Author: Eric Bergman

The Cumana cluster is named for the city of Cumana, capital of the
Venezuelen state of Sucre. This cluster, along with the Valencia cluster
replaces the Caracas cluster in GCCEL. The cluster includes a 6.1 MS
earthquake on June 12, 1974 and a 6.9 Mw earthquake on July 9, 1997.
With most events distributed along the coast and few seismograph
stations to the north, the azimuthal coverage for location calibration
is a bit weak, but the relocations were stable. All events have depth
control from near-source or local distance readings, or teleseismic
depth phases.

The Valencia cluster is named for the city of Valencia, capital of
Carabobo State in Venezuela. A cluster named Valencia in the same region
was published earlier in GCCEL but this version is a fresh start, based
on the large, damaging earthquake(s) of June 24, 2026. There has been
controversy as to whether it should be treated as two separate events or
one very complex event. It is treated as two events here. The initial
event of magnitude 7.2 Mw was followed by a second large pulse of energy
(7.5 Mw) after about 30 seconds. Because of interference from the coda
of the initial event, assembling a consistent arrival time dataset for
the second event was challenging. The arrival times reported by the ISC
were supplemented by a carefully curated set of arrival times by NEIC
analysts that helped resolve the second mainshock hypocenter, which
locates ~150 km east of the first one, just offshore of Caracas. The
cluster also includes a 6.6 MS earthquake on July 30, 1967 and a 6.4 Mw
earthquake on September 12, 2009. The June 24, 2026 earthquakes produced
surprisingly few aftershocks. 17 aftershocks over the following ~month
are included in the cluster but most of them do not have depth control
and are held at a default depth of 12 km. All other events in the
cluster have depth control from near-source and local-distance readings
and teleseismic depth phases in the case of the 1967 earthquake. The
distribution of local seismograph stations is quite good and the cluster
has a robust location calibration.

The Ambon cluster is named for the island of Ambon in the Maluku Islands
of Indonesia. The cluster consists of earthquakes with depths to a
little over 50 km. It contains three M6 earthquakes, a 6.0 Mw event on
December 30, 1996, a 6.5 event on March 19, 2001, and a 6.5 Mw event
(the so-called Ambon earthquake) on September 25, 2019. All earthquakes
are recorded to teleseismic distances. All events have depth control from
near-source and local-distance arrivals or teleseismic depth phases. The
number of local distance seismograph stations is small but they are well
distributed and provide a robust location calibration.

The Kochi cluster is named for the city of Kochi on the island of
Shikoku, Japan. The cluster includes most of Shikoku. It includes
earthquakes to a depth of ~50 km. The only M6 event is a 6.4 mb
earthquake on March 24, 2001. About half the events are recorded only to
near-regional distances (~5.0°) but they were retained because the
relocation suffered from instability without them. All events have depth
control from near-source and local distance readings, and many have
consistent estimates of focal depth from teleseismic depth phases. The
local seismic network is very dense and direct calibration is very
robust.

The Caucasia cluster is named for the town of Caucasia in northern
Colombia. It covers a relatively large area, which was necessary to
obtain adequate station coverage. Many of the events are subcrustal in
depth, as deep as 190 km. There are especially large earthquakes
(maximum magnitude is 5.7) but all events are observed at teleseismic
distances. All events have depth control, from teleseismic depth phases
in many cases, but also from near-source and local-distance
observations. The location calibration is robust overall but some of the
deeper events are susceptible to instability.

The Hawaii cluster is named for the island of Hawaii. It covers the
entire island and replaces the Keaoi cluster which included only 22
events along the southeast coast of the island. The cluster includes the
7.7 Mw Kalapana earthquake on November 29, 1975, and earthquakes with
magnitudes 6.0-6.7 on November 16, 1983 (6.6 Mw), October 15, 2006 (6.7
Mw) and May 23, 2026 (6.0 Mw). Many of the events included in this
cluster are observed only by seismic stations on the island, but are
retained for use in more detailed studies of the complex structure of
the island. All events are observed to a distance of at least 1.0°. All
events have depth constraint from near-source and local-distance
readings. The pattern of focal depths is bi-modal, with peaks at
approximately 8 and 34 km. Earthquakes that required focal depths above
sea-level have been left out of the cluster.

The Ziarat cluster is named for the city of Ziarat in northeast
Baluchistan Province, Pakistan. The cluster includes 7.0 Ms and 7.4 Ms
earthquakes in August 1931 and the 7.1 Mw event on February 27, 1997.
The cluster is motivated by a pair of 6.4 Mw earthquakes that occurred
on October 28 and 29, 2008 and that provide a basis for location
calibration. The source region is very poorly instrumented but the 2008
sequence has been studied by several groups using InSAR to try to
unravel what is evidently a very complex faulting pattern, featuring as
many as 5 major faults with conjugate geometry. The location calibration
is done with indirect calibration, fitting the relative locations of the
mainshocks and larger aftershocks to this complex pattern. The InSAR
signal of a more recent Mw 6.0 event (October 6, 2021) helps in
constraining the calibration. Origin time is constrained by Pg and Sg
readings from the seismograph station QUE at Quetta, the only station
within local distance of the cluster. All earthquakes have depth
control, from near-source and local distance readings from QUE and for
many events, teleseismic depth phases. All events are observed at
teleseismic distances.

The distribution package of mloc v13.0.6 (zipped archive ~450 MB) is now posted on the website, as well as an updated version of the Desktop User’s Manual (zipped archive, ~17 MB). This update contains mostly minor bug fixes and tweaks to some of the output files and plots.

The Karaman cluster is named for the city of Karaman in south-central
Turkey. The cluster covers a fairly large area but it is one of modest
seismicity. About a quarter of the earthquakes occur offshore, between
the southern coast of Turkey and Cyprus. The largest event is only 5.2
mb. The cluster is restricted to events observed at teleseismic
distances. Station distribution for location calibration is very good.
All events have depth control from near-source and local-distance
readings.

The Konitsa cluster is named for the town of Konitsa in northwestern
Greece, near the border with Albania. The cluster was motivated by a 5.5
Mw earthquake on March 8, 2026 and includes a 5.6 mb event on October
13, 1969 and several other events greater than magnitude 5.0. All events
are observed to teleseismic distances and all events have depth
constraint, mainly from near-source and local-distance observations. The
depths of several early events are constrained by teleseismic depth
phases. Two events with near-source observations are deep, around 60 km,
but the remainder are at crustal depths between about 5 and 25 km. The
distribution of stations is excellent and location calibration is very
robust.