April - June 2003


Executive Committee and Governing Council Meet
New Director Selected
Prompt Access to Collected Data
Testing New Travel Times to Improve Hypocentres

Executive Committee and Governing Council Plan Further ISC Development

The ISC Executive Committee and Governing Council each met in Sapporo, Japan during the IUGG Assembly. The Executive Committee evaluated ISC's recent work in detail and outlined plans for the next few years. The Governing Council met more briefly to review and revise the Committee's plans.

New Location Capabilities Hailed

A highlight of both meetings was iscloc, the program for computing hypocentres and magnitudes that ISC released this year. Hypocentres and magnitudes from iscloc match those from programs used to prepare the Bulletin. But iscloc's thorough documentation and modular design make it straightforward to use new travel times, attenuation relations, and algorithms. Already iscloc has been used to show that static corrections from a global crustal model would not make ISC hypocentres more accurate.

In contrast, use of S arrival times has successfully improved ISC hypocentres. Primarily thanks to local and regional S times, the ISC now more often computes both depths and epicentres from arrival times reported by all networks. Users also benefit from the ISC's specific identification of Sg, Sb and Sn phases.

The Executive Committee placed a high priority on developing a few auxiliary programs that would be required to support use of iscloc in ISC operations. After that, they recommended, the ISC should enhance iscloc to use travel times from 3-dimensional earth models. The Council endorsed this approach, and called for a workshop on hypocentral accuracy at the next IASPEI Assembly. Governing Council might then consider if the ISC should use new travel times to compute prime hypocentres.

Improving Data Collection

The Council noted that the ISC now completes routine data collection two months closer to real time than when the Council last met, and incorporates data for even the smallest earthquakes into its database. They decided that the ISC should continue aiming to improve the timeliness of data collection and the ISC's own final analysis, with a goal of completing collection about 17 months after real time by mid-2005.

The Executive Committee discussed strategies for promoting wider use of the archiving and re-distribution service that ISC now offers for catalogues and arrival time picks from research projects, such aftershock studies and other temporary deployments of on-land and ocean bottom seismic stations. Journal publishers and funding agencies will be asked to encourage authors and grant recipients to deposit catalogues in publicly accessible databases.

A figure from one of Dmitry Storchak's posters at IUGG shows effects described to the Governing Council from removing collection thresholds. ISC collected and re-distributed data for more than 180,000 earthquakes that occurred during year 2000 even though only 44,000 warranted review by ISC seismologists. Thanks partly to the use of S arrival times, the ISC was able to compute its own hypocentre for more than 80% of the reviewed events.

The meetings also featured discussion of the ISC making its own measurements from waveform segments that can be retrieved on the Internet. The ISC's initial efforts are related to a NERC-funded project to automatically measure SKS splitting. But the system that is being developed could be used to support automatic picking of secondary arrival times or to improve magnitudes by measuring more amplitudes.

Updating Data Distribution

The Governing Council confirmed its provisional decision of 2001 to stop printing the Regional Catalogue after volume 37, which had data for the events of 2000. The Council noted that users are still able to access Catalogue information readily using the range of services offered by the ISC on the Internet and from the ISC's CD-ROMS. The CD-ROMS will continue to include the Catalogue as PDF documents and as FFB files and ISF data files, which can be read from the WizMapII GUI.

The Governing Council also adopted a recommendation from the Executive Committee to reduce the size and frequency of the printed Bulletin. The suppression of arrival times and amplitudes from the printed Bulletin for the events of 1996-2001 will be extended to all events of 2002. This change facilitates cutting the frequency of printing from 6 per year to 3, making it affordable to include a CD-ROM with each printed issue. This change improves timeliness in making some data available, since the CD-ROMs include phase readings for even the smallest events.

New Director Selected

The ISC Governing Council has selected Dr Avi Shapira as the ISC's next director. Avi will assume leadership of the ISC in January 2004, when Ray Willemann plans to step aside.

Avi has extensive experience managing a scientific organisation, as director general of the Geophysical Institute of Israel (GII) since 2002 and before that as director of GII's Seismology Division since 1982. At GII Avi managed design and operation of Israel's national seismic network, national strong motion monitoring network, and a pool of portable seismic stations. He led, in parallel, development and implementation of new methods and procedures at GII to make full use of data from the networks.

