RESEARCH LETTER The 2017–2018 Maple Creek Earthquake Sequence

10.1029/2019GL082376
in Yellowstone National Park, USA
Key Points:
• Planar structures in the sequence
Guanning Pang1 , Keith D. Koper1 , J. Mark Hale1 , Relu Burlacu1, Jamie Farrell1 , and
align with the regional Quaternary Robert B. Smith1
fault system
1
• Low VP/VS values (1.50–1.66) imply Department of Geology and Geophysics, University of Utah, Salt Lake City, UT, USA
that CO2‐filled cracks exist in the
source region
• Many of the events are late Abstract We explore the detailed spatiotemporal evolution of 3,345 earthquakes that occurred near
aftershocks of the 1959 Mw 7.2
Hebgen Lake earthquake
Maple Creek, Yellowstone, for the time period of 12 June 2017 to 13 March 2018. We generate high‐
accuracy relocations and near source VP/VS ratios using 4.4 million P wave and S wave differential travel
Supporting Information: times derived from waveform cross correlation. The hypocenters can be subdivided geographically into two
• Supporting Information S1 major subpopulations: a northern cluster with planar structures striking mainly NW‐SE and a southern
cluster with planar structures striking mainly E‐W. We observe VP/VS ratios of 1.39–1.66 in the northern
cluster and a steady ratio of 1.50 in the southern cluster, suggesting the presence of CO2‐filled cracks. We
Correspondence to:
G. Pang,
interpret the northern earthquake cluster primarily as long‐lived aftershocks of the 1959 Mw 7.2 Hebgen
[email protected] Lake earthquake but with some influence of magmatic fluids. We interpret the southern earthquake cluster
as a more classic, swarm‐like sequence induced primarily by the migration of magmatic fluids.
Citation: Plain Language Summary Seismicity in the Yellowstone volcanic region dominantly consists of
Pang, G., Koper, K. D., Hale, J. M.,
Burlacu, R., Farrell, J., & Smith, R. B.
bursts of earthquakes, often referred to as swarms. The three largest swarms in modern times—in 1985,
(2019). The 2017–2018 Maple Creek 2008–2009, and 2010—were triggered by magmatic fluid flow. A fourth major burst of Yellowstone
earthquake sequence in Yellowstone earthquakes occurred in 2017–2018 near Maple Creek and consisted of over 3,000 events. We determined
National Park, USA. Geophysical
Research Letters, 46, 4653–4663. https://
accurate locations for these earthquakes and measured the ratio of P to S velocity in the source region. We
doi.org/10.1029/2019GL082376 find that the 2017–2018 Maple Creek sequence differs from the three previous large swarms in that many of
the earthquakes can best be categorized as late aftershocks of the 1959 Mw 7.2 Hebgen Lake earthquake.
Received 5 FEB 2019
Accepted 18 APR 2019
Accepted article online 30 APR 2019
Published online 13 MAY 2019 1. Introduction
The Yellowstone volcanic system is a dynamic geologic environment with active ground deformation,
frequent seismicity, extremely high heat‐flow, and an extensive hydrothermal system (Smith et al., 2009).
It has had three caldera‐forming eruptions in the last 2.1 million years and has some of the highest
earthquake hazard in the western United States (Petersen et al., 2014; White et al., 2009). Owing to these
factors, the Yellowstone region is monitored with a variety of technologies, including a permanent seismic
network, continuous global positioning system (GPS) receivers, campaign GPS and gravity surveys, stream
gauges, chlorine‐flux measurements, airborne gas sampling, and airborne infrared surveys (Yellowstone
Volcano Observatory, 2006).
Seismicity in Yellowstone dominantly consists of swarms of earthquakes. Farrell et al. (2009) examined
interevent times of Yellowstone earthquakes during 1984–2006 and identified 239 distinct swarms, which
together accounted for 51% of the 23,054 cataloged earthquakes in the region. In that study, swarms were
defined based on criteria originally developed by Mogi (1963) to categorize his Type III (swarm) sequences.
The key criterion is that the maximum daily seismicity rate (Nmax) initially grows with the sequence
duration time (T, in days), such that Nmax > (4 T)1/2. The three most active Yellowstone swarms in historic
times—in 1985 (Waite & Smith, 2002), 2008–2009 (Farrell et al., 2010), and 2010 (Shelly et al., 2013)—
occurred within or adjacent to the youngest (0.63 Ma) caldera, which covers an area of ~2,800 km2
(Figure 1). These swarms had rapid migration north/northwest from their initial location and have been
interpreted to result from the flow of magmatic fluids.
On 12 June 2017, the fourth major swarm‐like Yellowstone earthquake sequence in the instrumental era
(since 1972) began about 5–10 km east of Hebgen Lake near Maple Creek (Figure 1, red dots). The
©2019. American Geophysical Union.
earthquakes persisted through November 2017 in a rectangular area of ~9 km by 12 km (supporting
All Rights Reserved. information Figure S1). In February 2018, after ~2 months of low seismicity, the sequence briefly became

PANG ET AL. 4653

 19448007, 2019, 9, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019GL082376 by Cochrane Colombia, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Geophysical Research Letters 10.1029/2019GL082376

Figure 1. Yellowstone seismicity and the Yellowstone seismic network. Dots are earthquakes with magnitude ≥1.5,
triangles are seismic stations, and yellow triangles are the stations used in waveform‐based relative relocation and
VP/VS analysis. Black dots show earthquakes since 1980, with cyan dots for the 1985 swarm, purple dots for the
2008–2009 swarm, green dots for the 2010 swarm, red dots for the 2017 portion of Maple Creek sequence, and blue
dots for the 2018 portion of Maple Creek sequence. The nearby station WY.YMC is marked. The red beachball is the
focal mechanism of the largest event in the Maple Creek sequence (Mw 4.4 on 16 June 2017), and the black
beachball is the focal mechanism of the 1959 Mw 7.2 Hebgen Lake earthquake. The blue line shows the boundary of
the 0.64‐Ma caldera, and the purple line shows the boundary for the 2.1‐Ma caldera (Christiansen et al., 2007). The
red line in the inset shows the boundary of Yellowstone National Park.

