Earthquake Catalogs
SCSN Catalog Change History
These pages describe any changes in Southern California Seismic Network real time processing, post processing, or in Southern California Earthquake Data Center catalog archival that might be of interest to users of the earthquake catalog. Catalog users may also want to review our SCEDC Archive Information page. Since August 18, 2012, daily versions of SCSN and SCEDC format catalog files are stored on GitHub. Thank you for your interest in southern California seismicity!
2023
December 8, 2023 SCSN Now Using Machine Learning Picks in Post Processing
After extensive testing, the SCSN changed the phase picker used in its post-processing software for events below magnitude 3 beginning in November, 2023. The previous picker was a short-term average/long-term average (STA/LTA) algorithm that is used in the real-time seismic analysis system. This picker has been replaced by PhaseNet, which is a deep-learning algorithm (Zhu and Beroza, 2018). Automatic solutions using the new post-processing setup began on November 6, contributing to the SCEDC catalog and solutions sent out on PDL starting November 10. Fine-tuning to optimize use of the new picker will likely continue in the near future.
September 10, 2023 Catalog Processing Status files now available
The SCEDC is now hosting files that show the curation status of a given day in the earthquake catalog. Historically, SCSN analysts have made an effort to review every event in our catalog. With the growth of the network and catalog, however, complete review has become more difficult with our limited staff. This is particularly true of the 2019 Ridgecrest sequence, which remains largely unreviewed below M2.5. These files can be found at at this location. They are updated daily.
2020
August 17, 2020 Change of policy for Assignment of gtype for events near network boundary
A "Local" event (gtype = "l") is defined as an earthquake within the SCSN authoritative region. Local events are currently defined strictly geographically, i.e. as any event for which the epicentre falls within the SCSN’s official reporting region. Users should note that a) for sequences near the SCSN reporting region boundaries, events may be labelled as a mix of ‘Local’ and ‘Regional’, and disabling the ‘gtype’ restriction and using a polygon is a more reliable method to obtain complete results; b) the labelling of an event within the SCSN catalogue as ‘Local’ or 'Regional' does not necessarily reflect its status as the authoritative (or otherwise) origin in ComCat. Prior to 2015, in cases where a sequence or large event occurred near the border, events could be labelled ‘Local’ up to 10-25km (0.1 to 0.2 degrees) beyond the official boundary.
2016
January 12, 2016 Additional Mlr Magnitude Type Calculated
Starting at the end of December 2015, SCSN began calculating an additional magnitude type, labelled Mlr, which is a revised local, or Richter, magnitude (Ml). Mlr magnitudes are only calculated for events with Ml between 3.0 and 6.0, and are obtained by applying a linear adjustment to the Ml value. The adjustment is designed to bring initial magnitude values derived from Ml into closer agreement with moment magnitude (Mw) values (which take somewhat longer to calculate), because Mw is expected to be the preferred magnitude type for events above magnitude 3.5.
For most areas in southern California, Ml is systematically larger than Mw for magnitudes greater than 3.5, consequently the Mlr adjustment is a reduction of the Ml value of up to 0.5 unit (larger adjustment for larger events). Mlr is calculated using the following formula:
Mlr = Ml * 0.853 + 0.40125
For the majority of earthquakes, Ml will be the preferred magnitude for events smaller than 3.5 and Mw preferred for events greater than 3.5. Mlr may be the first reported magnitude for events larger than 3.5 and could be the preferred catalog magnitude if an Mw cannot be obtained. In general, catalog searches using the usual STP and web tools will not return an event's Mlr value unless it is the preferred magnitude.
2015
September 22, 2015 Change in Depth Datum in Earthquake Catalogs
On September 22, 2015, the SCEDC earthquake catalogs will report event depths relative to the WGS84 reference ellipsoid. Older versions of the catalog with the model depth datum will be available on the SCEDC repository on GitHub. Relative to the reference ellipsoid, shallow earthquakes may now have negative depths as ground elevations in California range approximately from +0.085 km (Death Valley, below sea level) to -4.421 km (Mount Whitney). The catalog search minimum depth will therefore be set to a default value of -5 km.
September 22, 2015 Change in Earthquake Catalog formats
On Septemeber 22, 2015, the SCEDC earthquake catalog formats "SCEDC" and "SCSN/Catread" will have the following format changes:
- A new column: geographic type is added.