Avi's wide range of experience in international seismology includes his participation in governing the ISC and the European-Mediterranean Seismological Centre. Avi was vice president of the European Seismological Commission (1996-2000), is Israel's national representative to the Asian Seismological Commission, and is a long-serving member of several IASPEI Commissions. He has participated in seismological field experiments in many regions, including Central America, the Caribbean, Central Asia, the South Pacific, and the Middle East.

Avi's expertise in seismology will be an important asset in the coming years as the ISC undertakes development to improve hypocentres and magnitudes using its new program "iscloc". Avi is an author of more than 40 peer-reviewed papers and nearly 50 invited reviews and publications in conference proceedings. Under his leadership, GII has been an important participant in efforts to improve hypocentres computed from seismic stations used to monitor the Comprehensive Test Ban Treaty.

Most importantly, perhaps, Avi brings a strong commitment to the ISC's mission and to improving ISC's services to seismologists. "The ISC has its footprints in my professional activities and career since the days I picked phases on WWSSN film records and telexed them to the prestigious Centre," says Avi. "Now ISC must explore improvements based on the development of new location methods, new travel time models, and new techniques for location error estimates."

Prompt Access to Collected Data

The ISC makes data available promptly after it receives them, which in some cases is many months before analysis for the Bulletin. Collected data are available from the ISC's web server as the recent part of the "Comprehensive Bulletin".

The ISC's data capture system automatically parses data from most e-mail, so new data may be added to recent months of the Comprehensive Bulletin at any time. Inserting data triggers a procedure to group hypocentres that are likely to be different estimates for the same event. Another procedure runs during off-peak hours to associate phase arrivals with the prime hypocentre for each event.

The figure at right shows the status of phase readings in the Comprehensive Bulletin at the end of July. At that time, the database included the results from the ISC's final analysis through 2001 September, data from NEIC's PDE through 2002 July, and data from IDC's REB 2002 December, and data from the JMA bulletin through 2003 March.

The figure reflects that

  • Apart from the REB and PDE, most readings are contributed within a few months after the events.
  • NEIC's PDE is the source of most duplicate phase readings, but also of about 20% of distinct readings.
  • ISC's automatic association had not been re-run recently on Mar - Apr 2002, so recently collected readings were not yet associated with events.
  • Otherwise, nearly all distinct phases are automatically associated with an event, and thus are included in the Comprehensive Bulletin.

The only things missing from the preliminary part of the Comprehensive Bulletin are hypocentres computed by the ISC from the union of all associated phase readings and, of course, the judgement of ISC seismologists about the accuracy of the automatic grouping and association. For example, ISC seismologists disassociate as many as 10% of the distinct phases that automatic processing had deemed might be related to particular events.

Testing New Travel Times to Improve Hypocentres

Mike Ritzwoller of the University of Colorado and Adam Dziewonski and Mike Antolik of Harvard University have contributed software to the ISC for computing travel times in 3-dimensional earth models. Istvan Bondar of the Centre for Monitoring Research verified versions of the programs delivered to the ISC.

These programs newly available to the ISC compute true minimum travel times, including the effect of lateral variations in wave speed on the geometry of ray paths. Richard Luckett of the ISC has used the programs to compute travel times through models of the crust and the upper and lower mantle that Ritzwoller and Dziewonski contributed with the software. Richard incorporated his results into functions that rapidly re-compute times by interpolating among tabulated values.

In contrast to results when he ran tests with crustal models that incorporate lateral variability, in this case Richard found that both residuals and hypocentres for ground truth events were improved.

The ISC would face many challenges if it were to use these sorts of models to compute hypocentres all around the world. From a scientific perspective, the models may lack sufficient resolution in regions without many earthquakes or local stations. Of course, earthquakes or stations are sparse in most oceans and even some broad continental regions. From a technical perspective, station-specific tables may become awkwardly large when many hundreds of stations are involved and when travel times must be tabulated for secondary phases and depth phases, both at several depths.

An example of source-specific station corrections computed from the Harvard University model. The size of each red or blue triangle represents the size of the positive or negative correction for P travel time from a source at the surface