Last update Nov 2022
Continuous data for all CERI stations is available at IRIS as of late 2013. An unfortunate error prevented archiving at IRIS of all but the vertical channels for CERI broadband stations until 2013 so only the vertical is available at IRIS as of 2004. Event data are archived locally. See the IRIS metadata aggregator for NM and ET.
Hypocenters and arrivals for all CERI located earthquakes are available in the USGS Comprehensive Catalog (Comcat). Hypocenters and associated errors are also available in the CERI online catalog.
The analog backbone (the SCx shortperiod analog channels) was discontinued on May 8, 2025. These are the subset of channels that were digitized at CERI and transmitted by various analog means from the MKTA and LNXT nodes to CERI. The LNXT channels used the WKNO (and subsequently KLOV) SCA subcarrier (sca = subsidiary communications authority). Hence the channel designator SCx for these channels. They were primarily useful for automated locations and rapid alerts because they were not gain-ranged (gain-ranged channels do not play well with auto pickers). As we convert more gain-ranged analog stations to digital, we are adding them into the auto picker for rapid locations and the analog backbone is no longer needed. It also allows us to retire the clock and obsolete PC hosting the digitizer at CERI.
We migrated our AQMS processing system from oracle/solaris to postgres/linux on Nov 15, 2022. We made no changes to processing other than the new OS and DBMS.
As of Jan 1, 2015 we now use the ANSS Quake Management System (AQMS) instead of Xpick/Hypoellipse. The primary difference for the data, is that output files are now in Hypoinverse format (ftp://ehzftp.wr.usgs.gov/klein/hyp1.40/) rather than Hypoellipse. There is a .hyp file instead of a .arc file and there is no .out file. The .hyp is a similar format but slightly different from .arc. Sac format files for all the associated waveforms are also in the event directories and there are no longer any subdirectories (all files are in the top level event directory).
When we made AQMS our production system on 1/1/2015, we used mapping of hypoinverse weight to arrival error in seconds that made errors too small (i.e. too few arrivals with weight of 1) so we loosened the error seconds beginning 1/1/2017. All events located with AQMS back to 1/1/2015 have been reprocessed with the new parameters. The specific AQMS parameter in the analysis tool (jiggle) is hypoinvWgt2Deltatime:
# values used prior to 1/1/2017 (and beginning 1/1/2015)
#hypoinvWgt2DeltaTime=.02 .05 .10 .20 .40
# values beginning 1/1/2017
hypoinvWgt2DeltaTime=.05 .10 .25 .50 1.0
CERI waveform data from 1996 to present is online. The data are in sac format and the headers do not contain complete station information. Arrival times and hypocenter information are available in hypoellipse format files. In most cases there is single directory per event. Delays sometime exist in getting data from the staging area to the archive area so lack of a directory for a particular event does not necessarily imply its lack of existence; check back later. The disks containing the data are mounted on gaia and accessible through other systems on the NIS (sometimes known as gaia) domain. Data are in sun byte order and do not use xdr.
The primary data directory is /gaia/data/seisnet/. There are four directories in the primary directory:
Within each of the network directories, data are further subdivided by year. Each year contains further subdirectories based on event type. For events prior to 1/1/2000, NMSZ events are in the Reg subdirectory. For events after 1/1/2000, NMSZ events are in the Loc subdirectory. For example, data for the Pocahontas earthquake are in: /gaia/data/seisnet/NM/1999/Reg/199910210818.
There may be a delay of several days to several weeks getting some events to the archive directories on /gaia/data (usually the smaller it is, the longer it takes). Those data for more recent earthquakes reside in /gaia/smeagol/dat2/Working. Within Working are NM and ET subdirectories. Each of those have the data further distributed by month (e.g. 200607). Event subdirectories are in the respective month parent directory. As of 1/1/15 under AQMS, working directories are staged at /gaia/smeagol/dat2/Working/PP prior to being moved to /gaia/data/seisnet. Also as of the start of AQMS processing, there is a file ending in .evid for each event directory. The first field of the filename is the event id which corresponds to each entry in the CERI catalog in column 17.
The Teleseis directory is also subdivided by year and also contains a TOC file for each year. The TOC file is a table of contents detailing the events contained within each year. Events are chose somewhat subjectively and are not intended to be a comprehensive list of global earthquakes. In most cases only broadband data has been archived. For example, the 7.8Ms Turkey earthquake data are in: /gaia/data/seisnet/Teleseis/199908140030. Most teleseismic event directories contain a README file describing event parameters and expected arrivals at Memphis. Some data have been compressed to save space. Compressed data files will end in .gz. Use gunzip to uncompress (though you will have to uncompress a copy of the data).