active again a few kilometers south‐southeast of the 2017 region. We refer to the combined earthquakes
between 12 June 2017 and 13 March 2018 as the Maple Creek sequence. This sequence differs from the
Yellowstone swarms of 1985, 2008–2009, and 2010 in that it is well outside of the youngest (0.63 Ma)
caldera and is adjacent to the source region of the 1959 Mw 7.2 Hebgen Lake earthquake (Doser, 1985) in
an area of increased Coulomb failure stress (Chang & Smith, 2002; Figure S2).
In this study, we compute and interpret high‐resolution earthquake locations for the 2017–2018 Maple
Creek sequence using data from the Yellowstone seismic network. We compute single‐event locations with
a regional 1‐D velocity model and use these absolute locations as starting points in a multievent, relative
relocation process known as GrowClust (Trugman & Shearer, 2017). We also estimate near‐source VP/VS
ratios from the cross correlation‐derived differential travel times measured for the relative relocation. We
examine the space–time evolution of the seismicity and the VP/VS ratios to constrain the mechanism that
drove the 2017–2018 Maple Creek earthquake sequence.

2. Absolute Earthquake Locations and Magnitudes

Earthquakes in the Yellowstone region are routinely detected and located by the University of Utah
Seismograph Stations using HYPOINVERSE (Klein, 2002) with a 1‐D velocity model (Table S1). Owing to
the high density of seismometers in the region, the absolute locations of the 3,345 events in the Maple
Creek sequence have low epicentral uncertainties and well‐determined focal depths. The median horizontal
standard error (ehr) is 500 m, and the median standard error in depth (ehz) is 730 m (Figure S3). Over 97% of
the events are located at depths of 6–14 km beneath the surface (mean elevation of 2.3 km) in a region of
relatively high VP, which is overlain by a region of relatively low VP that has been interpreted as a CO2
gas‐saturated body (Husen et al., 2004). Coda duration magnitudes (MC) were calculated for all the events

PANG ET AL. 4654

 19448007, 2019, 9, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019GL082376 by Cochrane Colombia, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Geophysical Research Letters 10.1029/2019GL082376

in the sequence and varied from −1.7 to 4.4. We estimate a magnitude of completeness of MC 0.5 and a b
value of 0.96 from least squares fitting of the magnitude‐frequency data between MC 0.5 and 3.5 (Figure S1).
Only one earthquake was large enough that its regional waveforms could be inverted for a moment tensor
(16 June 2017 00:48:47, 44.7813°N, 111.033°W, 11.6 km, ML 4.36). Using the methodology of Minson and
Dreger (2008), and assuming a deviatoric source mechanism, we obtained an Mw 4.4 oblique‐normal solu-
tion (Figure S4) with one nodal plane dipping 62° to the northeast and striking to the northwest at 327°,
which is subparallel to the 1959 Mw 7.2 Hebgen Lake fault scarp (Figure 2). The northeast‐southwest
oriented T axis is consistent with the GPS‐determined regional strain field as well as historic earthquakes
(Payne et al., 2012; Puskas et al., 2007; Schmeelk et al., 2017). The best fitting moment centroid depth is
15 km, ~6 km deeper than the focal depth derived from arrival times. This difference may result from the
difference in velocity models used in the two procedures, although the centroid depth is relatively poorly
constrained. We also attempted to invert the waveforms for a full moment tensor with six degrees of free-
dom. We found an isotropic component accounting for 33% of the moment; however, the F test statistic
for the isotropic component is significant at a confidence level of only 57%, much lower than the 95% value
that is used to accept an apparent isotropic component as genuine (e.g., Whidden & Pankow, 2012).

3. High‐Precision Relative Relocation of Earthquakes

To explore small‐scale structures in the seismicity, we measured ~4.4 million high‐quality waveform‐based
differential travel times (27% S and 73% P) with a bispectrum cross‐correlation package, BCSEIS (Du et al.,
2004), from 22 local stations and 2 regional stations (Figure 1, yellow stars) that were selected based on data
quality and azimuthal coverage. We inverted the differential times for relative locations with the hierarch-
ical clustering relocation algorithm GrowClust (Trugman & Shearer, 2017). Before starting relocation,
GrowClust fixes the centroid of the entire sequence and sorts event pairs by similarity coefficients obtained
from waveform cross correlation. The event pairs are relocated sequentially with a grid search, starting with
the most similar pair. When a new event is paired with previously relocated events, it will remain at its ori-
ginal location if the overall residual does not decrease during the grid search. With our high‐quality P wave
and S wave differential travel times, we were able to relocate 3,257 out of 3,345 events with GrowClust.
Because of the proximity of station YMC, and the large number of differential S‐P times, we obtained
high‐precision relative relocations with low uncertainties. Using 100 bootstrap resampling runs, we estimate
a median vertical error of 86 m and a median horizontal error of 62 m (Figure S3). The true uncertainties
may be slightly higher because of simplifications made in the forward problem such as the use of a 1‐D iso-
tropic velocity model.

