising the Anchorage strong-motion network and the recording that network made of the November 30 earthquake. Strong-motion sensors are most often located in urban environments and around high-value infrastructure. They can also show details such as shaking variability across an area. Notice how the shaking in some parts of Anchorage exceeded 40% of gravity and others experienced much less shaking. This aligns with areas more heavily damaged during the earthquake

Figure 3: This image shows a comparison of recordings of the 2013 M7.5 Craig earthquake that struck southeast Alaska in 2013. The seismogram at the top shows the earthquake as recorded by a conventional broadband seismometer, while the seismogram at the bottom is from a high-rate GPS instrument at roughly the same distance from the epicenter. Note that the broadband recording exhibits an unnatural “spiky” signal; the result of the signal being “clipped” due to the instruments’ exposure to strong shaking. In contrast, the GPS has not been clipped, and has also accurately recorded the permanent component of the ground motion, often referred to as the “static offset.”

Figure 4: Seismic site KNK (Knik Glacier) has both a broadband (green instrument) and strong motion sensor (grey instrument).