Skip to main content
U.S. flag

An official website of the United States government

What causes earthquake swarms at Yellowstone?

Earthquake swarms are common at Yellowstone, but why do they occur?  Are they driven by magma migration?  Water?  Steady creep along faults?  All three are possibilities, and tracking the style of the earthquakes can reveal the causes.

Yellowstone Caldera Chronicles is a weekly column written by scientists and collaborators of the Yellowstone Volcano Observatory. This week's contribution is from David Shelly, seismologist with the U.S. Geological Survey.

Earthquake swarms—sequences of elevated earthquake activity with no clear mainshock—are common at Yellowstone and many other places.  Swarms occur in a variety of volcanic and tectonic settings and have several possible causes.   Some swarms are driven by slow fault slip that causes earthquakes on few sticky patches of the fault.  Other swarms are generated when magma-filled cracks push their way through the crust.  And perhaps most commonly, swarms can be generated when aqueous fluids (water) enter and interact with pre-existing faults in the crust.  Sometimes, combinations of these mechanisms might be active in a given swarm. 

image related to volcanoes. See description
Evolution of the 2017 Maple Creek earthquake swarm. Plots show earthquake locations colored by time. a) Map view. b) West-east cross-section. c) Three-dimensional view, looking from the east-southeast, along the axis for much of the swarm activity.

But how can we distinguish among these processes?  Such a determination is not simple and remains an important topic of ongoing research, but we can examine several aspects of swarms to guide our interpretations. 

At a volcano, everyone wants to know, “is the swarm driven by magma?”   This question is particularly pressing if the swarm is occurring at shallow depths (within a couple miles of the surface), because a shallow, magmatically driven swarm could potentially be a precursor to an eruption.  

A primary distinction between magma and water as a driving mechanism is the width of the crack required for the fluid to move through the crust.  While water can travel through very small preexisting cracks or faults within the crust, magma requires a much thicker crack to allow the magma to continue to propagate without quickly cooling and solidifying.  

We can potentially observe several properties of the seismicity related to this difference in crack widths.  For example, earthquakes triggered by water will usually occur as fault slip on the crack(s)/fault(s) along which the fluids are propagating.  That is, the two sides of the crack hosting the fluids are still in contact with each other, but the fluid reduces the clamping forces and lubricates the fault enough to allow it to slip. 

In contrast, the walls of a crack hosting magma are generally not in direct contact with each other—they are separated by the magma itself.  So, rather than occurring by slip between the walls of the crack, earthquakes will instead occur near the crack tip (the point ahead of the magma where the crack is starting to open) or off to the sides of the crack (because the opening of the crack stresses the surrounding rock).   

In the case where a swarm is caused by slow fault slip, earthquakes repeatedly occur on the same small sticky patches within a dominantly creeping fault.  There might be no progression of seismicity at all—just the same small patches constantly generating earthquakes.

Cartoon showing differences between magma- and water-driven earthquake swarms
Simplified cartoon showing differences between patterns of earthquakes expected for a swarm driven by a magma-filled crack (left) versus one driven by pressurized water in a fault (right).

In addition to seismic observations and the pattern of earthquake locations, we can also use deformation measurements to examine changes in the shape of the surface above the earthquake swarm.  Opening of thick cracks required for magma propagation creates a warping of the earth’s surface, which becomes increasingly large and easy to observe for the shallow cracks that would be of most concern.  Swarms driven by aqueous fluids pressurizing preexisting cracks, in contrast, would cause only very small surface warping, mostly due to slip in the cracks themselves, and is usually too small to be observed.  Slow fault slip, if large and shallow enough, can be observed on the surface, but the pattern of deformation is very distinct from that caused by a magma-filled crack and therefore easy to distinguish.

When interpreting the process(es) that might be causing an earthquake swarm, it is also important to consider the context.  At Yellowstone, for example, the last magmatic eruption was a lava flow that occurred about 70,000 years ago, but the area is home to one of the most vigorous hydrothermal systems on Earth. So even though eruptions are rare, we observe many small earthquake swarms at Yellowstone every year, and relatively large earthquake swarms every few years.  The characteristics of the swarms, and their context, indicate that the vast majority are driven by water moving through the subsurface. 

As an example of putting this all together let’s consider the June-September 2017 Maple Creek swarm—the second-largest earthquake swarm ever recorded in Yellowstone!  About 2400 earthquakes, the largest of which was M4.4, were located over about 3 months in the northwest part of Yellowstone National Park, between Norris Geyser Basin and Hebgen Lake.  The earthquakes moved around over time but were occurring in an area that hosts numerous existing faults.  The migration of the seismicity was rapid, and no ground deformation occurred.  Observations like these suggest that water was the cause of the swarm.  Water was also the most likely cause of an intense swarm that occurred near Madison Junction, in the western part of Yellowstone National Park, in 2010.

A more complex case was the 2008-2009 Yellowstone Lake swarm.  Rapid migration of earthquakes in this swarm suggest it might have been driven by a low-viscosity fluid (such as water or CO2) that could move easily through small cracks in the subsurface.  But unlike other recent swarms in the park, this swarm was accompanied by a small amount of observed surface deformation, making it difficult to completely rule out a magmatic source.  Fortunately, with improvements in the Yellowstone monitoring network since 2009, we could learn even more from similar swarms that might occur in the future.

Each Yellowstone earthquake swarm is unique, lasting minutes to weeks and including a few to thousands of earthquakes that might or might not move around over time. With long-term investments in seismic and deformation monitoring, we continue to learn more about the Yellowstone system by studying this common form of seismicity!

Related Content