3.1. Spatial Structure and Geometry of Relocated Earthquakes

Based on geographic and temporal considerations, the relocated hypocenters can be divided into two major
sequences. The northern sequence, which we refer to as Cluster n, occurred exclusively in 2017 to the NNW
of station YMC (Figure 2). It forms an oblique V‐shaped pattern with a NW‐SE strike similar to that of the
Hebgen Lake fault scarp and the northeast dipping nodal plane of the Mw 4.4 earthquake. On a smaller scale,
Cluster n also contains orthogonal planar features that strike NE–SW, similar to the strike of the southeast‐
dipping nodal plane of the Mw 4.4 earthquake. Based on finer‐scale features, we divided Cluster n into three
smaller clusters (Figure 2): n‐I, containing the earliest events in 2017, which expanded outward from the
sequence origin near the Mw 4.4 epicenter; n‐II, containing the more scattered events to the NNW that
showed northwest‐to‐southeast migration; and n‐III, containing the less scattered events in the eastern part
of the sequence with northwest‐to‐southeast migration. We used principal component analysis (PCA) on the
spatial covariance matrices of the clusters to quantify their geometrical structure (e.g., Michelini & Bolt,
1986). Our PCA implementation returns: (1) a scalar measure of planarity that varies from 0 for a perfectly
spherical distribution to 1 for a perfectly planar distribution, (2) the orientation of the optimal plane in a least
squares sense, and (3) bootstrap‐derived uncertainties for the orientation of the optimal plane (Pang et al.,
2018). All three clusters are described by planes striking NW‐SE and dipping steeply to the southwest
(Figure 2 and Table S2), but only Cluster n‐I has a relatively high planarity (0.89) indicating a robust struc-
ture. The lower planarities of Clusters n‐II and n‐III (0.77 and 0.78) are caused by smaller‐scale features (e.g.,
NE–SW striking planes) that combine to give artificially steep dips when fit together as a single plane.

PANG ET AL. 4655

 19448007, 2019, 9, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019GL082376 by Cochrane Colombia, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Geophysical Research Letters 10.1029/2019GL082376

Figure 2. (a) Relocated earthquakes in the 2017–2018 Maple Creek sequence. The relocations were computed using ~4.4
million P and S differential travel times. The red colors mark Clusters n‐I, n‐II, and n‐III. Dark green circles mark Cluster
s‐I, which occurred in 2017. Blue, orange, purple, and black circles mark Clusters s‐II, s‐III, s‐IV and s‐V, which all
occurred in 2018. Best fitting planes for each cluster are shown in matching colors, with uncertainties determined from
500 bootstrap resamples. Gray circles are the remaining earthquakes that were not assigned into a subcluster. The
beachball is the focal mechanism of the largest (Mw 4.4) earthquake in the sequence. The red line is the surface rupture of
the 1959 Mw 7.2 Hebgen Lake earthquake (Johnson et al., 2018), black lines show regional faults, and the purple line is the
boundary of the 2.1‐Ma caldera. (b) A‐A′ cross section with depth relative to sea level. (c) B‐B′ cross section with depth
relative to sea level.

The southern portion of the 2017–2018 Maple Creek sequence, which we refer to as Cluster s, occurred
mostly in 2018 to the SSE of station YMC (Figure 2). Based on origin times, dip direction, and epicentral dis-
tribution, we divided Cluster s into five smaller clusters (s‐I, s‐II, s‐III, s‐IV, and s‐V) and used PCA to quan-
tify their geometrical properties (Figure 2 and Table S2). These clusters were well described by planes—four

PANG ET AL. 4656

 19448007, 2019, 9, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019GL082376 by Cochrane Colombia, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Geophysical Research Letters 10.1029/2019GL082376

had planarities >0.95, and the fifth had a planarity of 0.85—which were rotated slightly counterclockwise
relative to those in the north, such that they had nearly EW strikes (Figure 2 and Table S2). The dips were
generally less steep than the planes in Cluster n, with two of the planes dipping to the south (s‐II and s‐V)
and three dipping to the north (s‐I, s‐III, and s‐IV).

3.2. Space–Time Migration of Relocated Earthquakes

We examined the space–time migration of hypocenters to determine if there were patterns consistent with
the diffusion of fluids. We calculated separate migration rates for earthquakes occurring in 2017 (Clusters
n‐I, n‐II, n‐III, and s‐I) and 2018 (Clusters s‐II, s‐III, s‐IV, and s‐V). The migration rates for both sequences
are smaller than 0.1 km/hr (Figure S5), which has been suggested as a critical value that, if exceeded, would
be too fast for a fluid‐driven sequence (Chen & Shearer, 2011; Hauksson et al., 2016). The migration rate for
earthquakes in 2018 was slightly faster than for earthquakes in 2017. By fitting a homogeneous 3‐D fluid dif-
fusion model (Shapiro et al., 1997) to the space–time distribution of hypocenters, we obtained an overall
hydraulic diffusivity (D) of 0.02–0.40 m2/s for the 2017 earthquakes and a slightly higher value, D ≥
0.40 m2/s for the 2018 earthquakes (Figure S5).
We also examined migration rates for some of the individual clusters that showed especially high organiza-
tion: Clusters n‐II and n‐III, which showed northwest to southeast linear migration, and Cluster s‐I, which
occurred in July 2017 and showed west to east migration (Figure S6). For Clusters n‐II and n‐III, we only
chose earthquakes occurring in August 2017, which is a majority of events in each cluster, to fit the 3‐D dif-
fusion models. We obtained well‐fitting diffusion models with D ~ 1 m2/s for Clusters n‐II and s‐I, which is
close to the diffusion rates of previous Yellowstone earthquake swarms (Shelly et al., 2013). In contrast,
Cluster n‐III seems less diffusive‐like with D of only ~0.02 m2/s.

4. Imaging VP/VS in the Source Region

We used the approach of Lin and Shearer (2007) to estimate VP/VS ratios in the source region of 2017–2018
Maple Creek sequence. Instead of removing the mean differential time like Lin and Shearer (2007), we sub-
tracted the median P and S differential time in each event pair to mitigate the origin time bias. We selected
event pairs with S‐P differential times measured at a minimum of three stations, resulting in 177,697 pairs
from 2,401 events that yielded a VP/VS ratio of 1.602 ±0.011. An error of 0.001 is estimated from 200 boot-
strap resamplings, and another 0.01 uncertainty is the estimated bias from the takeoff angle inequality
between P wave and S wave raypaths. Our VP/VS estimate is within the range of 1.57–1.61 that was pre-
viously observed for this portion of the Yellowstone region from local earthquake tomography (Husen
et al., 2004). This consistency supports the idea that the assumptions of the Lin and Shearer (2007) method
related to similarity of near‐source raypaths are being met.

