October - December 2002

P and S, together at last!

Starting with the events of 2001, the ISC is using S arrival times in computing hypocentres. The use of S times is an important step in improving the results from ISC's data analysis. The large majority of S times being used at the ISC are local and regional phases, Sg and Sn, since relatively few teleseismic S arrival times are reported.

One issue that the ISC had to address is the well known inconsistency between P and S travel times in the J.-B. tables, which are still used at the ISC. But tests showed that, compared to hypocentres computed from P times alone, using S produced no systematic change in depth or epicentre. Based on this result, the Executive Committee concluded that S times will not bias ISC hypocentres.

True locations are not known for a large representative sample of events, so we cannot prove that S times improve ISC hypocentres. But we did compare test results to well-determined hypocentres from national monitoring agencies, which often use region-specific travel time tables. We also compared focal depths with depths computed from pP-P times. Using S times moved the ISC's hypocentres closer to these external standards on average, and moved very few of them away from the standards.

For the ISC, which uses nonlinear weighting to accommodate outliers, an important advantage from using S is improved stability in inversions for epicentres, origin times and focal depths. As shown in the figure at the right, in first three months using S arrival times, the proportion of events for which the ISC could not compute an hypocentre was cut by more than 1/3 compared with the previous year. With additional constraint from S times, spurious convergence on local minima occurs less often. In addition, ISC seismologists need less often resort to fixing depths and epicentres to reported values. This increases the proportion of events for which the Bulletin includes epicentres computed using the ISC's globally consistent location procedures, weighting algorithm, and travel times.

S times were introduced by modifying the ISC's existing location program. Many of the modifications were made by Chen Qi-fu of the China Seismological Bureau, who worked at the ISC for a year as a Royal Society visiting fellow. Richard Luckett made further changes to improve identification of regional S phases and ensure that S travel times are computed accurately.

Seismologists who follow developments at the ISC may recall that Chen Qi-fu's goal in modifying the ISC location program was to use travel times from laterally heterogeneous earth models. Consensus on a 3-dimensional mantle model may still be a few years off but, apart from the effects of subducted lithosphere, much of the variability in travel times is due to crustal structure. But Richard Luckett is investigating the effect that crustal corrections would have on ISC hypocentres. Results from his work using the Mooney-Laske "Crust 2.0" model are likely to be discussed during the ISC's Governing Council meeting at the IUGG Symposium in June.

2000 Bulletin Completed

In early November the ISC completed its analysis of the seismic events of 2000.

The total number of events for which the ISC collected data grew from 152,000 in 1999 to 195,000 in 2000. The increase was partly due to the Honshu "mega-swarm", which resulted in JMA reporting twice the normally expected number of events for several months. But much of the increase was due to improved collection, which included establishing regular contributions from agencies in Korea, Ecuador, Costa Rica, Syria, Saudi Arabia, Iran, Vietnam and more than a dozen other countries around the world. The ISC also reached agreements for more complete bulletins from long-time contributors, including mining-related events from the Geological Surveys of Canada and the USA, and teleseismic readings from the Geophysical Institute of Israel.

Apart from Japan, the most prolific swarm reported for 2000 was on the East Pacific Rise, but almost all of these events were too small to be recorded except by hydrophones. The seismicity of 2000 was also notable for numerous aftershocks following large earthquakes in Sumatra during June and near New Britain during December.

The ISC significantly improved its system for selecting the events of 2000 that were reviewed manually. The new system takes advantage of the ISC's own preliminary processing, which groups hypocentres and associates readings with events. Even if no magnitude is known, the ISC's processing now gives us a good idea of how widely each event was recorded. Most events reported without a magnitude are too small to be detected beyond a single network, and these are no longer reviewed. ISC seismologists had to check only 45,000 events of 2000 to ensure that all reported M>=3 events and all multiply-recorded events were properly analysed. By comparison, 52,000 events of 1999 were manually reviewed at the ISC.

New CDs Include GUI

The ISC's annual Bulletin and Catalogue CDs were released soon after analysis of December 2000 was completed. As last year, the new CDs include Bulletin files for the latest year in IASPEI Seismic Format (ISF). In addition, however, this year's CDs include ISF Catalogue files back to 1904. All of the data are also included in ISC's own "FFB" format, although these files are compressed.

