Author: Eric Bergman

The Kamchatka cluster is named for the Kamchatka Peninsula in the far
east of Russia, and covers the subduction zone along southern
Kamchatka’s Pacific coast. The region is intensely active, most notably
with the 9.0 Mw earthquake of Novermber 4, 1952 and the 8.8 Mw
earthquake of July 29, 2025. The cluster includes both earthquakes
(indicated by larger red stars in the base map), as well as the
September 13 (7.4 Mw) and September 19 (7.8 Mw) aftershocks of the 2025
mainshock, indicated by smaller red stars in the base map, but the point
of this cluster is not to represent the full pattern of seismicity in
the region; it is to provide a framework of calibrated hypocenters which
can be a reference point for more detailed studies of the seismicity. As
in most subduction zone settings, most of the seismicity in southern
Kamchatka is off-shore and therefore is not amenable to direct
calibration because of poor azimuthal coverage. However, by careful
selection of earthquakes, including events that occurred at greater
depths (to ~190 km) in the subduction zone, a cluster with good
azimuthal coverage for direct calibration can be assembled. It is
possible to add additional events to this kernel, of course, but adding
too many offshore events destabilizes the location calibration and even
if that were not the case, the shear number of large earthquakes in this
region overwhelms the computational capacity of the relocation code.
Therefore the strategy for placing more of the seismicity of this
important seismic zone in a location-calibrated context is to first
establish this calibrated framework which includes well-recorded events
from 1952 through 2025. Then subsets of the seismicity in the southern
Kamchatka subduction zone, such as a fuller representation of the 1952
sequence, can be relocated with calibrated locations through the
indirect calibration process, making use of the reference locations
determined in this cluster. All earthquakes in the cluster have depth
control, often from both direct arrivals at short epicentral distance
and from teleseismic depth phases. After the 1952 and 2025 mainshocks,
events are selected mainly on the basis of being well-recorded
teleseismically and also having observations at local seismograph
stations that provide the azimuthal coverage needed for a stable direct
calibration. Another criterion is temporal continuity from 1952 through
2025, so that the cluster adequately represents the evolving network of
regional and global seismograph stations. Direct calibration is based on
a simplistic flat-layered crustal model which is obviously a poor
representation of the true velocity structure of a subduction zone. To
some extent the resulting increased scatter in residuals is reflected in
the uncertainties of hypocentral parameters but there is undoubtedly
some remaining systematic bias in the hypocenters.

The Anchor cluster is named for Anchor Point at the western end of the
Kenai Peninsula, Alaska. The cluster spans the Cook Inlet. It includes
events in two depth ranges, events shallower than ~25 km, mainly along
the coast of the Alaskan mainland on the west side of Cook Inlet and
events in the range 30-65 km on the eastern side of Cook Inlet, and
beneath the western coast of the Kenai Peninsula. Many events are small
and only observed to regional distances, but all events are observed to
at least 5° epicentral distance. All events have depth control from
near-source and local-distance arrival times.

The Simav-Gediz cluster is named for two significant earthquakes in
central Turkey, the 7.2 Mw Gediz earthquake on March 28, 1970, and the
5.9 Mw Simav earthquake on May 19, 2011. Initial work on these clusters
was done by Ezgi Karasözen. The cluster includes both sequences as well
as other well-recorded events in the vicinity. All events were recorded
to at least 10° epicentral distance and most were recorded
teleseismically. The seismograph network in this area was sparse before
about 2007, but quite dense afterwards, so all events after that time
have depth control from near-source and local-distance readings. Depth
control for earlier events is a mixture of local-distance readings,
teleseismic depth phases, and, in 17 cases lacking any other constraint
(all prior to 1977), a default depth sampled from a probability density
function that was based on the distribution of events with depth
constraint. The probability density function has a peak at 9 km and a
spread of 2.5 km. Sampled depths are required to be within 2-sigma of
the mean (i.e., 4-14 km).