4.1. Spatial Variations in VP/VS

We next examined spatial variations in the VP/VS ratio. Considering just the northern portion of the
sequence that we previously defined as Cluster n in Figure 2, we have 139,740 pairs of P and S differential
times from 1,672 events, leading to a VP/VS ratio of 1.624 ±0.012.
We experimented with different methods of subdividing this sequence and ultimately settled on the three
smaller clusters defined previously as n‐I, n‐II, and n‐III (Figure 3). We obtained 236 events for Cluster n‐
I yielding 18,458 viable differential time pairs, which gave a VP/VS ratio of 1.494 ±0.014. The 387 events
in Cluster n‐II yielded 16,580 viable differential time pairs, which give an extremely low VP/VS ratio of
1.393 ±0.017. The 930 events in Cluster n‐III yielded 95,019 viable differential time pairs, resulting in a rela-
tively high VP/VS ratio of 1.658 ± 0.003 (Figure 3).
For the southern Cluster s, from which we previously defined clusters s‐I, s‐II, s‐III, s‐IV, and s‐V (Figure 2),
we have 37,394 viable differential time pairs from 724 events, leading to a VP/VS ratio of 1.505 ± 0.002
(Figure 3). We found VP/VS ratios that varied from 1.48 to 1.55 for the five smaller clusters; however, the
values were poorly resolved with errors of ~0.02 from 200 bootstrap resamplings. In addition, there was
0.01 error estimated from takeoff angle inequality between P and S raypaths. Thus, we consider there to
be little evidence for significant spatial variation in VP/VS ratios within Cluster s.

PANG ET AL. 4657

 19448007, 2019, 9, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019GL082376 by Cochrane Colombia, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Geophysical Research Letters 10.1029/2019GL082376

Figure 3. (a) Estimates of “in situ” VP/VS ratio for clusters of earthquakes in the Maple Creek sequence. (b) Cross section
along A‐A′, with the same color scale as in (a). (c) Cross section along B‐B′ with the same color scale as in (a). Depth is
relative to sea level in both cross sections.

4.2. Interpretation of Observed VP/VS Ratios

We observe VP/VS ratios of 1.40–1.66 throughout the source region of the 2017–2018 Maple Creek sequence,
much lower than the average value of 1.77 for rocks in the continental crust (Christensen, 1996). Partial melt
would lead to above average VP/VS ratios and our low VP/VS ratios instead imply the presence of aqueous
fluids (Watanabe, 1993). Similarly low VP/VS observations and interpretations have been made in other vol-
canic regions, including Hawaii (Lin et al., 2015) and northeastern Japan (Nakajima et al., 2001). The pre-
sence of fluids in the source region of the 2017–2018 Maple Creak sequence is also suggested by
earthquake migration patterns that can be fit by a 3‐D homogenous diffusion model with hydraulic diffusiv-
ities of 0.02–1 m2/s (Figures S5 and S6).

PANG ET AL. 4658

 19448007, 2019, 9, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019GL082376 by Cochrane Colombia, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Geophysical Research Letters 10.1029/2019GL082376

It is difficult to determine from our seismic observations whether the fluids in the source region are liquid
water, H2O/CO2 gases, or a mixture of the two. Gregory (1976) obtained VP/VS ratios of 1.42–1.98 for
water‐saturated rocks and 1.30–1.69 for gas‐saturated rocks. More recently, Lin et al. (2015) calculated VP/
VS ratios for a host rock with a VP of 6.0 km/s, a VP/VS ratio of 1.74, and ellipsoidal cracks with aspect ratios
from 0.001 to 0.30. From their simulations, the lowest VP/VS ratio obtained from a water‐filled crack model
was 1.594, but the CO2‐filled crack models could reduce the ratio to as low as the value of 1.412 observed in
the Kilauea (Hawaii) caldera. Thus, while not unique, our preferred interpretation of the low VP/VS values
near the Maple Creek sequence is the presence of CO2‐filled cracks.
Lin et al. (2015) also found that for CO2‐filled cracks with constant aspect ratio, VP/VS decreases with
increasing porosity. This provides a possible explanation for our observation of a reduction in VP/VS from
1.66 to 1.39 between the southeastern and northwestern portions of the Cluster n (Figure 3). The VP/VS
decrease we observe would require a porosity increase of ~80%, which might be expected if there were
increased damage in the northwestern segment of Cluster n (i.e., Cluster n‐II) because of its closer proximity
to the Hebgen Lake rupture zone. Johnson et al. (2018) noted significant scarp height in the southeastern
portion of the rupture zone, and because of the relatively low strain rates, it is plausible that the damage
in the rupture zone has not yet had time to heal. The increased epicentral scatter in Cluster n‐II is also con-
sistent with a highly fractured region.

