International Seismological Centre Bulletin

Raymond J. Willemann, Director, International Seismological Centre

A typical bimonthly issue of the ISC Bulletin lists more than 10,000 earthquakes and 300,000 associated readings.

As recently as the 1960's, seismological investigations often began by compiling readings from several global bulletins and re-computing epicentres, depths and magnitudes using the combined data. The global seismic bulletins available in the 1960's included the International Seismological Summary (ISS) by the British Geological Survey in Edinburgh, the Preliminary Determination of Epicenters by the US Coast and Geodetic Survey in Washington, the Bulletin of the Bureau Central International de Seismologie in Strasbourg, and the Seismological Bulletin of the Soviet Academy of Sciences in Moscow. The International Seismological Centre (ISC) was founded in 1964 to compile readings from these and other sources efficiently and routinely to a high standard, allowing individual seismologists working from the ISC Bulletin to concentrate on further investigation and interpretation.

There were few international precedents for organizing such work, and substantial commitments for direct support were required from the founding participants. Seismologists from national agencies of the US, UK, SovietUnion, France, Canada, Australia and New Zealand agreed to provide their data freely to the new Centreno small feat in an era of growing efforts to monitor nuclear weapon tests. In addition, each participant made a commitment to help support the ISC financially so that the new Bulletin could be offered at low cost, thus promoting wider distribution.

As the number of seismic networks and regional agencies around the world has grown, the ISC has always aimed to provide a comprehensive global summary. The Bulletin now includes more than 60,000 events annually.

The result of these efforts was profound: more than 350,000 phases were used to compute the locations of 10,770 events in 1964, the first year of ISC Bulletins. This was a better-than three-fold increase from the number of events reported in the 1963 ISS. Within a few years, the ISC was more formally organized with a set of working statutes that UNESCO helped to formulate, and a Governing Council of representatives from each member agency contributing significant financial support. Each of the founding members has continued supporting the ISC for more than 30 years, joined regularly by agencies from other countries. Today the ISC receives direct financial contributions from nearly 50 countries and readings from an even broader cross-section of the global seismological community.

Thanks to this comprehensive commitment, the ISC now offers a Bulletin comprised of more than 60,000 events annually as illustrated here.

Preparing the ISC Bulletin requires international collaboration. Seismologists at individual stations and national or regional agencies read records, associate phase data within networks, and compute preliminary hypocentres. Thus, there is no "ISC network" of stations. Instead, the ISC has always aimed to be as comprehensive as possible, associating and reporting data from otherwise under-represented regions whenever they are offered. The ISC does maintain an International Registry of stations but, in order to encourage the broadest possible reporting, imposes only minimal requirements for registration. The number of registered stations has grown to nearly 10,000, of which more than 3,000 have contributed readings in each recent year. The location of these stations is shown on page 6.

The nine scientists on staff at the ISC carry out the final work of grouping independent hypocentres for the same events, re-associating phase readings, and re-computing hypocentres. The staff is comprised of experts from around the world. Seismologists working currently or within the last three years at the ISC have come from New Zealand, Germany, Russia, Canada, the United States, the Malagasy Republic, Algeria, China and the Philippines.
Britain has offered a variety of in-kind support in addition to its financial contributions. The ISC was originally housed in the Edinburgh facility of the British Geological Survey, and moved to southern England in the early 1970's to take advantage of computing facilities at Rutherford-Appleton Laboratory. Today the ISC works in a building it purchased in 1987 in the small town of Thatcham and its home institution is the University of Oxford, whose faculty members have offered critical programming and other support in recent years.

The format of the printed Bulletin has changed over the years to offer the ever-growing volume of readings in an affordably compact layout. For events since 1995, the printed Bulletin is issued bi-monthly and readings are included only for shallow events larger than mb 5.5 and deeper events larger than mb 5.0. The ISC has always made its data available in computer-readable formats, formerly on custom-written tapes but now on a set of standard CDs. A new CD is issued annually with the Bulletin data (hypocentres and phase readings) for a new year and Catalogue data (hypocentres only) for the entire history of the ISC. For the sake of compatibility with the existing programs of seismologists around the world, the ASCII files on the CDs continue to be formatted in "96-column fixed format" used in ISC tapes for several decades. The ISC also operates a web site (www.isc. ac.uk). In the original spirit of making the Bulletin available at the cost of reproduction, the complete set of ISC Bulletin data are freely available from the ISC web site. In an effort to foster more widespread use of modern standards, the on-line Bulletin is offered in IMS 1.0 format.