An exciting feature of the new CDs is a version of "Wizmap II". Roger Musson of the British Geological Survey (BGS) wrote the program, which runs under Microsoft Windows, and modified it to use ISF files. Wizmap II can be used

to make maps of epicentres from a catalogue, select subsets of those events by drawing polygons on the map, and analyse the magnitude distribution of the selected events. Wizmap II can write an ISF file of the selected events, so it can also be used as a Graphical User Interface (GUI) to select events for other programs that read ISF files.

The ISC wrote files for the new CDs by selecting data from its database rather than re-formatting the FFB files from previous CDs. The advantage of this approach is that it includes recently collected data for older events. The most widespread newly collected data are the Engdahl, van der Hilst and Buland (EHB) hypocentres for large events from 1964 to 1999 and located all over the world. The most numerous new data are the PMEL hydroacoustic epicentres in the Pacific Ocean during 1996 to 2000. Some newly collected data on the CDs are events recorded only by temporarily deployed seismometers, while others are new source parameters for events that were already in the Bulletin.

Want an ISC CD?

ISC's web site, http://www.gxjydl.com/, has several pages that may be of interest to CD users. To buy CDs for libraries, the order form at /doc/sales/cds_sales.html can be posted or faxed to the ISC. A pair of CDs comprising the 2000 Bulletin and

1904 - 2000 Catalogue can be purchased for individual use for £10 at /doc/sales/on-line.html. Authors who have published recent papers that use data from the ISC can get a free copy of the Catalogue CD by completing the on-line profile at /users/profile.htm. Enquiries by telephone, fax, post and e-mail are also welcome.

ISC Posts Maps on the Web

The ISC has started providing custom-made maps to its web-site users, using software contributed by Mission Research Corporation and Science Applications International Corporation. Users can obtain a map of the events that they selected in an on-line query by clicking on the phrase "Make a map!", which is now printed near the top of each set of selected data posted by the ISC web server.

For now the maps only plot a single symbol at the prime epicentre for each event, using fixed scales for symbol size with magnitude and symbol colour with depth. But the software could be used to create "clickable" maps that might retrieve further information about individual events, or to prepare other maps such the stations that recorded an event or several epicentres for one event with their uncertainty ellipses.

Users can comment on the utility of the service or suggest maps that might be widely used by sending e?mail to suggestions@isc.ac.uk.

Project to Transform "Editing" Concludes Successfully

Collaboration between ISC and Reading University has resulted in a transformation in the way that ISC seismologists work.

Tom Bayliss of Reading's Department of Computer Science worked at the ISC for two years to develop a new editing system. The project was funded by the Teaching Company Scheme, which is a UK Government programme that aims to spread the use of British academic expertise beyond universities.

Tom is demonstrating his system in the photo at the right, using a bar code scanner to record edits. Commands are scanned from a summary page. Objects of each command, such as hypocentres or readings, are scanned from bar codes that are now printed in listings. Advice from Reading faculty with expertise in human-computer interaction helped in designing a listing format that could be scanned efficiently.

The "coolest" part of the system comes into play when a seismologist makes a custom entry and links it to a particular event. Tom built a simple, platform-independent of carrying out this complex task by using a web browser to interact with the seismologist and an internal web server to record the entry.

An immediate benefit from the system is to eliminate the need for manually keying edit commands. This both frees up staff for other tasks and eliminates a potential source of error in analysis.

Further advantages are expected over time. The system might be extended, for example, to display a map of stations that recorded (or failed to record) the event that a seismologist is editing. Further down the road, the system might be modified to include true decision support functions, such as showing results for provisional editing or suggesting the depth at which a poorly constrained event should be fixed.

Friends of Robin Adams, ISC's senior seismologist from 1978 to 1995, may recall that he sometimes brought a listing home to edit in the evening while watching the telly. Does the new system make such "family-friendly" editing a thing of the past? Temporarily, perhaps. But in light of ongoing developments in mobile computing, who could doubt that ISC seismologists will someday use wireless scanners at home? Perhaps they will pop up a window on their digital TV from time to time to display a map, record a custom entry, or even consult with other seismologists!