5. Discussion
5.1. The Northern Portion of the 2017–2018 Maple Creek Sequence
The northern portion of the 2017–2018 Maple Creek earthquake sequence, which we designate as Cluster n,
exhibits a V‐shaped structure that mainly strikes northwest‐southeast (Figure 2). On a smaller scale, the
numerous “X”‐shaped structures associated with the PCA results for three subclusters are consistent with
composite focal mechanisms from Shelly and Hardebeck (2019), which suggest the existence of a mesh‐like
fracture network in the region. Magmatic fluids likely played a major role in earlier Yellowstone swarms
(Farrell et al., 2010; Shelly et al., 2013; Waite & Smith, 2002), and Shelly and Hardebeck (2019) suggested that
this portion of the sequence is also related to fluids. However, fluid diffusion alone does not explain the
change in migration rates between Cluster n‐II and Cluster n‐III. The low overall diffusivity (0.4 m2/s) for
the northern portion of the sequence, the less‐diffusive‐like properties of Cluster n‐III, and the
double‐couple nature of the moment tensor of the largest event imply that there may be an alternative force
driving the seismicity. Several observations suggest that Cluster n is primarily a continuation of the long‐
lived aftershock sequence of the 1959 Mw 7.2 Hebgen Lake earthquake: (1) The earthquakes are adjacent
to (within 1–2 rupture lengths) the southeast end of the surface rupture of the Hebgen Lake earthquake,
(2) the earthquakes are in a region of positive Coulomb stress perturbation from the Hebgen Lake event
(Figure S3, Chang & Smith, 2002), (3) the overall NW‐SE structure of the cluster is consistent with the orien-
tation of fault scarps of the Hebgen Lake event, and (4) the T axis of the Mw 4.4 event on 16 June 2017 is con-
sistent with T axes of Hebgen Lake event and one of its aftershocks (Doser, 1985).
We also studied the temporal evolution of Cluster n using an approach similar to Vidale and Shearer (2006).
We divided Cluster n into three subsequences based on their origin time and location (Figure 4a and Figure
S7). We calculated their daily rates and duration time independently and then normalized the time scale by
the duration (in days), making the maximum duration time equal to 1. The normalized time evolution pat-
terns show that more earthquakes occurred in the early stage for all three sequences (Figure 4b). After sum-
ming the three time‐normalized sequences, we obtained a robust time evolution measurement showing that
Cluster n is an aftershock‐like sequence in which most earthquakes occur early in the sequence and there is
clear decay with time, supporting the idea that Cluster n represents late aftershocks of the Hebgen
Lake earthquake.
Overall, we interpret Cluster n as part of the long‐lived aftershock sequence of the 1959 Mw 7.2 Hebgen Lake
earthquake, similar to our interpretation of the 2014–2017 seismicity near Challis, ID, as primarily after-
shocks of the 1983 Mw 6.9 Borah Peak earthquake (Pang et al., 2018), with some influence of fluid diffusion.
Aftershock sequences in low‐strain‐rate regions can last hundreds of years (Castro et al., 2010; Ebel et al.,
2000; Pang et al., 2018; Stein & Liu, 2009; Toda & Stein, 2018), and there is evidence that aftershock

PANG ET AL. 4659

 19448007, 2019, 9, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019GL082376 by Cochrane Colombia, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Geophysical Research Letters 10.1029/2019GL082376

Figure 4. (a) Plot of magnitude versus time in color‐matched subsets of earthquakes. The warm colors mark earthquakes
in Cluster n (Figures 4 and S6) and the cool colors mark the earthquakes in Cluster s (Figures 4 and S6). (b) Number of
daily events as a function of normalized time so that each subset starts at time 0 and ends at time 1. The light red, red, and
purple histograms are color matched with (a), and the gray histogram is the result of summing the three earthquake
subsequences together. (c) Similar to (b) but for the earthquake subsequences in colors of dark blue, blue, and cyan. Again,
the gray histogram is the average of three individually normalized subsequences.

sequences in extensional tectonic settings are especially long (Valerio et al., 2017). The aftershock duration
time in years, ta, can be estimated by applying ta = 314/v from Stein and Liu (2009), where v is the loading
rate in millimeters per year. Considering the regional extensional strain rate determined from GPS measure-
ments (Payne et al., 2012), we estimate a loading rate of ~3 mm/year, corresponding to an aftershock dura-
tion of ~100 years for the 1959 Mw 7.2 Hebgen Lake earthquake.

5.2. The Southern Portion of the 2017–2018 Maple Creek Sequence

In contrast to the northern portion, the southern portion of the 2017–2018 Maple Creek sequence (Cluster s)
consists mainly of EW‐striking planar structures. Many of these structures dip to the north, opposite to the
dip of the Maple Creek section of the Red Canyon fault. These earthquakes are also shallower than those in
the northern sequence, making them closer to the hypothesized CO2‐filled body at shallow depths (Husen
et al., 2004). In addition, the southern epicenters are adjacent to the boundary of the 2.1‐Ma caldera
(Figure 2), and the south‐to‐north migration pattern for Clusters s‐II, s‐III, s‐IV, and s‐V is similar to the
migration pattern of the three previous major Yellowstone swarms, all of which have been associated with
magmatic fluids (Massin et al., 2013).
We applied the same temporal evolution analysis to these southern events as we did for the northern events.
We divided the earthquakes into three subsequences based on their occurrence time (Figure 4a) and equal-
ized the time scales by dividing by the duration times. The normalized time evolution pattern is significantly

PANG ET AL. 4660

 19448007, 2019, 9, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019GL082376 by Cochrane Colombia, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Geophysical Research Letters 10.1029/2019GL082376

different from the northern events. Instead of occurring early in the sequences, the earthquake rates reach a
maximum in the middle or later portion of the normalized duration (Figure 4c). The average daily seismicity
rate in the sequences gradually increases to a maximum and then slowly decays to the background level indi-
cating that the sequences are more swarm like than aftershock like. Cluster s‐I initiated during Cluster n and
the very small sequence (blue in Figures 4 and S7) in late October 2017 looks like a precursor to Cluster s‐II.
Combined with the south‐to‐north migration feature and the stable in situ VP/VS ratio, the clusters in the
south are likely a swarm induced by fluids or gases, perhaps with some stress interaction from the
aftershock‐like events in Cluster n.