There is no formal threshold for including events in the Bulletin. The ISC publishes a location for each event that it can locate with a reasonable degree of reliability using all of the reported readings. Because the ISC aims to be comprehensive, the number of events in the Bulletin has grown over time. Unfortunately, many agencies reporting to the ISC fail to include amplitude measurements so the ISC is able to compute mb or MS for only about 25% of the events it can locate. Based on this minority of events, the practical threshold of the Bulletin estimated from magnitude-frequency relations varies from one region to another as illustrated on the map on the next page. The threshold for including events in the Bulletin without a computed magnitude is significantly better, but also variable from one place or one year to another.

Despite the inevitable changes in coverage over time as seismic stations are opened and closed, the ISC has maintained a uniformity in its processing in order to make it as straightforward as possible to compare earthquakes from different years. Thus, the ISC continues to use the Jeffreys-Bullen travel time tables and to compute hypocentres exclusively from first-arrival times. Similarly, magnitudes


Much of the growth in reporting through the early 1990's occurred in already well-monitored regions. Further improvements are required in many parts of the world, including some areas with moderate or high seismic hazard.

Senior seismologist Dmitry Storchak (left) is responsible to uphold editing standards for the Bulletin. He works with seismologists on 2 to 3 year appointments at the ISC, including Mamy Andrianarinna (right) of the Malagasy Republic.

The ISC's threshold for computing mb varies with the density of reporting stations and on how comprehensively their records are read and reported. Only about 25% of the events reported to the ISC include magnitude estimates. The threshold based on this fraction is shown above.

continue to be computed from teleseismic (21­100 for mb, 21­160 for MS) amplitudes using the Gutenberg-Richter attenuation tables. The Bulletin in print, on CDs and on the Internetalso continues to include macroseismic data (intensities, felt and damage reports) and other widely recognized event parameters, such as the Harvard Centroid-Moment Tensor solutions.

While maintaining continuity remains important, advances in seismological practice and information technology have made it important and practical to begin updating the Bulletin. Building on the tradition of drawing together efforts around the world, the ISC is reporting more source parameters computed elsewhere, such as radiated energy, stress parameters and source time functions. Starting in 2000, the ISC is testing new location procedures, including alternative algorithms, additional phases, and more accurate travel times. The ISC is also working towards making reported readings and preliminary hypocentres available on its web site before they are analyzed in the Bulletin. Recognizing that advances in earthquake physics and earth structure often rely on specialized measurements from raw waveforms, the ISC is working to integrate its Bulletin with waveform archives, for example by implementing the AutoDRM software developed at Swiss Seismological Centre and the NetDC software developed at the IRIS Data Management Center.

Phase picks and event catalogues are likely to remain the starting point for many seismological investigations for the foreseeable future. By providing this information and up to date services, the ISC will continue as an important resource to seismologists and earthquake engineers for years to come.

Additional Information

For additional information about the ISC and their bulletins, see the ISC website: www.isc.ac.uk and Willemann, R.J., Regional Thresholds of the ISC Bulletin, Seismol. Res. Lett., 70, 313-321, 1999.

The threshold for completeness of the ISC Bulletin varies from place to place, as well as from year to year. The "roll-off" of the global magnitude-frequency distribution (shown) between mb 4.2 and 4.8 shows that the improvement in the threshold from 1994 to 1995 was followed by a further improvement in 1996.
At higher magnitudes, however, the 1996 distribution is offset from earlier years, and the 1997 distribution differs even more. The offset is far too large to be attributed to natural variations in seismicity; it is implausible that the typical number of mb 5.0 earthquakes was nearly 400 through 1995, but that there were fewer than 200 in 1997. The alternative is that the 1996 and 1997 curves are offset not downward, but leftward. Looking at the offset this way suggests that recent mb values are biased. That is, earthquakes are being assigned smaller magnitudes than in the past.

Magnitude calculation procedures have not changed at the ISC, so the most likely cause of a bias may be that the mixture of stations reporting amplitudes to the ISC is changing, with newer stations more likely to report small amplitudes. This could be a result of the growing number of stations located on carefully selected low-noise sites. The same properties that reduce noise at these sites might also turn out to reduce signal amplitudes. Alternatively, the low noise stations may be simply more likely to report amplitudes of small signals that previously would have been lost in the noise.