The distribution of mloc v13.0.2 was posted today. It can be downloaded from https://seismo.com/mloc/distribution/. The accompanying Desktop User’s Manual can be downloaded from https://seismo.com/mloc/desktop-manual/.

Due to extensive changes to the source code since the last distribution, cohesion between the current version of mloc and the User’s Manual is not as complete as I would like. I will be glad to be informed of places in the Manual that need further work.

The Oran cluster is named for the city of Oran on the Mediterranean
coast of Algeria. It includes a few events across the border in coastal
Morocco. The cluster contains four events with magnitude > 5, the
largest being the 5.5 Mw event on June 6, 2008. Most events were
observed to teleseismic distances but some smaller events that were
recorded only to near-regional distances were retained to improve the
azimuthal coverage and statistical power of the location calibration.
All events have depth control, from near-source or local-distance
readings, or in a several cases, a waveform modeling study or
teleseismic depth phases.

The Afar cluster is named for the Afar region of Ethiopia, Eritrea and
Djibouti. The cluster occupies the northern part of what has been termed
the Danakil or Arrata microplate. Much of the seismicity has occurred in
swarms associated with volcanic activity. The cluster includes five
events with magnitude greater than 6.0, in swarms that occurred in April
1969 and August 1989. All but five events in the cluster were recorded
at teleseismic distances. All events have depth control from near-source
and local-distance readings or (more frequently than usual) teleseismic
depth phases. Location calibration is based mainly on a seismic swarm
that was well-instrumented with a temporary network from November 2007
to February 2009, and included events large enough to be recorded at
teleseismic distances, which established the necessary connectivity with
events that lack local observations.

The Sumba cluster is named for the island of Sumba in the Lesser Sunda
Islands of eastern Indonesia. The cluster is restricted to events with
constrained focal depths of 40 km or less since 2008. Earlier events
lack station coverage to adequately link them to the more recent
earthquakes. The largest earthquake has magnitude 5.8 Mw, on February
21, 2022. With one exception (depth constrained by teleseismic depth
phases) all events have depth constraint from near-source or
local-distance readings. Many small events that were recorded only to
near-regional distances are retained to improve the statistical power of
the location calibration.

The Lucia cluster is named for the island of St. Lucia in the Windward
Islands. It includes events near Barbados and St. Vincent and the
Grenadines. Station coverage is fairly good since 2008. Many events are
recorded only locally but there are useful amounts of data at far
regional and teleseismic distances. The largest event is 5.7 mb on
August 12, 1987. All events have depth control, mostly from near-source
and local-distance readings, but several events are constrained in depth
by teleseismic depth phases.

The Sparta cluster is named for the city of Sparta in northwestern North
Carolina. It covers a rather broad area and includes events in eastern
Tennessee, western Virginia and southernmost West Virginia. The cluster
is based on a 5.1 Mw earthquake on August 9, 2020 very close to Sparta.
The Sparta earthquake is remarkable for its shallow depth (with various
estimates in the 2-4 km range) and the observation of surface faulting.
It has a depth of 4 km in the calibrated relocation, based on arrivals
at local distance, but a shallower depth cannot be ruled out. The
cluster as a whole is at quite shallow depths, with only a single event
as deep as 10 km. It is quite likely that many of the events are mining
blasts. Most events have magnitudes in the 2-3 range. The distribution
of seismic stations is very good and the location calibration is very
strong, with an uncertainty of ~500 m for the hypocentroid at the 90%
confidence level. However the calibrated epicenter, like most other
location-based estimates that were reported, is a few km south of the
surface faulting. This could be a consequence of lateral heterogeneity
in the region. Seismograph stations to the southeast are on the coastal
plain, while the rest are in the Appalachian Mountains.

The Chardon cluster is named for the city of Chardon, Ohio. The cluster
covers the northeastern corner of Ohio along the southern bank of Lake
Erie, and includes some events in western Pennsylvania. The largest
event is a 5.3 mb event on January 31, 1986; no other event is larger
than 3.9 mb. Most events are recorded only to regional distances. All
events have depth control from near-source or local-distance readings.