6. Conclusions
The 2017–2018 Maple Creek earthquake sequence was the fourth major burst of seismicity in the
Yellowstone region since it was first instrumented with seismometers in 1972. It can be divided into two dis-
tinct subsequences: an aftershock‐like sequence for the northern portion and a swarm‐like sequence for the
southern portion. Overall, we interpret the 2017–2018 Maple Creek earthquake sequence as a tectonic after-
shock sequence of the 1959 Mw 7.2 Hebgen Lake earthquake that was influenced by magmatic fluids asso-
ciated with the nearby Yellowstone volcanic system.
We used waveform‐derived differential travel times to generate high‐precision relative relocations and esti-
mate VP/VS ratios in the near‐source region. We observed several planar structures and low VP/VS ratios
with significant spatial variations. The low VP/VS ratios preclude significant amounts of partial melt in
the source region and instead indicate the presence of cracks filled with either CO2 gas or aqueous solutions
(e.g., CO2/H2O mixture). In the northern portion of the sequence, the planes tended to dip steeply and strike
either NW‐SE or NE–SW, similar to the two nodal planes of the largest event in the sequence (Mw 4.4 on 16
June 2017). This geometry is consistent with the regional stress field and the faulting pattern in the nearby
1959 Mw 7.2 Hebgen Lake rupture zone. The VP/VS reduction in this region from 1.66 in the southeast to 1.40
in the northwest may be a manifestation of increased fault porosity/density as one approaches the still‐
damaged 1959 Mw 7.2 Hebgen Lake rupture zone. In the southern portion of the sequence the planar struc-
tures had shallower dips and nearly EW strikes, consistent with Quaternary faults located southeast of the
Hebgen Lake rupture zone, closer to the 0.63‐Ma caldera boundary. The VP/VS ratio of 1.505 ± 0.002 for this
region is lower than the north portion of the sequence, but we did not observe significant variations within
the southern portion.
We suggest that the southern portion of the 2017–2018 Maple Creek earthquake sequence is a swarm
Acknowledgments
The authors thank Hongjian Fang for induced by migrating fluids associated with the Yellowstone volcanic system, similar in nature to the prolific
help and suggestions related to VP/VS 1985 Yellowstone swarm. This interpretation is supported by the south‐to‐north space–time migration of
determination and Diane Doser for
events, the EW orientation of planar structures in the seismicity, the proximity of the 2.1‐Ma caldera bound-
discussion about the source parameters
of 1959 Mw 7.2 Hebgen Lake ary, and the relative proximity of the upper crustal magma reservoir beneath the 0.63‐Ma caldera boundary.
earthquake. This work was partially We interpret the northern portion of the 2017–2018 Maple Creek earthquake sequence as primarily late
funded by the United States Geological
aftershocks of the 1959 Mw 7.2 Hebgen Lake earthquake but with some influence of fluid diffusion
Survey under cooperative agreement
G16 AC00029 and partially supported (Figures S5 and S6) as noted by Shelly and Hardebeck (2019). These earthquakes are deeper and closer to
by the Brinson Foundation and the the Hebgen Lake rupture zone and form primarily NW‐SE planar structures, similar to the Hebgen Lake
Carrico Funds. Many of the figures
fault system. Their magnitude‐time distribution is more front loaded and Omori like compared to the clas-
were made with the Generic Mapping
Tools (GMT; Wessel & Smith, 1998). All sical swarm‐like distribution observed for the southern sequence. The northern sequence of earthquakes
of the data used in this study are openly also occurs in a region of increased positive Coulomb failure stress created by the 1959 Mw 7.2 Hebgen
available from IRIS (www.iris.edu). We
used data from the following seismic
Lake earthquake (Figure S3).
networks: MB (Montana Bureau of
Mines and Geology/Montana Tech
(Montana Bureau of Mines and References
Geology/Montana Tech (MBMG, MT Albuquerque Seismological Laboratory (ASL)/USGS (1990). United States National Seismic Network. International Federation of Digital
USA), 2001), PB (Plate Boundary Seismograph Networks. Other/Seismic Network. https://doi.org/10.7914/SN/US
Observatory), TA (IRIS Transportable Castro, R. R., Shearer, P. M., Astiz, L., Suter, M., Jacques‐Ayala, C., & Vernon, F. (2010). The long‐lasting aftershock series of the 3 May 1887
Array, 2003), US (Albuquerque Mw 7.5 Sonora earthquake in the Mexican Basin and Range Province. Bulletin of the Seismological Society of America, 100(3), 1153–1164.
Seismological Laboratory [ASL]/USGS, https://doi.org/10.1785/0120090180
1990), and WY (University of Utah, Chang, W. L., & Smith, R. B. (2002). Integrated seismic‐hazard analysis of the Wasatch Front, Utah. Bulletin of the Seismological Society of
1984). We thank Diane Doser, Gavin America, 92(5), 1904–1922. https://doi.org/10.1785/0120010181
Hayes, and an anonymous referee for Chen, X., & Shearer, P. M. (2011). Comprehensive analysis of earthquake source spectra and swarms in the Salton Trough, California.
detailed reviews. Journal of Geophysical Research, 116, B09309. https://doi.org/10.1029/2011JB008263

PANG ET AL. 4661

 19448007, 2019, 9, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019GL082376 by Cochrane Colombia, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Geophysical Research Letters 10.1029/2019GL082376