All event directories begin with the nominal data starttime. This is not necessarily the start time of the individual traces and traces may have different start times. The correct time is in the sac header for each trace.
All traces are named by the so-called scn convention (station-channel-network). Station is the station code. Channel is the channel designator that includes information about sensor type, sample rate, sensitivity, and orientation, and follows the SEED convention. Network is the network code that designates more or less the source of the data (NM includes both CERI and St Louis, ET is CERI, US is the USGS national network).
Beginning January 1, 2009 we switched, from the scn file naming convention to the scnl convention. Now instead of using three fields, station.channel.network to identify a trace, we use four: station.channel.network.location. The location code is a two character field that conforms with the SEED location code. It is generally used when there are more than one channel of a given channel name at at station. For example, station LNXT has two strongmotion instruments so we use LNXT.HH?.NM.00 and LNXT.HH?.NM.01 to distingiush them. In most cases CERI uses 00 locotion codes as the default. Unfortunately there is no standard and different networks use different defaults and its all more or less arbitrary.
Channel names (the channel part of station.channel.network file names) follow the seed convention and are 3 letters. The third letter indicates which direction (Z for vertical, N for north/south, E for east/west). The first two letters indicate several things including the low corner, sample rate, type of instrumentation, gain, etc. The first two letter combinations commonly found at CERI include:
Additionally, within each event directory you will find additional subdirectories and non-waveform files that include:
Every event listed in this directory contain several subdirectories. After the earthquake is located traces that aren't used are moved to the "unused" directory. Traces that do not show a clear signal or show that the station is is down are moved to the "bad" directory. The "alarms" directory contains the settings (ex.-temperature) that are monitored at the nodes. "report" is a file that is included with each event. It shows a summary of the location data. If a location from the University of Saint Louis is provided, there will be a "slu.data" directory. When a minor event (3.0 and up occurs) other data is usually included in the event directory. Each set of data, if saved, will be lableled as follows: mkta.data-mkta node nmad.data-nmad node lnxt.data-lnxt.node ceri.data-ceri backbone data anss.data- strong motion data aztn.data-autozone data
A summary of local events is available via the search engines at http://folkworm.ceri.memphis.edu/catalogs/html/cat_nm.html. and the ANSS comprehensive catalog at: https://earthquake.usgs.gov/data/comcat/.
Station information is available at the IRIS MDA http://ds.iris.edu/mda/NM and http://ds.iris.edu/mda/ET. See note on response changes for NMSZ broadbands in 2006 at the end of this page.
There are some useful manipulation programs, some downloaded and some written by Withers, available in /gaia/smeagol/local/bin. (2019 update: we have decommissioned most suns including smeagol and deagol. Most of the programs have been ported to Linux and Mac, ask Mitch). The shortperiod stations are gain-ranged. That is if the amplitude approaches the clip limit, the gain is electronically reduced. An upgrade began in 2000 to add a fourth channel to track gain levels. These channels are designated by EHG. If available, two programs in /gaia/smeagol/local/bin can be used to degain. Use calldegain for the NM network and calldegain2 for the ET network. These caling scripts have the 00 location code hard-wired, so for pre-2009 data you much copy and edit to remove the location code variable from the script.
If you do not have access to /gaia, then you have a problem and either need to have your computer become supported, or use one that is. An alternative is to check the IRIS DMC.
During the course of the 2006 calendar year, the dm16 digitizers on the broadband sensors located in the NMSZ were swapped and are now dm24 digitizers. The result is greater amplitude resolution and dynamic range. There are now only 3 channels for each broadband (HH?) instead of the previous 6 channels (i.e. BL? channels are no longer necessary). The digitizer gain is thus changed from about 6.3 microVolts/count to about 1.127 microVolts/count.
In the fall of 2010, ARRA upgrades were completed on the CERI broadband stations. These are all now 6-c Reftek RT130 digitizers with Trillium T120PA broadband seismometers and Reftek RT147 accelerometers. Similar upgrades were performed at SWET, CPCT, and FPAL (though FPAL has an episensor not an RT147).
Contact Mitch if you have any questions or experience problems and especially if you find problems with the data. Believe it or not, I want to know about problems with the data so that they can be corrected.