October - December 1998

Sytem Administrator Hired
NSF Funds Development
Working with the IDC
Plans for GSN Readings
Web Site Usage Still Growing
Editing Effort Tracked
Magnitudes Assessed

Sytem Administrator Hired

To help take best advantage of its new computers, the ISC has hired an experienced system administrator. James Harris holds an honours degree in physics from the University of Manchester. Since earning his B.Sc. in 1991, James has worked in the electronics hardening division of the UK Atomic Weapons Establishment (AWE). There, James developed and maintained computer applications using a wide variety of numerical algorithms. More recently with AWE, he was responsible for administering his group's DEC computer and PCs, running Unix and Windows operating systems.

In addition to administering the Sun workstations, James has primary responsibility for installing and configuring the Oracle database management system at the ISC. "This a great opportunity gain experience with state-of-the-art information technology," says James. "I'm eager to put the skills that I've learned at AWE to use in a challenging new way."

James has his hands full just now with two further Sun workstations that arrived in December, bringing the Centre's total to seven. Each staff member is now equipped with a PC or workstation.

The Centre's newest employees, seismologist Mamy Andrianirinna and system administrator James Harris, discuss configuration of Mamy's Sun workstation.

NSF Funds Development

The US National Science Foundation has provided supplementary funding to accelerate development of ISC software. The supplement is for the salary of a developer for one year, concentrating on improving automatic association of readings with earthquakes. An extra developer should also help to free up the time of other staff members to develop tools for interactive editing. For both projects, a relational database is an important tool to implement effective new programs quickly.

Working with the IDC

Dr Igor Chernobay, Operations Manager at the Provisional International Data Centre in Vienna, visited the ISC in November. The Provisional IDC is part of the Provisional Technical Secretariat (PTS), which is building the system to monitor the Comprehensive Test Ban Treaty. Dr Chernobay's group plans a bulletin of its analysis beginning in August 1999, progressively replacing that of the Prototype IDC in Arlington, USA.

Dr Chernobay was in the UK as a member of a delegation that included Mr Wolfgang Hoffmann, Secretary General of the PTS. Working together, the IDC and the ISC could each provide greater benefits to their users, including a more comprehensive Bulletin. At a meeting in London, ISC Director Ray Willemann discussed the advantages of co-operation with Mr Hoffmann and Dr Chernobay. Partly as an outcome of this meeting and Dr Chernobay's visit, several new activities are underway. Initially, the ISC is sending the IDC a monthly letter to briefly compare the IDC and ISC locations for common events. In addition, the IDC and ISC plan to co-ordinate development of data exchange formats that meet the needs of both organisations.

Plans for GSN Readings

Digitally recorded broadband data from the Global Seismic Network are routinely archived at the IRIS Data Management Center (DMC). Despite extensive investment in operating these stations and archiving and distributing the data, many of the station records are not routinely read. Starting over the last several months, the quality assurance procedures of the IRIS Data Collection System have been modified to carefully review records of large earthquakes.

While visiting the DMC, Ray Willemann worked with Bob Woodward of the Albuquerque Seismic Laboratory and Ken Creager of the University of Washington to alter these reviews so that readings can be reported to the ISC. These large events are already among those with the most precise locations in the Bulletin. But the additional GSN readings will include amplitudes and later phases, which are reported less often, but useful for a variety of earth structure studies. In addition, Ray and DMC Director Tim Ahern discussed working co-operatively to develop CORBA-compliant object classes for seismic data.

Albuquerque Seismic Lab is using their Windows program "DIMAS" to read GSN records that will be sent to the ISC.

Web Site Usage Still Growing

More people than ever obtained data from the ISC web site during the last quarter of 1998. The number of visitors began climbing with the start of the new academic year, and the average number of on-line Bulletin searches grew from 120 in August to 380 in November and early December. Many searches are for recent data, some not yet been released on CD, but two-thirds of queries are for events from before 1990.

Currently, users can search over time periods spanning several calendar months only by repeating otherwise identical queries. New tools to streamline this frequently used search pattern are planned within a few months, after the complete Catalogue from 1964 is loaded into a relational database.

Editing Effort Tracked

The ISC has implemented a new programme to track the effort devoted to each stage of editing the Bulletin. ISC seismologists recognise that some months are much more difficult due to large aftershock sequences or other circumstances. At the same time, there is a steady increase in the average editing load due to the receipt of more data. Tracking will aid decisions on software development, including tools for on-line editing.

One early surprise is that during the first few months ISC seismologists have spent less than 10% of their time reviewing the outcome from "Search". Search is the ISC's process to generate new event hypotheses from readings that cannot be associated with reported hypocentres. Over 95% of the events in the Bulletin are from events previously located by regional agencies using subsets of the data available at the ISC. It had been thought that Search events require an inordinate fraction of editing time, and the value of the effort has been questioned. Many Search events may have been previously located, but assigned magnitudes falling below a threshold for reporting to the ISC. Tracking sets a benchmark for the performance of alternative event formation algorithms.

Magnitudes Assessed

From year to year, there are changes in the mix of stations that contribute amplitudes used to compute magnitudes at the ISC. How much does this changing mixture influence the magnitudes? The question is of special importance in comparing magnitudes from different years during the 1990's, as the GSETT-3 experiment starting in 1995 may have had a singular influence on magnitudes. Half of mb amplitudes in the 1995 ISC Bulletin are from stations that participated in GSETT-3 which, on average, had station mb values 0.3 units smaller than other stations.

Ray Willemann addressed this question in talks at the Prototype IDC, Lamont-Doherty Earth Observatory and the American Geophysical Union meeting in December. He showed that while the GSETT-3 stations influenced 1995 ISC mb values, the same set of stations was important before 1995 and had similar effects in earlier years. What's more, some other sets of stations had even larger, if less widespread, effects. Surprisingly, some individual GSETT-3 stations turn out to have systematically high station magnitudes, in contrast to the network average. Questions remain about magnitudes, but the GSETT-3 influence was not unique. Those requiring accurate comparisons of magnitude must take careful account of the influence of data selection in all cases. Slides presented at the meeting are available on the ISC web site here.

Bulletin mb values are systematically lower than they would be if GSETT-3 amplitudes were excluded.