Christensen, N. I. (1996). Poisson's ratio and crustal seismology. Journal of Geophysical Research, 101(B2), 3139–3156. https://doi.org/
10.1029/95JB03446
Christiansen, R. L., Lowenstern, J. B., Smith, R. B., Heasler, H., Morgan, L. A., Nathenson, M., et al. (2007). Preliminary assessment of
volcanic and hydrothermal hazards in Yellowstone National Park and vicinity. U. S. Geological Survey.
Doser, D. I. (1985). Source parameters and faulting processes of the 1959 Hebgen Lake, Montana, earthquake sequence. Journal of
Geophysical Research, 90(B6), 4537–4555. https://doi.org/10.1029/JB090iB06p04537
Du, W. X., Thurber, C. H., & Eberhart‐Phillips, D. (2004). Earthquake relocation using cross‐correlation time delay estimates verified with
the bispectrum method. Bulletin of the Seismological Society of America, 94(3), 856–866. https://doi.org/10.1785/0120030084
Ebel, J. E., Bonjer, K. P., & Oncescu, M. C. (2000). Paleoseismicity: Seismicity evidence for past large earthquakes. Seismological Research
Letters, 71(2), 283–294. https://doi.org/10.1785/gssrl.71.2.283
Farrell, J., Husen, S., & Smith, R. B. (2009). Earthquake swarm and b‐value characterization of the Yellowstone volcano‐tectonic system.
Journal of Volcanology and Geothermal Research, 188(1–3), 260–276. https://doi.org/10.1016/j.jvolgeores.2009.08.008
Farrell, J., Smith, R. B., Taira, T. A., Chang, W. L., & Puskas, C. M. (2010). Dynamics and rapid migration of the energetic 2008–2009
Yellowstone Lake earthquake swarm. Geophysical Research Letters, 37, L19305. https://doi.org/10.1029/2010GL044605
Gregory, A. R. (1976). Fluid saturation effects on dynamic elastic properties of sedimentary rocks. Geophysics, 41(5), 895–921. https://doi.
org/10.1190/1.1440671
Hauksson, E., Andrews, J., Plesch, A., Shaw, J. H., & Shelly, D. R. (2016). The 2015 Fillmore earthquake swarm and possible crustal
deformation mechanisms near the bottom of the Eastern Ventura basin, California. Seismological Research Letters, 87(4), 807–815.
https://doi.org/10.1785/0220160020
Husen, S., Smith, R. B., & Waite, G. P. (2004). Evidence for gas and magmatic sources beneath the Yellowstone volcanic field from seismic
tomographic imaging. Journal of Volcanology and Geothermal Research, 131(3–4), 397–410. https://doi.org/10.1016/S0377‐
0273(03)00416‐5
IRIS Transportable Array (2003). USArray Transportable Array. International Federation of Digital Seismograph Networks. Other/Seismic
Network. https://doi.org/10.7914/SN/TA
Johnson, K. L., Nissen, E., & Lajoie, L. (2018). Surface rupture morphology and vertical slip distribution of the 1959 Mw 7.2 Hebgen Lake
(Montana) earthquake from airborne lidar topography. Journal of Geophysical Research: Solid Earth, 123, 8229–8248. https://doi.org/
10.1029/2017JB015039
Klein, F. W. (2002). User's guide to HYPOINVERSE‐2000, a Fortran program to solve for earthquake locations and magnitudes (No. 2002–
171). US Geological Survey. https://doi.org/10.3133/ofr02171
Lin, G., Amelung, F., Shearer, P. M., & Okubo, P. G. (2015). Location and size of the shallow magma reservoir beneath Kīlauea caldera,
constraints from near‐source Vp/Vs ratios. Geophysical Research Letters, 42, 8349–8357. https://doi.org/10.1002/2015GL065802
Lin, G., & Shearer, P. (2007). Estimating local VP/VS ratios within similar earthquake clusters. Bulletin of the Seismological Society of
America, 97(2), 379–388. https://doi.org/10.1785/0120060115
Massin, F., Farrell, J., & Smith, R. B. (2013). Repeating earthquakes in the Yellowstone volcanic field: Implications for rupture dynamics,
ground deformation, and migration in earthquake swarms. Journal of Volcanology and Geothermal Research, 257, 159–173. https://doi.
org/10.1016/j.jvolgeores.2013.03.022
Michelini, A., & Bolt, B. A. (1986). Application of the principal parameters method to the 1983 Coalinga, California, aftershock sequence.
Bulletin of the Seismological Society of America, 76(2), 409–420.
Minson, S. E., & Dreger, D. S. (2008). Stable inversions for complete moment tensors. Geophysical Journal International, 174(2), 585–592.
https://doi.org/10.1111/j.1365‐246X.2008.03797.x
Mogi, K. (1963). Some discussions on aftershocks, foreshocks and earthquake swarms—the fracture of a semi‐infinite body caused by an
inner stress origin and its relation the earthquake phenomena. Bulletin of the Earthquake Research Institute, 41, 615–658.
Montana Bureau of Mines and Geology/Montana Tech (MBMG, MT USA) (2001). Montana Regional Seismic Network. International
Federation of Digital Seismograph Networks. Other/Seismic Network. https://doi.org/10.7914/SN/MB
Nakajima, J., Matsuzawa, T., Hasegawa, A., & Zhao, D. (2001). Three‐dimensional structure of VP, VS, and VP/VS beneath northeastern
Japan: Implications for arc magmatism and fluids. Journal of Geophysical Research, 106(B10), 21,843–21,857. https://doi.org/10.1029/
2000JB000008
Pang, G., Koper, K. D., Stickney, M. C., Pechmann, J. C., Burlacu, R., Pankow, K. L., et al. (2018). Seismicity in the Challis, Idaho, Region,
January 2014–May 2017: Late aftershocks of the 1983 Ms 7.3 Borah peak earthquake. Seismological Research Letters, 89(4), 1366–1378.
https://doi.org/10.1785/0220180058
Payne, S. J., McCaffrey, R., King, R. W., & Kattenhorn, S. A. (2012). A new interpretation of deformation rates in the Snake River Plain and
adjacent basin and range regions based on GPS measurements. Geophysical Journal International, 189(1), 101–122. https://doi.org/
10.1111/j.1365‐246X.2012.05370.x
Petersen, M. D., Frankel, A. D., Harmsen, S. C., Mueller, C. S., Haller, K. M., Wheeler, R. L., et al. (2014). Documentation for the 2014
update of the United States national seismic hazard maps (No. 2014–1091). Geological Survey (US). https://doi.org/10.3133/
ofr20141091
Puskas, C. M., Smith, R. B., Meertens, C. M., & Chang, W. L. (2007). Crustal deformation of the Yellowstone–Snake River Plain volcano‐
tectonic system: Campaign and continuous GPS observations, 1987–2004. Journal of Geophysical Research, 112, B03401. https://doi.org/
10.1029/2006JB004325
Schmeelk, D., Bendick, R., Stickney, M., & Bomberger, C. (2017). Kinematic evidence for the effect of changing plate boundary conditions
on the tectonics of the northern US Rockies. Tectonics, 36, 1090–1102. https://doi.org/10.1002/2016TC004427
Shapiro, S. A., Huenges, E., & Borm, G. (1997). Estimating the crust permeability from fluid‐injection‐induced seismic emission at the KTB
site. Geophysical Journal International, 131(2), F15–F18. https://doi.org/10.1111/j.1365‐246X.1997.tb01215.x
Shelly, D. R., & Hardebeck, J. L. (2019). Illuminating faulting complexity of the 2017 Yellowstone Maple Creek earthquake swarm.
Geophysical Research Letters, 46, 2544–2552. https://doi.org/10.1029/2018GL081607
Shelly, D. R., Hill, D. P., Massin, F., Farrell, J., Smith, R. B., & Taira, T. A. (2013). A fluid‐driven earthquake swarm on the margin of the
Yellowstone caldera. Journal of Geophysical Research: Solid Earth, 118, 4872–4886. https://doi.org/10.1002/jgrb.50362
Smith, R. B., Jordan, M., Steinberger, B., Puskas, C. M., Farrell, J., Waite, G. P., et al. (2009). Geodynamics of the Yellowstone hotspot and
mantle plume: Seismic and GPS imaging, kinematics, and mantle flow. Journal of Volcanology and Geothermal Research, 188(1–3),
26–56. https://doi.org/10.1016/j.jvolgeores.2009.08.020
Stein, S., & Liu, M. (2009). Long aftershock sequences within continents and implications for earthquake hazard assessment. Nature,
462(7269), 87–89. https://doi.org/10.1038/nature08502