- Changes in values for Event Type. Events previously with values "le","re","ts" will now have an event type of "eq" and geographic types of "l", "r", and "t" respectively.
For users of the premade SCSN/SCEDC catalog files, all files are generated in the new formats, but a branch of the files before the change was made exists in the SCEDC repository on GitHub.
January 1, 2015 Changes in Post Processing Procedures
Due to reduced funding and personnel, the analysis practices (or post processing) of the earthquake catalog were revised in order to continue handling, in a timely manner, the full, continuous, seismic workload. While aiming to preserve as much data quality as possible, starting as of January 1st, 2015 every detected event will no longer go through the obligatory and thorough hand-timing by an analyst. While every event is still reviewed by a seismologist, a significant portion goes through only a visual inspection, before it is finalized in the catalog. All events M2.5 or greater or those failing the initial summary inspection, will undergo full, manual analysis. Under the new practices, there will be events processed only (and completely) by the automated system. To ensure quality, all events with automated solutions that have an RMS of 0.4 or greater, or with anomalous depths or magnitudes will be fully analyzed. There is no identification in the catalog for which events have been manually timed and which have been only automatically timed and inspected. We will be reviewing the catalog to assess if there is any impact of this change in procedure. We thank you for your understanding.
2013
June 13, 2013 Changes in Real Time Processing Parameters
The Southern California Seismic Network (SCSN) made changes to its real time event detection software which should improve its accuracy in detecting events as well as determining accurate locations and magnitude. The change was to add the no_S_on_Z parameter to binder.
April 23, 2013 Update to Data Processing Protocol
Due to a recent confluence of moderately high seismicity, more stations to process, and level staffing, we have been forced to make some changes in the earthquake catalog processing. These changes went into effect at the time of the March 2013 Anza sequence. Prior to April 2013, we made every reasonable effort to place every impulsive pick at the correct time, and some emergent picks as well. This procedure improved the locations, but also caused long delays in getting large sequences processed.
- Each earthquake must still be reviewed by an analyst, to make sure that the location and magnitude are within reasonable accuracy (which depends on the geographic location), and quarry blasts are appropriately marked.
- More reliance on an automatic picker that is tailored to the post-processing environment.
- Nearest 5 to 10 stations are accurately timed.
- Since more stations means denser coverage, most picks beyond 100 km distance not included (again, this depends on the geographic area).
- Earthquakes with 2.5 or larger magnitude get the same attention as before.
- Because the automatic picker may not choose the correct wave, there may be a degradation in the focal mechanisms for small quakes. New software probably offsets this problem, by also incorporating S amplitudes into the solution.
If an earthquake passes through this process, it has an “I” or “intermediate” processing state, which flags the event for further review at a later time. If activity is low, or there is some reason it is efficient or necessary to completely time an earthquake, it moves to an “F” or “finalized” processing state. This means analysis has been completed for this event.
As we have time, we may be able to go back and finalize most of the earthquakes in question.
We also continue to work on finalizing the unfinished El Mayor Cucupah earthquakes in April and May 2010.
April 17, 2013 Changes in real time processing parameters
The Southern California Seismic Network (SCSN) made changes to its real time event detection software which should improve its accuracy in detecting events as well as determining accurate locations and magnitude. A detailed list is as follows:- new binder_ew binary Version <1.0.11 2013.03.22>
- binder_ew, eqproc and eqassemble using PICK_RING (instead of FILTERPICK_RING)
- changed the grid weight from 0,0 to 2,1 for picks with quality 2 and 3
- no_P_on_Horiz
- Channel number Maps:
- ChannelNumberMapByNet NEZ PB
- ChannelNumberMapByNet ZNE CI
- ChannelNumberMapByNet ZNE CE
- ChannelNumberMap ZNE
- rstack 60.0
- new trimag binary - VERSION: TriMag2 v0.2.8, 2011-05-06
- parameters for trimag:
- MinCutoffDist 20.0
- MaxCutoffDist 200.0
- StepCutoffDist 20.0
- DefaultCutoffDist 150.0
- UseCITwindows 0
- PostSMult 1.0
- PostBuffer 1.0