PANG ET AL. 4662

 19448007, 2019, 9, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019GL082376 by Cochrane Colombia, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Geophysical Research Letters 10.1029/2019GL082376

Toda, S., & Stein, R. S. (2018). Why aftershock duration matters for probabilistic seismic hazard assessment. Bulletin of the Seismological
Society of America, 108(3A), 1414–1426. https://doi.org/10.1785/0120170270
Trugman, D. T., & Shearer, P. M. (2017). GrowClust: A hierarchical clustering algorithm for relative earthquake relocation, with appli-
cation to the Spanish Springs and Sheldon, Nevada, earthquake sequences. Seismological Research Letters, 88(2A), 379–391. https://doi.
org/10.1785/0220160188
University of Utah (1984). Yellowstone Wyoming Seismic Network. International Federation of Digital Seismograph Networks.
Other/Seismic Network. https://doi.org/10.7914/SN/WY
Valerio, E., Tizzani, P., Carminati, E., & Doglioni, C. (2017). Longer aftershocks duration in extensional tectonic settings. Scientific Reports,
7(1), 16403. https://doi.org/10.1038/s41598‐017‐14550‐2
Vidale, J. E., & Shearer, P. M. (2006). A survey of 71 earthquake bursts across southern California: Exploring the role of pore fluid pressure
fluctuations and aseismic slip as drivers. Journal of Geophysical Research, 111, B05312. https://doi.org/10.1029/2005JB004034
Waite, G. P., & Smith, R. B. (2002). Seismic evidence for fluid migration accompanying subsidence of the Yellowstone caldera. Journal of
Geophysical Research, 107(B9), 2177. https://doi.org/10.1029/2001JB000586
Watanabe, T. (1993). Effects of water and melt on seismic velocities and their application to characterization of seismic reflectors.
Geophysical Research Letters, 20(24), 2933–2936. https://doi.org/10.1029/93GL03170
Wessel, P., & Smith, W. H. (1998). New, improved version of Generic Mapping Tools released. Eos, Transactions American Geophysical
Union, 79(47), 579–579. https://doi.org/10.1029/98EO00426
Whidden, K. M., & Pankow, K. L. (2012). A catalog of regional moment tensors in Utah from 1998 to 2011. Seismological Research Letters,
83(5), 775–783. https://doi.org/10.1785/0220120046
White, B. J. P., Smith, R. B., Husen, S., Farrell, J. M., & Wong, I. (2009). Seismicity and earthquake hazard analysis of the Teton–
Yellowstone region, Wyoming. Journal of Volcanology and Geothermal Research, 188(1–3), 277–296. https://doi.org/10.1016/j.
jvolgeores.2009.08.015
Yellowstone Volcano Observatory (2006). Volcano and earthquake monitoring plan for the Yellowstone Volcano Observatory, 2006–2015.
U.S. Geological scientific investigation report 2006–5276 (13 p.). Reston, VA.

References From the Supporting Information

Pechmann, J. C., Bernier, J. C., Nava, S. J., Terra, F. M. (2006). Correction of systematic time‐dependent coda magnitude errors in the
Utah and Yellowstone National Park region earthquake catalogs, 1981–2001. Appendix C in Arabasz, W. J., R. B. Smith, J. C.
Pechmann, K. L. Pankow, & R. Burlacu, Integrated regional and urban seismic monitoring—Wasatch Front Area, Utah and adjacent
intermountain seismic belt, Final Tech. Rept., U.S. Geological Survey Cooperative Agreement 04HQAG0014 (137 pp.). Retrieved from
http://earthquake.usgs.gov/research/external/reports/04HQAG0014.pdf
Lin, G., & Shearer, P. M. (2009). Evidence for water‐filled cracks in earthquake source regions. Geophysical Research Letters, 36, L17315.
https://doi.org/10.1029/2009GL039098

PANG ET AL. 4663